Contact Us | Intranet | Search
 

CI-Tutor Evaluation Results

AG00:Overview for Decision Makers
Total number of evaluations submitted for this course is: 57
This web-based course as a whole was:
ResponseCountPercent
Fair 1 1.8%
Good 21 36.8%
Very Good 22 38.6%
Excellent 13 22.8%
The effectiveness of the web-based format was:
ResponseCountPercent
Fair 3 5.3%
Good 18 31.6%
Very Good 25 43.9%
Excellent 10 17.5%
No response 1 1.8%
The quality of the course content was
ResponseCountPercent
Fair 4 7%
Good 19 33.3%
Very Good 20 35.1%
Excellent 13 22.8%
No response 1 1.8%
The organization of the course was
ResponseCountPercent
Fair 1 1.8%
Good 20 35.1%
Very Good 19 33.3%
Excellent 16 28.1%
No response 1 1.8%
Usefulness of self-tests was
ResponseCountPercent
Not Applicable 31 54.4%
Good 10 17.5%
Very Good 9 15.8%
Excellent 5 8.8%
No response 2 3.5%
Usefulness of references provided was:
ResponseCountPercent
Not Applicable 12 21.1%
Poor 2 3.5%
Fair 2 3.5%
Good 15 26.3%
Very Good 17 29.8%
Excellent 6 10.5%
No response 3 5.3%
Usefulness of audio/video resources was:
ResponseCountPercent
Not Applicable 22 38.6%
Very Poor 1 1.8%
Poor 2 3.5%
Fair 3 5.3%
Good 11 19.3%
Very Good 5 8.8%
Excellent 8 14%
No response 5 8.8%
What was your knowledge of this topic before this course?
ResponseCountPercent
No knowledge at all 23 40.4%
Basic knowledge 19 33.3%
Intermediate knowledge 7 12.3%
Advanced knowledge 3 5.3%
No response 5 8.8%
What is your knowledge of this topic now?
ResponseCountPercent
Basic knowledge 24 42.1%
Intermediate knowledge 22 38.6%
Advanced knowledge 6 10.5%
No response 5 8.8%
Do you think the knowledge you gained will improve the quality of your work?
ResponseCountPercent
No response 5 8.8%
No 4 7%
Yes 48 84.2%
Would you take another web-based course?
ResponseCountPercent
No response 4 7%
No 2 3.5%
Yes 51 89.5%
Why or why not?
Yes I would to learn more, computer is always a learning thing to stay on top of the field.
I would like to get more information about the subject to apply to my work.
To continue learning about Access Grid
because i want and need to be updated with the national centre for supercomputing applications.
Another way to learn
They are fairly useful
I think it is a very good to share information
I would take another course - because, though slow, it covers the topic well.
Yes because of the convenience involved.
It is very convenient.
Because I need the information.
Found the first course easy to assimilate, and I believe taking further courses will enhance my knowledge even more; giving me the ability to build, debug, and operate grid systems at the speed and quality I expect from myself and improve my skills the way employers also appreciate.
Allow the access of educational material from different locations
to improve my self and to serve my organization SESAME
 
What did you like most about this course and why?
Organization of the course
the video was the best, it explained everything for the basic knowledge. Now to learn more.
It provide useful information.
the course provided the basics understanding of the technology and how it would be implemented
ACCESS GRID is a supercomputing development technolgy and i think developing it is a need .
Topic I am interested in
My interest in the topic was piqued by an inquiry for services which I received. I appreciate very much that this information is made available, open, as it is to the educational community. The WebCT format, in this my first user experience, is effective in content delivery.
The content. It was of easy understanding
The way it builds up from the basics
Short and to the point
It was fairly well organized.
The information is formatted in portions just the right size for the subject matter at hand, and it's organized in a way that makes it easy to remember. Also including vocbulary links makes it much faster to read through and understand.
The course highlights the potential of the AG technology
on-line sound conversation that's to be some people responsible for users questions
Easy understanding.
 
What did you like least about this course and why?
none
nothing...
nothing leats just por sound audio.
The link to Community resources did not work and that was my most important question.
It could use sounds.
Slowness and not describing the overview of the systm (such as venues and nodes) and their relationships
It could have one practical example
No obvious dislikes
Very clean design and writing
A little sparse. I could deduce most of the functions from the basic structure of the system.
The photos of people working on a grid system are atractive. They give your eyes and brain a break, allowing you time to absorb the material. However, graphic presentations of the subject matter would be far more valuble when it comes time to recall usefull information.
Links to audio and video presentations as well agoo number of images.
 
Additional Comments:
keep up the good work
thanks to NCSA. IFEANYI-CHUKWU MARK-JESUS [ugbor} is my name
No comments
Under acknowledgements, you need Kevin's last name. For monitoring yourself (when no one is available to get in a room with you), it was actually helpful when the Sandiego Supercomputing Center left their node in the test room with echo cancellation turned off. That way you could verify you were receiving audio fine and you could also get an idea of the quality of your audio as heard by others.
There may not be a better way of presenting the material.
I know I learned something from this course but I'm not certain how much, as I wasn't offered the self-test. Plus, because my computer is really slow I skipped the Video, fearing it wouldn't run properly. As a teacher I realize the importance of graphic reinforcement. Hopefully you will too.
The course should differentite the AG technology from other technologies and try to introduce advnatages over real time video streaming and video conferencing.
 
AG01: How to Build and Install an Access Grid Node (AGN)
Total number of evaluations submitted for this course is: 19
This web-based course as a whole was:
ResponseCountPercent
Poor 1 5.3%
Fair 1 5.3%
Good 3 15.8%
Very Good 5 26.3%
Excellent 9 47.4%
The effectiveness of the web-based format was:
ResponseCountPercent
Poor 1 5.3%
Fair 1 5.3%
Good 3 15.8%
Very Good 3 15.8%
Excellent 11 57.9%
The quality of the course content was
ResponseCountPercent
Poor 1 5.3%
Fair 2 10.5%
Good 4 21.1%
Very Good 4 21.1%
Excellent 8 42.1%
The organization of the course was
ResponseCountPercent
Fair 1 5.3%
Good 5 26.3%
Very Good 4 21.1%
Excellent 9 47.4%
Usefulness of self-tests was
ResponseCountPercent
Not Applicable 9 47.4%
Poor 1 5.3%
Fair 1 5.3%
Good 2 10.5%
Very Good 1 5.3%
Excellent 4 21.1%
No response 1 5.3%
Usefulness of references provided was:
ResponseCountPercent
Not Applicable 1 5.3%
Very Poor 1 5.3%
Poor 1 5.3%
Fair 3 15.8%
Good 1 5.3%
Very Good 5 26.3%
Excellent 7 36.8%
Usefulness of audio/video resources was:
ResponseCountPercent
Not Applicable 2 10.5%
Poor 1 5.3%
Fair 3 15.8%
Good 3 15.8%
Very Good 4 21.1%
Excellent 5 26.3%
No response 1 5.3%
What was your knowledge of this topic before this course?
ResponseCountPercent
No knowledge at all 8 42.1%
Basic knowledge 9 47.4%
Intermediate knowledge 2 10.5%
What is your knowledge of this topic now?
ResponseCountPercent
Basic knowledge 11 57.9%
Intermediate knowledge 8 42.1%
Do you think the knowledge you gained will improve the quality of your work?
ResponseCountPercent
No response 1 5.3%
No 2 10.5%
Yes 16 84.2%
Would you take another web-based course?
ResponseCountPercent
No response 4 21.1%
Yes 15 78.9%
Why or why not?
I found this very useful
I have to master collaboration and I have very little time to do it in.
It was GREAT !!
Web courses allow me to learn at my own pace. (skip the things I know and focus on the things that I don't)
Covenience. I'm at home with a laptop. No driving, parking etc.
 
What did you like most about this course and why?
The ease at which the lesson are delivered
The small windows with basic concepts the URLs
It arranged everything I needed to know about the basic setup of the AccessGrid.
Easy and self paced
The attention to detail
Covered tecnical information with specific step-by-step instructions including components with links to manufacturers' web sites.
 
What did you like least about this course and why?
The discussion introduces jargon without explanation!!what os a vv server? what is arm ?The are/is (at least one) error in text. Section 1.6.2 is referencedprobably should have been 6.2Section6.2 is full of dead links --- makes the entire discussion some what moot. There is sw that will automatically detect dead links !!
nothing
There was nothign to dislike.
Most of the links to outside resources need to be updated because they are broken.
Can't thing of anything annoying.
 
Additional Comments:
The link to MuralGeom is broken
It makes this project look disorganized to have so many broken links in the tutorial.
Now I have to build this Access Grid system for ICC.
 
AG02: How to Operate an Access Grid Node (AGN)
Total number of evaluations submitted for this course is: 7
This web-based course as a whole was:
ResponseCountPercent
Good 4 57.1%
Very Good 1 14.3%
Excellent 2 28.6%
The effectiveness of the web-based format was:
ResponseCountPercent
Good 4 57.1%
Very Good 2 28.6%
Excellent 1 14.3%
The quality of the course content was
ResponseCountPercent
Fair 1 14.3%
Good 2 28.6%
Very Good 1 14.3%
Excellent 3 42.9%
The organization of the course was
ResponseCountPercent
Good 6 85.7%
Very Good 1 14.3%
Usefulness of self-tests was
ResponseCountPercent
Not Applicable 2 28.6%
Good 3 42.9%
Very Good 1 14.3%
No response 1 14.3%
Usefulness of references provided was:
ResponseCountPercent
Good 4 57.1%
Very Good 2 28.6%
No response 1 14.3%
Usefulness of audio/video resources was:
ResponseCountPercent
Fair 2 28.6%
Good 2 28.6%
Very Good 1 14.3%
No response 2 28.6%
What was your knowledge of this topic before this course?
ResponseCountPercent
No knowledge at all 4 57.1%
Basic knowledge 3 42.9%
What is your knowledge of this topic now?
ResponseCountPercent
Basic knowledge 4 57.1%
Intermediate knowledge 2 28.6%
Advanced knowledge 1 14.3%
Do you think the knowledge you gained will improve the quality of your work?
ResponseCountPercent
Yes 7 100%
Would you take another web-based course?
ResponseCountPercent
Yes 7 100%
Why or why not?
I want to learn more. It would be helpful if abreviations such as moo would be defined.
I would like to learn the whole process involved in the Access Grid environment
Access grid being deployed into new campuses. Further training will be required for support personnel to successfully assist non-technical customers using AG.
This format is very convenient
 
What did you like most about this course and why?
AG Nodes seem to be the future in communications.
It was pointed to the right questions, and it had areas that one could get more information if they wish. Of particular interest were the "definitions".
It was my first. It was quite informative.
Simple lay out and quick and easy access to glossary and AG terms.
Format and organization was good, and easy to follow.
 
What did you like least about this course and why?
Not enough examples
I cant think of anything
no sound option.
Pages are rather busy.
Technical level was extremely low.
 
Additional Comments:
Maybe later
It was good. I will visit again. I am learning this technology to increase my desirable criteria count to qualify for a position as staff at a University.
 
AG03: How to Produce an Access Grid Event
Total number of evaluations submitted for this course is: 1
This web-based course as a whole was:
ResponseCountPercent
Excellent 1 100%
The effectiveness of the web-based format was:
ResponseCountPercent
Good 1 100%
The quality of the course content was
ResponseCountPercent
Excellent 1 100%
The organization of the course was
ResponseCountPercent
Excellent 1 100%
Usefulness of self-tests was
ResponseCountPercent
Not Applicable 1 100%
Usefulness of references provided was:
ResponseCountPercent
Not Applicable 1 100%
Usefulness of audio/video resources was:
ResponseCountPercent
Not Applicable 1 100%
What was your knowledge of this topic before this course?
ResponseCountPercent
Basic knowledge 1 100%
What is your knowledge of this topic now?
ResponseCountPercent
Intermediate knowledge 1 100%
Do you think the knowledge you gained will improve the quality of your work?
ResponseCountPercent
Yes 1 100%
Would you take another web-based course?
ResponseCountPercent
Yes 1 100%
Why or why not?
only soure of this material
 
What did you like most about this course and why?
 
What did you like least about this course and why?
 
Additional Comments:
 
AG07: Access Grid Audio Troubleshooting
Total number of evaluations submitted for this course is: 2
This web-based course as a whole was:
ResponseCountPercent
Poor 1 50%
Very Good 1 50%
The effectiveness of the web-based format was:
ResponseCountPercent
Good 2 100%
The quality of the course content was
ResponseCountPercent
Very Poor 1 50%
Very Good 1 50%
The organization of the course was
ResponseCountPercent
Fair 1 50%
Good 1 50%
Usefulness of self-tests was
ResponseCountPercent
Fair 1 50%
Good 1 50%
Usefulness of references provided was:
ResponseCountPercent
Not Applicable 2 100%
Usefulness of audio/video resources was:
ResponseCountPercent
Not Applicable 1 50%
Good 1 50%
What was your knowledge of this topic before this course?
ResponseCountPercent
No knowledge at all 1 50%
Intermediate knowledge 1 50%
What is your knowledge of this topic now?
ResponseCountPercent
Basic knowledge 1 50%
Intermediate knowledge 1 50%
Do you think the knowledge you gained will improve the quality of your work?
ResponseCountPercent
No 1 50%
Yes 1 50%
Would you take another web-based course?
ResponseCountPercent
Yes 2 100%
Why or why not?
problem unsolved
 
What did you like most about this course and why?
 
What did you like least about this course and why?
diagnosis was mostly obvious things. where is the detail?
 
Additional Comments:
 
AG09: Collaborating over the Access Grid
Total number of evaluations submitted for this course is: 3
This web-based course as a whole was:
ResponseCountPercent
Very Good 3 100%
The effectiveness of the web-based format was:
ResponseCountPercent
Poor 1 33.3%
Very Good 2 66.7%
The quality of the course content was
ResponseCountPercent
Good 2 66.7%
Very Good 1 33.3%
The organization of the course was
ResponseCountPercent
Good 1 33.3%
Very Good 2 66.7%
Usefulness of self-tests was
ResponseCountPercent
Not Applicable 1 33.3%
Fair 1 33.3%
No response 1 33.3%
Usefulness of references provided was:
ResponseCountPercent
Fair 1 33.3%
Very Good 1 33.3%
No response 1 33.3%
Usefulness of audio/video resources was:
ResponseCountPercent
Not Applicable 1 33.3%
Fair 1 33.3%
No response 1 33.3%
What was your knowledge of this topic before this course?
ResponseCountPercent
No knowledge at all 1 33.3%
Advanced knowledge 2 66.7%
What is your knowledge of this topic now?
ResponseCountPercent
Basic knowledge 1 33.3%
Advanced knowledge 2 66.7%
Do you think the knowledge you gained will improve the quality of your work?
ResponseCountPercent
Yes 3 100%
Would you take another web-based course?
ResponseCountPercent
Yes 3 100%
Why or why not?
I need to know more about what AG can support,and how I can use some of its functions to support my job!
convenience, easy to use
 
What did you like most about this course and why?
why AG need collabortion
 
What did you like least about this course and why?
if it can privde more cases,it 's better
 
Additional Comments:
I want to use AG to develop new collaboration through portal,but I don't know what to do. I need more technical details about how to develop collaboration apllication using by portal
 
AG10: Introduction to AG Toolkit 2.x
Total number of evaluations submitted for this course is: 13
This web-based course as a whole was:
ResponseCountPercent
Fair 1 7.7%
Good 6 46.2%
Very Good 6 46.2%
The effectiveness of the web-based format was:
ResponseCountPercent
Fair 1 7.7%
Good 4 30.8%
Very Good 4 30.8%
Excellent 4 30.8%
The quality of the course content was
ResponseCountPercent
Fair 1 7.7%
Good 7 53.8%
Very Good 4 30.8%
Excellent 1 7.7%
The organization of the course was
ResponseCountPercent
Fair 2 15.4%
Good 4 30.8%
Very Good 7 53.8%
Usefulness of self-tests was
ResponseCountPercent
Not Applicable 4 30.8%
Poor 1 7.7%
Fair 3 23.1%
Good 2 15.4%
Very Good 2 15.4%
Excellent 1 7.7%
Usefulness of references provided was:
ResponseCountPercent
Not Applicable 2 15.4%
Poor 1 7.7%
Fair 3 23.1%
Good 2 15.4%
Very Good 4 30.8%
No response 1 7.7%
Usefulness of audio/video resources was:
ResponseCountPercent
Not Applicable 11 84.6%
Good 2 15.4%
What was your knowledge of this topic before this course?
ResponseCountPercent
No knowledge at all 2 15.4%
Basic knowledge 6 46.2%
Intermediate knowledge 4 30.8%
Advanced knowledge 1 7.7%
What is your knowledge of this topic now?
ResponseCountPercent
Basic knowledge 4 30.8%
Intermediate knowledge 8 61.5%
Advanced knowledge 1 7.7%
Do you think the knowledge you gained will improve the quality of your work?
ResponseCountPercent
No 4 30.8%
Yes 9 69.2%
Would you take another web-based course?
ResponseCountPercent
No response 2 15.4%
Yes 11 84.6%
Why or why not?
To expand my knowledge.
Knowledge is power!
Because it is often the only available option.
I live in hope that I can actually find some useful documentation somewhere
It's online when you need it and cost-effective. Con is the preference for IE. It should work on standard (X)HTML browser if possible. Mozilla on Linux at least if not Opera.
I am interested in Access Grid and online courses are good tutorials to start to get the knowledge enough to understand the features of the system.
Yes, whenever it helps to improve my work by self studying.
To reinforce what Iv'e already learned, and increase my level of computer skills and knowledge even further.
convenience, easy to use/schedule
 
What did you like most about this course and why?
Introductory info because I'm a novice.
It is easy to access and always available.
It assumed that you knew almost nothing
Compact
It is easy to understand. That is a very good point.
It was easy to follow.
Compact and hands-on approach
The small addition of graphics helped.
this course will be used as a reference when working with people interested in acquiring the AG software to establish a node and/or people using the AG for meeting.
 
What did you like least about this course and why?
Nothing to compare to; give me time, though!
It describes what the Access Grid is, but not how to use it. I took this with the intention of learning how to use the AG software because I can learn about what it is from their own web page.
It didn't teach me anything I wanted to know
Not enough pictures giving a good view of the whole display area to get a feeling what it really looks like...
Room based seems to assume multinode and personal single node even though that may not be the case. No answer for what the other types of certificates besides personal are used for. And is there a difference between personal and room based except perhaps in the fact the other might have better EC and more cameras covering a whole room? No mention on data and application setup, maybe that's not for a novice?
The lack of graphic representation.
 
Additional Comments:
Thnx.
Keep up the good work
Or, if you prefer to work from the command line, it is also possible to use a small program, certmgr.py. WHERE does this parenthesis end? (This program is included in AG 2.1.2 and higher; if you need to export a certificate from an installation of an earlier version you can download it. If you are using Windows, save this file into the directory: /Program Files/Access Grid Toolkit/bin on the machine on which you have a certificate already installed; on linux you may save it to any convenient directory.
Once again I wasn't given the self test, and I've been left with a very weak mental picture of a grid system.
 
AG11: AG 2.1 2 on Personal and Room-Based Nodes
Total number of evaluations submitted for this course is: 2
This web-based course as a whole was:
ResponseCountPercent
Good 1 50%
Excellent 1 50%
The effectiveness of the web-based format was:
ResponseCountPercent
Good 2 100%
The quality of the course content was
ResponseCountPercent
Good 1 50%
Very Good 1 50%
The organization of the course was
ResponseCountPercent
Fair 1 50%
Excellent 1 50%
Usefulness of self-tests was
ResponseCountPercent
Poor 1 50%
Fair 1 50%
Usefulness of references provided was:
ResponseCountPercent
Good 1 50%
Very Good 1 50%
Usefulness of audio/video resources was:
ResponseCountPercent
Not Applicable 1 50%
Poor 1 50%
What was your knowledge of this topic before this course?
ResponseCountPercent
Basic knowledge 2 100%
What is your knowledge of this topic now?
ResponseCountPercent
Intermediate knowledge 2 100%
Do you think the knowledge you gained will improve the quality of your work?
ResponseCountPercent
Yes 2 100%
Would you take another web-based course?
ResponseCountPercent
Yes 2 100%
Why or why not?
They are available whenever I want them.
 
What did you like most about this course and why?
It's ease of use because I am pressed for time.
 
What did you like least about this course and why?
The images used to explain various steps did not match those on my software.
 
Additional Comments:
 
Debugging Serial and Parallel Codes
Total number of evaluations submitted for this course is: 8
This web-based course as a whole was:
ResponseCountPercent
Good 1 12.5%
Very Good 1 12.5%
Excellent 6 75%
The effectiveness of the web-based format was:
ResponseCountPercent
Good 2 25%
Very Good 1 12.5%
Excellent 5 62.5%
The quality of the course content was
ResponseCountPercent
Good 1 12.5%
Very Good 1 12.5%
Excellent 6 75%
The organization of the course was
ResponseCountPercent
Good 1 12.5%
Excellent 7 87.5%
Usefulness of self-tests was
ResponseCountPercent
Not Applicable 1 12.5%
Fair 2 25%
Very Good 2 25%
Excellent 3 37.5%
Usefulness of references provided was:
ResponseCountPercent
Good 3 37.5%
Very Good 1 12.5%
Excellent 4 50%
Usefulness of audio/video resources was:
ResponseCountPercent
Not Applicable 4 50%
Fair 1 12.5%
Excellent 3 37.5%
What was your knowledge of this topic before this course?
ResponseCountPercent
No knowledge at all 1 12.5%
Basic knowledge 3 37.5%
Intermediate knowledge 3 37.5%
Advanced knowledge 1 12.5%
What is your knowledge of this topic now?
ResponseCountPercent
Basic knowledge 1 12.5%
Intermediate knowledge 2 25%
Advanced knowledge 4 50%
No response 1 12.5%
Do you think the knowledge you gained will improve the quality of your work?
ResponseCountPercent
Yes 8 100%
Would you take another web-based course?
ResponseCountPercent
No 1 12.5%
Yes 7 87.5%
Why or why not?
LEARNING IS AN EVERY DAY ACTIVITY
Its incredible. For a lot of time Ive been looking for good and organized information just like it.
sure and should be this is super computing and i need to be a super somputer uer for my career and goals as a CEO and corporate president of MOTTO-95 CORPORATE AMERICA ORGANIZATION.
Well i have taken the course DEBUGGING SERIAL AND PARALLEL CODE earlier befor and i still reregisterd beacuse i opened my student web CT and found out that the course has been removed , How ever the course have been attached to my CURRICULM VITAE which is an upgrade. i am so glad with the University of Illinois and the National centre for super computing applications.
Convenience and useful
 
What did you like most about this course and why?
VERY INFORMING
The organization, examples and the use of various debuggers.
Detailed explanation about why one use a particular command of the debugger.
it is very explanative.
Its is very exceptional and advancing , it has made my CV a good one and my technolgy super comuted.
the course content
 
What did you like least about this course and why?
EXCELLENT IN QUALITY INFORMATION
Nothing really. All was very, very good.
Errors. There are a lot of errors in this course: Source codes are not consistent to each other for the same program. (That is, there are different versions of the same program that confuse the readers). The solution provided in section 3.11.4 is incorrect. There are data dependence in the solution so that the program produces wrong answer when more than 1 thread is used.
no certificates is what i disgust as a certifiate will upgrade and give value of time and academy.
The course i difficult and conprehensibly articulated ,but need a lot of attention no distractions.
audio/video resources are not enough
 
Additional Comments:
NCSA IS DOING VERY WELL.
The parallel debugging section was cool. Im particularly intersted in parallel programming so my next choice will be some course on it.
The course will be excellent if all the errors are removed.
i need a certificate to prove i have done this work.
the course have been attached and listed in My CV so the University of Illinois and NCSA if contacted should be able to defend me as i have taken the ocurse earlier last year and expected to stil have it on my web CT but foud that is is out and still reregisterd for the course.
 
Display Wall-in-a-Box
Total number of evaluations submitted for this course is: 6
This web-based course as a whole was:
ResponseCountPercent
Good 2 33.3%
Very Good 2 33.3%
Excellent 2 33.3%
The effectiveness of the web-based format was:
ResponseCountPercent
Fair 1 16.7%
Good 1 16.7%
Very Good 3 50%
Excellent 1 16.7%
The quality of the course content was
ResponseCountPercent
Good 2 33.3%
Very Good 2 33.3%
Excellent 2 33.3%
The organization of the course was
ResponseCountPercent
Good 1 16.7%
Very Good 1 16.7%
Excellent 4 66.7%
Usefulness of self-tests was
ResponseCountPercent
Not Applicable 6 100%
Usefulness of references provided was:
ResponseCountPercent
Good 2 33.3%
Very Good 1 16.7%
Excellent 3 50%
Usefulness of audio/video resources was:
ResponseCountPercent
Not Applicable 4 66.7%
Fair 1 16.7%
Excellent 1 16.7%
What was your knowledge of this topic before this course?
ResponseCountPercent
No knowledge at all 2 33.3%
Basic knowledge 4 66.7%
What is your knowledge of this topic now?
ResponseCountPercent
Basic knowledge 2 33.3%
Intermediate knowledge 4 66.7%
Do you think the knowledge you gained will improve the quality of your work?
ResponseCountPercent
No 2 33.3%
Yes 4 66.7%
Would you take another web-based course?
ResponseCountPercent
No response 1 16.7%
Yes 5 83.3%
Why or why not?
this was quick and easy to do from home.
I found the ability to learn online at home at a time and pace of my own choosing to increase my desire to learn. It doesn't require as much comittment to start a course and makes it more likely that one will try something new.
 
What did you like most about this course and why?
It explained the settup of the wall in logical path so I could understand why it was designed the way it was.
This course was logically laid out: it was easy to navigate.
 
What did you like least about this course and why?
I might like info on failings while trying to build the wall although some of this was covered.
Some of the links to other institutions pointed to unreachable pages and although they might be found by searching outside this course it breaks up the easy flow of the course. The point being that the course is there to provide all the relevant information so that it is easily at hand.
 
Additional Comments:
This information will save me time on a project.
I came across this course on my search for a program or at least more knowledge on blending crt projectors for a personal home theatre. Prior to this AVS's (www.avsforum.com) forum on crt projectors had the best practical information for doing this on the cheap. For me the commercial solutions are prohibitively expensive and it is heartening to see programming effort on this being released into the public domain. A day may come when I get to see my two Marquee 8500's blended.
 
Getting Started on the TeraGrid
Total number of evaluations submitted for this course is: 12
This web-based course as a whole was:
ResponseCountPercent
Good 2 16.7%
Very Good 9 75%
Excellent 1 8.3%
The effectiveness of the web-based format was:
ResponseCountPercent
Good 2 16.7%
Very Good 8 66.7%
Excellent 2 16.7%
The quality of the course content was
ResponseCountPercent
Good 4 33.3%
Very Good 7 58.3%
Excellent 1 8.3%
The organization of the course was
ResponseCountPercent
Good 4 33.3%
Very Good 6 50%
Excellent 2 16.7%
Usefulness of self-tests was
ResponseCountPercent
Not Applicable 2 16.7%
Poor 1 8.3%
Fair 6 50%
Good 1 8.3%
Excellent 2 16.7%
Usefulness of references provided was:
ResponseCountPercent
Not Applicable 1 8.3%
Good 4 33.3%
Very Good 6 50%
Excellent 1 8.3%
Usefulness of audio/video resources was:
ResponseCountPercent
Not Applicable 9 75%
Fair 1 8.3%
Excellent 1 8.3%
No response 1 8.3%
What was your knowledge of this topic before this course?
ResponseCountPercent
No knowledge at all 5 41.7%
Basic knowledge 6 50%
Intermediate knowledge 1 8.3%
What is your knowledge of this topic now?
ResponseCountPercent
Basic knowledge 5 41.7%
Intermediate knowledge 7 58.3%
Do you think the knowledge you gained will improve the quality of your work?
ResponseCountPercent
Yes 12 100%
Would you take another web-based course?
ResponseCountPercent
Yes 12 100%
Why or why not?
Very efficient way to learn.
I find the course format a good way to learn new techniques, plus the self-test at the end helps reinforce core concepts.
Great way to learn--at my pace!
These courses are by far the simplest, easiest tutorials I could find anywhere for the topics covered. Also, they are free.
To become more comfortable with different programs.
Am alone in the wilderness with no one to ask
 
What did you like most about this course and why?
Good intro to TerGrid I will need in the near future. All resources are efficiently placed in one spot.
Revisiting and references
I really liked the example inputs and expected outputs provided.
A good summary of the basics of working on the Teragrid in a tutorial manner, as opposed to accessing the Teragrid support pages directly, which I feel sometimes have an overwhelming amount of information.
concise and straightforward
I was able to understand the information provided in this course and will be able to retain it.
good overview of the topic, helpful
easy acccess
 
What did you like least about this course and why?
Little inapplicable at the beginning regarding logons.
the presentation order of material which I could have changed but did not assuming the course should be done in sequence. I had to revisit the later sections.
Some of discussion about specific RP environments I do not have allocations on were not applicable to my interests.
The expected level of proficiency is very uneven--some topics are a bit difficult to follow (for instance, the portion on modules), and some are unnecessary (for example, explaining step-by-step how to edit a batch script with pico, in my opinion, was not needed; it would have sufficed to indicate what changes to make to the template script with a text editor).
very small number of test questions
I think the couse could have been a little bit more clear for those of use that are just getting started in type of course.
self-tests need more questions
links are stale in few places.I do not have access to ABE
 
Additional Comments:
could you add a feature to allow a PDF of the course to be downloaded? Then I could use it later (it would have search capability).
I found this course very useful, thank you.
Overall, great course. I know that when I need more details I can go to the support pages, but this goes a long way in getting me started using the Teragrid resources!
 
Intermediate MPI
Total number of evaluations submitted for this course is: 11
This web-based course as a whole was:
ResponseCountPercent
Not Applicable 1 9.1%
Fair 1 9.1%
Good 3 27.3%
Very Good 4 36.4%
Excellent 2 18.2%
The effectiveness of the web-based format was:
ResponseCountPercent
Fair 2 18.2%
Good 2 18.2%
Very Good 5 45.5%
Excellent 2 18.2%
The quality of the course content was
ResponseCountPercent
Fair 1 9.1%
Good 3 27.3%
Very Good 4 36.4%
Excellent 3 27.3%
The organization of the course was
ResponseCountPercent
Fair 1 9.1%
Good 7 63.6%
Very Good 1 9.1%
Excellent 2 18.2%
Usefulness of self-tests was
ResponseCountPercent
Fair 1 9.1%
Good 6 54.5%
Very Good 4 36.4%
Usefulness of references provided was:
ResponseCountPercent
Not Applicable 1 9.1%
Fair 1 9.1%
Good 4 36.4%
Very Good 3 27.3%
Excellent 1 9.1%
No response 1 9.1%
Usefulness of audio/video resources was:
ResponseCountPercent
Not Applicable 7 63.6%
Poor 1 9.1%
Fair 2 18.2%
No response 1 9.1%
What was your knowledge of this topic before this course?
ResponseCountPercent
Basic knowledge 9 81.8%
Intermediate knowledge 2 18.2%
What is your knowledge of this topic now?
ResponseCountPercent
Basic knowledge 1 9.1%
Intermediate knowledge 9 81.8%
Advanced knowledge 1 9.1%
Do you think the knowledge you gained will improve the quality of your work?
ResponseCountPercent
No 1 9.1%
Yes 10 90.9%
Would you take another web-based course?
ResponseCountPercent
Yes 11 100%
Why or why not?
Convenient, well presented, good exercises
learn interesting things
because the way these courses go on are structured perfectly. It starts with the basic concepts and even for a people who don't know anything about the subject, it may create a big difference. It's an amazing tutorial for the beginners, but in a short time it improves the learner a lot. No, more than a lot!
I needed to use many statements with mpi, so I worked with the material called "introduction to mpi", and I thought too much interesting, but I think I'm not prepared to learn "Intermediate MPI". I will use this material, but in the future.
In this way one can study anytime any day.
I've been programming for 40 years, originally mainly in Fortran. I have little experience with MPI, PVM, etc.
It's a nice way to study at my own pace, and I can make the "book" [browser] look the way I prefer.
it is time efficient
 
What did you like most about this course and why?
Good concise notes on the topics. Solutions given to most of the exercises.
This course was interesting, in the sense that knowing basic MPI you can pick and choose what you need from each chapter without loosing continuity
organizationcode examples for every lesson
I did not read all the material.
The course provides easy to understand (albeit somewhat contrived) examples of how to use somewhat esoteric MPI calls.
I didn't take enough time to read every screen-- I used the self-tests as a study guide.
color coding the source examples is a great use of color--good job!
Examples presented in C
 
What did you like least about this course and why?
Not so many examples of using the different IO strategies in the MPI IO chapter
short self tests
self-tests are not enough and compared to Introduction to MPI, they were weaker.
It's necessary many examples to help us to understand the statements.
These comments are based on my impressions from the 1st three chapters, which are all I had the chance to complete by Oct. 18th. I'll try to get through more sections in the next few days. First of all, the course lacks a "mission statement". There needs to be a "Chapter 0" that lays out the reason for creating an intermediate level course and why the topics covered were chosen and structured the way they were. Give examples whose coding is simplified by making use of these calls. Second, the English (grammar and spelling) is not as polished as I would have expected for a draft for which you are seeking comments. (I'll send a hard copy to Leslie with suggested corrections.) Third, your discussions and/or examples you chose give the impression that while you may understand MPI programming, you do not understand the underlying theory behind the algorithms being employed. For example, in the Monte Carlo example of Chapter 3, you use a random number generator to generate the NUMBER of points to be sampled. Why would anyone do this? Certainly not based on anything from the theory of Monte Carlo integration. The relative accuracy of a typical Monte Carlo code (not employing variance reduction techniques) varies as the number of points sampled, N, to the -1/2 power. To get 3 digit accuracy you would want to sample about 10**6 points. Hence, there is no rational reason for making N random. If the convergence is so slow, why do Monte Carlo? In addition to odd volumes, another more common reason for using the method is simply the dimensionality of the integration to be performed. If the dimension of space is larger than about 4-6, then doing the integrations as iterated 1-D integrals becomes prohibitively expensive as the number of operations grows as M**d where M is the typical number of sample points in each dimension. Similarly, the discussion of "norms" in section 1.2 was confusing. For example, I was confused by the notion of the norm of a vector being a vector. This is not correct. The norm of a vector is scalar (as a norm is essentially the generalization of the concept of the length of a vector). The norm of a set of vectors might be a vector. (This might be applicable if you are generating multiple solutions in parallel.) Finally, a few opportunities for useful comments were missed. E.g., you never explained why the parallel implementation of the 1-norm needed to be coded as abs(a)+abs(b). Granted the reason is pretty obvious, but at least you should have left it as a problem to the reader. In section 2.4 on message probing, some comment on the necessity that messages be received in order is needed. Section 2.4.3 seems to implement a really weird method for handling messages of different datatypes. The more natural and flexible method would be to encode the datatype in a tag (say in its low order bits) and then extract the information via some sort of cheap mod operat! ion. I apologize if I sound too negative here. I think this can be a very good course with a little more work.
Fragmented codes. Completed working codes should be given. That way we don´t waste time trying to figure out why something does not work.
Clear examples, lots of materials.
C++ and mpi seems like a great idea from a computer-science/theory point of view. But for applications developers, and scientists, I'm afraid it will yield code that's quite difficult to understand [though very compact and pretty] six months after the last developer quits working on it.
-
 
Additional Comments:
some video examples for the code written for a better understanding
Thanks to the material.
In section 2.1, what about MPI_Irecv, etc? In 2.4.1, can something other than calloc be used? Some might get the impression that calloc is the only method that can be used.
One of the best MPI resources when used with the Intro MPI WebCT course. Lucy Carruthers U of Arizona
C++ is an abomination of C. Real programmers avoid it. [my 1 free flame]
-
 
Introduction to MPI
Total number of evaluations submitted for this course is: 184
This web-based course as a whole was:
ResponseCountPercent
Not Applicable 1 0.5%
Very Poor 1 0.5%
Poor 2 1.1%
Fair 3 1.6%
Good 30 16.3%
Very Good 68 37%
Excellent 79 42.9%
The effectiveness of the web-based format was:
ResponseCountPercent
Not Applicable 1 0.5%
Poor 2 1.1%
Fair 8 4.3%
Good 35 19%
Very Good 79 42.9%
Excellent 59 32.1%
The quality of the course content was
ResponseCountPercent
Not Applicable 1 0.5%
Very Poor 1 0.5%
Poor 1 0.5%
Fair 8 4.3%
Good 31 16.8%
Very Good 71 38.6%
Excellent 70 38%
No response 1 0.5%
The organization of the course was
ResponseCountPercent
Not Applicable 1 0.5%
Poor 2 1.1%
Fair 5 2.7%
Good 45 24.5%
Very Good 68 37%
Excellent 62 33.7%
No response 1 0.5%
Usefulness of self-tests was
ResponseCountPercent
Not Applicable 16 8.7%
Very Poor 1 0.5%
Poor 3 1.6%
Fair 20 10.9%
Good 48 26.1%
Very Good 48 26.1%
Excellent 45 24.5%
No response 3 1.6%
Usefulness of references provided was:
ResponseCountPercent
Not Applicable 24 13%
Very Poor 1 0.5%
Fair 15 8.2%
Good 53 28.8%
Very Good 57 31%
Excellent 24 13%
No response 10 5.4%
Usefulness of audio/video resources was:
ResponseCountPercent
Not Applicable 113 61.4%
Very Poor 2 1.1%
Poor 8 4.3%
Fair 12 6.5%
Good 15 8.2%
Very Good 11 6%
Excellent 5 2.7%
No response 18 9.8%
What was your knowledge of this topic before this course?
ResponseCountPercent
No knowledge at all 94 51.1%
Basic knowledge 82 44.6%
Intermediate knowledge 5 2.7%
Advanced knowledge 1 0.5%
No response 2 1.1%
What is your knowledge of this topic now?
ResponseCountPercent
No knowledge at all 4 2.2%
Basic knowledge 66 35.9%
Intermediate knowledge 103 56%
Advanced knowledge 9 4.9%
No response 2 1.1%
Do you think the knowledge you gained will improve the quality of your work?
ResponseCountPercent
No response 5 2.7%
No 2 1.1%
Yes 177 96.2%
Would you take another web-based course?
ResponseCountPercent
No response 6 3.3%
No 3 1.6%
Yes 175 95.1%
Why or why not?
It's an easy way of gaining knowledge and this course was very good. Compared to finding all this information my self would have been difficult and I would not be aware of some of the stumbling blocks common to this topic.
easier delivery mode
Basically because of the fact that I have all the information available all the time.
I think I got a lot out of this course. I began at almost zero knowledge and now I could work out some simple problems on my own. I will probably take the follow up during the next few days.
to find and learn new things.
Because I really got a lot of things from this web-based course. I can't wait to see the other courses online. I've already enrolled to OpenMP and I'm gonna start that right now.
Because I am successfully able to get information I need from the web course (or, alternative variant of answer, it is sufficiently good for learning what I want to learn).
If it is web based and maintained this neatly, I can take my time to understand the things and I can spend time whenever I want.
I hope other courses are as good as this.
Easy to understand the information
This is an extremely convenient way to learn new computing skills. One can take the course at exactly the right time to begin applying the skills in his/her daily work.
It provides a basis of knowledge to take off.
It was free, and available to me at times when I could spend time on it.
I want know more about this topic
self study, spare time
It was very insightful and was flexible in time needed to take it.
More advanced is needed.
They're an effective introduction to a topic.
Allows me to study the topic as I have time.
It went well. Perhaps I took longer that I should have to finish everything though.
lots of coding examples-> great!
Convenience (scheduling) and ease of review of the on-line material.
self paced
It provides the greatest support for self-based learning. This is the first place I would check for learning something new.
I will take intermediate MPI
You can do the course at your own speed, when you have the time. It is very useful to do it online, since every reference and suggested link can immidiately be checked.
The web-course was free, and well-written. It provided a more structured learning approach than a HOWTO or text-based introduction. It is convenient to do the course at one's own pace. The self-tests are very useful.
I think that it is very useful for the following reasons: 1) presents the course-taker with methodically organized information, 2) allows skipping some chapters that are less needed at the start, although these are not numerous, 3) self-tests and course problems are good for establishing a better knowledge base, and make the course more interesting.
Good introduction to topics.
I thought the information presented was to the point and effectivly established directness at presenting the material.
AS it gives me more information and can immediatly if any thing donot know check the links on the net
- own scheduling - quick copying of relevant paragraphs / pictuers (compared with reading a book)
Yes, it is quite comfortable and useful.
convenience
Though the course structure was desgined very well but there are only few examples. With out examples it is very difficult for any beginers to learn such advanced topic like MPI.
Easy access and easily scheduled. It is good for practice programming, because of the source code available.
it will help me
I benefit a lot from the course And I want to go forward, then I choice Intermediate MPI.
its interesting, useful, and very easy and clear explaining of ideas.
Because these courses are designed to give the computational scientist exactly what he needs, while also providing some references for more in-depth explorations
because I can study when i have time, and can go back if i need to review something
Well I am at the first class of the course, but I think it will be very usefull for me in the near future.
To get more information about parallel algorithms.
yes i will the tutorial was more useful since it contained source code too.
This has been a very efficient method of instruction. The wall-clock time spent on any major section (e.g., all of Section 4) was an appropriate amount (i.e., it didn't take 5 hours to complete). In other words, I could comfortably fit in various chapters for study throughout my day and complete them to my satisfaction.
This course provides me very helpful knowledge, I am longing for some advanced cources more.
Convenience
fast, convenient
Because in a week time I start a work in which parallel programming (MPI, OpenMP, HPF etc.) is involved. But even if I hadn't been given this work and if I simply discovered the courses, I would probably take some of them in order to learn something new.
Sure, because I can read the outline and print up to read just what I need.
easy to see what is interesting for me or not
I did gain knowledge and the course fit into my schedule. I was only able to work on this an hour or two a week.
This course was not boring and was very useful in my real life programming, I expect another one will be the same
I found this course extremely useful both as an initial learning tool, but also as a reference. Thank you very much.
I found it useful. Like talking about MPI step by step which i think is not present any where else. There are papers which discuss some things in details and there are books which would my whole life to read, but this one just tells what you need to know. Thats why to save my timne and learn more i would love to take some other coursese on this site.
its effective and easy to do while at work. I like how you can stop the course at any time and automatically return to that bookmarked location later.
I'm now taking the course ('Introduction to OpenMP and its very good too ...
It saves lots of my precious time.
The self-paced format is excellent.
Easy way to pick up information I need.
The web-based format is convenient & easily accessible. Your interface is very smoothly designed, with minimal obstacles.
Because 1. It allows me to search information faster; 2. The same piece of information can be observed faster than for the writtten version; 3. The study of web-based course is more active
Because I can access it anywhere and more effective.
Very efficient way to learn at user's pace.
to get more Information
Yes I would. I was seraching the web for information on MPI as I am doing a dissertation on Parallel computations and my programs are in Fortran. This has been theonly place where i got the info I needed to proceed with my programming. I am very pleased.
They have been quite usefull to reach my goals at the moment.
Such courses are great for further education and satisfying curiousity.
Self paced and was easy to follow.
I will take another course. its convenient, I can take from anywhere anytime. Instead of reading the whole book, the course is faster and to the point.
It's a time saver
its an efficient method for learning.
This is a good form of education.
Cause it rocks :)
It is so good. :-)
Ease of use
If you would be kind enough to provide C samples of the programs this course would be VERY beneficial. I can pace myself and can print out what I want to.
was well worth 30 minutes to pick up a bit of basic background on the topic.
I want to learn more about high performance computing.
I enjoyed the way in which the content was presented and especially enjoyed the feedback on the self tests. The content is also good.
The other courses available here seem interesting and useful.
Much effort was put into this (unlike other web resources) and this shows.
This is excellent. I had "no" clue about MPI but needed to use ScaLapack in order to speed up calculations. This way exactly what I needed, since it is difficult to understand how ScaLapack works without any knowledge of MPI.
because the information is concentrated and integral
need to learn certain topics
It is effective.
If I have to take another online course I would, but I prefer a lecture recording or some sort of human interaction. I feel this way because I am not familiar with C++, therefore I struggled with understanding this material.
 
What did you like most about this course and why?
I liked the appearance, because there was never too much stuff on one page.
simple language and good examples
Pretty simple and straight to the point topics.
I liked the way the game of life and search problems kept building up as the chapters progressed, incorporating the new techniques learnt.
examples for every new object described.
its structure - well organized
Organization and the way it goes on. The concepts are explained well and right after the explanations, the code samples are incredibly useful to understand. As a beginner to MPI, I got a lot of knowledge in very short time.
Good, plain-text explanations of the MPI framework at the start of the course.
The way it constructed is very good, starting from scratch and building the code is really a good idea.
Clear identification of MPI nomenclature, and understandable definitions of terms.It has enabled me to start to write a specification that will use MPI.
Well-thought selection of information
Easy to understand the information
Self paced.
The example problems built on the lessons in an easy-to-follow manner.
Examples and thorough explanation
progressivegives the most essential information one needs to starts practical work with mpi
The worked examples and plentiful examples.
I liked the web-based format.
Easy to understand.
completeness
good introduction to MPI, contains all the essential and have clear explanations. brief and clear.
Course Problems
Examples
It was concise and effective at conveying the material.
Clearly stated explanations. Easy to follow code samples. Excellant progression of information and layering of topics.
The simple example of searching an array was modified after each chapter to illustrate the chapter's new material. This was effective. Also, having the Fortran99 solution available to be peeked at made the C programming pretty straightforward. I also appreciated the way the material was broken up into small screen-size chunks. For most of the course, scroll bars were not needed.
lots of coding examples-> great! the references to C code.
easy navigation. self-tests. good, understandable and comprehensive introduction to the topics.
writing, explanation, gradual build up
I was able to move at my own pace and skip material I will not need.
It brought me up to speed and allowed me to implement a simple parallelized scheme for a serial Monte Carlo program I had developed.
I like the examples in Fortran and C respectively for each specific MPI routine. They helped me a lot.
That it worked its way forward a little at a time, starting out with letting you write the pseudo-code, then the serial code and then adding more and more of what you learn to the parallel code. This was a very useful way of learning.
It was free. It was well-written and aimed at not only introducing the student to the concepts of MPI programming, but also challenging them to write their own programs using the tests at the end of the chapters. It was aimed at both Fortran and C programmers.
The good methodical approach that allowed me (a total beginner in MPI) to grasp the basic concepts of MPI and parallel programming, and then to go a step further and learn some practical stuff.
Neat and well-organized.
about learning more on MPI programming.very sraight forward and easy to grasp.
Getting to the point, available.
I liked the examples of actual code. I liked the self tests. I like the step by step format (its much easier for me to learn like this than sitting down and reading a book).
its smiple and organized way of explaining
- it exactly fits my needs (themes and level)
It is always stright to the important concepts
chapter "projects". Very useful.
Theory was good.
Practical examples plus source code
it was a great eye opener
It describes the communication in detail which is what I like most.
sample codes and diagrams
simlesity
The simplicity and organization of the course.
See above: I attended SDSC's 5-day workshop on introduction to parallel computing and I really gained nothing out of it. This WebCT course gave me exactly what I needed to develop a solid understanding of MPI and apply it to my Computational Fluid Dynamics code.
organization and quality of content
it's very detailed, excellent for begginers
I didn't begin it yet...
Brief, but enough detailed curse focussing on the most important topics.
the contents were informative with references
The convenience of the web based teaching approach, the self tests and the various sample codes.
I like every parts of the cource, I think you have done a unvaluable thing.
The way the topics were laid out made it very easy to learn.
detailed self-tests
The content is easy to read and understand
Its existence;) - I didn't expect such a course can be available for everyone and for free - and that it can be prepared with such care.
See above.
I don´t have a good access to books on MPI. So, this course is very usefull for me. The topics are very well organized.
The examples were good.
Self tests and course problems help you not to forget material just after reading :)
Well organised, very clear explanations, good examples and useful tests.
The most important this is defintely, being comprehensive and informative. It doesnt take much of a time to parse throuhg and course and simple practical exmaples make it even more cool.
Concentrated on teh most useful, basic topics, explained well.
The pacing was just right. I did not feel like I was being spoon-fed little bits of information, nor did I get lost in the middle of advanced topics.
its very easy to understand the topics ... The text is very clear ...
I have no comment on this one.
The course problems were relevant to the material and to real programming situations. I also liked how they built on each other.
Size of the modules was good. It was relatively easy to stop and then pick up the thread of the class again later.
Good depth and breadth of topics, without requiring too much background knowledge. Well thought-out examples, excellent explanation for the reasoning behind various approaches.
repetition of the important things
The implementation's examples.
Organization went from basic to more detailed.This provided a good base to expand understanding.
its organization
Simplicity and relevance of the contents including the exercises and tests.
Detailed explanations, even a beginner can understand it. The acces to complete programmas are a good help to beginners implementation of new software.
Lots of realistic examples problems were provided.
Self-contained material and readability.
Very in depth chapters explaining the material
Explained all topics concisely, what is needed at the ('Introductory level. That way it is fast to learn the topic.
The organization was good.
Examples and clear explanations of relevant topics.
Quality of the material
It was easy to read and understand. I found the stuff, that I was looking for.
The material is very clear and well organised. The examples are very helpful.
Study at own pace
Good explanations
very clear exposition. Examples.
Very well structured and organized
New ideas were presented in a logcial manner.
A lot working examples, short description, very descriptive
The logical guidance through the material and then the self tests afterwards. I also think the practical at the end of each chapter is a good idea even though I didn't feel like spending time in doing them. I also enjoyed the short pages for each section and subsection instead of a 1000 line page for the whole chapter.
The problems gave an opportunity really learn the material.
Comprehensiveness.
availability
The "layer-by-layer" structure of increasing complexity and knowledge at the very right pace with useful self-tests in order to check that I had understood what I needed to proceed.
the organization and completeness
The exercises
exercise programs
Split into small slices, so it's easier to understand each slice.
I thought it was good that homework problems were assigned because it gives you the opportunity to practice. However, they were not helpful because I have no experience with programming.
 
What did you like least about this course and why?
I didn't like the exercises because they were too large/heavy so I didn't have the patience to read them trough carefully or to try them on my own computer.
The course could have provided more reasons why we need MPI
That is on HTML. It would be nice to have a version in PDF.
that the self tests did not store the results over time
The self-tests could be extended, I guess there are tricky and good questions but more questions could be added. When I came to the end, I always feel sorry to see the last question.
The cookie-cutter pattern in which most pages are written, starting at about 20% through the course.
The content in the slides (ppt and pdf) is very less compared to content in the web version so for future reference I will not have access to very good explanation in the web.
This questionnaire - Using this course as a starting point with MPI, it is very difficult to be objective about it's deficiencies.
Cannot complain at all
Easy to understand the information
I would have preferred to see all examples in C++ instead of C.
Writing was sometimes confusing. Some sections lacked prior exposure to some of the functions/variables that's critical to understand.
just one specific thing: the C++ chapter is not good C++ : looks more like C++ written by a C-programmer. Which makes it more difficult to understand the MPI library from a C++ programmer's perspective.
There are some bugs. Also, the ordering of topics was a little confusing at times.
More complex issues missing, only basic level.
Structure, mixing of different aspects related to mpi, organization could be improved
lack some global references (list of all MPI routines for example). the part about ScaLAPACK is not as clear as the rest.
No C solutions to Course Problems
Nothing.
C examples would have been more useful than Fortran examples in the situations where both weren't given.
Several examples are in Fortran only. They are however simple enough to follow for the course objective. Many MPI 2 features are not covered.
I can't name anything right now.
Didn't care for text wrapping in code(personal preferance maybe, however I felt it made it harder to quickly deteremine the head and tail of the function calls.
a general disadvantage of web based courses: missing live discussions. a personal problem: i was looking especially for a how-to for pmpi... maybe this is something you could add one day??? (or at least a reference?)
couldnt retrieve my old password, had to create a new account
The online MPI documentation could be organized better.
If the course can give me a total list of MPI routines, I will more appreicate this course.
Some of the later MPI-commands where not well documented. What I mean is that there where some commands which where only described by them being used in the example programs.
Please see the "additional comments" section for a list of errors I encountered, and my suggestions on how to improve the course.
Some of the MPI routines that interested me were not explained fully, but than again I guess this course is not the right place for a detailed reference.
Vedio or Audio should help.
nothing...
A few of the examples were a bit difficult.
I'm not familiar with Fortran, and there are many examples that were only shown in Fortran. They were more or less readable to me, but for my job at this point we don't use fortran so I didn't really want to bother with it. It may be nice to try to have both Fortran and C examples every time an example is given.
Few examples
...would prefere to select on top Fortran / C
Nothing I did not like.
poor indexing of commands
no comments
I dont have any dislike.
i cannot get access to other couses on your website
Becaue that's just a introduction course, so I can't ask more.
fortran
It was very fortran oriented. It would have been better if equal importance to C material was given.
I'd like more exercises
I just hope it's updated...
Well, I hpe I will begin the first class...
the content display
Coming into the course with "No knowledge" (question 8, above), I followed along quite well for the first few chapters. Some of the later material was a bit more than I could digest, so I will have to do some additional review.
Can't think of anything for this question.
too much fortran, not enough code inline or linked to a popup/viewer
sometimes the explanation is not adequate
That the text is often so much concentrated on signatures of MPI functions and there are unnecessary repetitions (i.e. two separate sections describing very similar functions Mpi_File_read_shared/write_shared; I believe it would be better to describe them together).
When looking at the pages in web browser there is not very much shown on a single screen and to get to another screen is a bit counter-intuitive
I would like to have this course on my natural language, portuguese.
I would like to have seen the course problems done in both Fortran and C.
Some errors in PDF format of the course
There was not very much that I didn't really like!
I dont tihnk i didnt like any thing. Yeah one thing, i wanted to know what Communicators does, NO well i know what it does, i need to know, how is it being impelemtned, like u can say, its implementation notes, but i dont find them any hwere. Coz i am thinking of implememntating MPI in some other language, and for that i need to know, how ppl have implemented communicator in a certain way and why ? Or-May-be it is nolt necessary to implement it in a certain way. Actually my knowledge is not quite good in this regard.
Too many Fortran examples. All of our team programs in C or C++.
I have no comment on this one.
The ScaLAPACK section annoyed me. I know its useful for many of the people who might take this course but it just didnt seem to fit in this course.
Tests could have suggested particular paragraphs to reread on missed questions
The self-tests seem to have been made as an afterthought, and werent all that helpful. This is, however, a minor complaint.
Sometimes, questions for self-test are a bit formal and scholastic
some examples are fortran only which is really annoying if you dont care about fortran
The derived datatype section. its unclear. I must get another resource to understand it.
its largeness
I just liked the course.
None.
Not enough C-code examples.. Need more....
Nothing in particular.
The exercises omitted key concepts that remain as questions for me. For example, the relationship between groups and communicators was not explored effectively. I also had to appeal to a book to figure out how to use a derived type with relative addressing (subtracting the base address). The example stops at the point of definition and before use.
ScaLAPACK chapter was overly specialized for the ostensible thrust "('Introduction to MPI". Also, Course exercises were somewhat repetitive (intentionally?)
The examples - cause the were good choosen. I liked the self-tests, too.
There wasnt anything I didnt like.
nothing
No ANSI C examples of the course problems. Only Fortran. This has basically made what would have been a great course quite useless.
Not enough graphs and pictures. Chapters 4 and 5 were wordy. (Note: I haven't finished the whole course yet.)
very descriptive, since sometimes written thing is very hard to understand but it is easer if you can present some corresping virtual schemes
The slow South African Internet connection. Maybe make the course downloadable in some format so that one must be connected to the internet and logged in, but the file are remotely available to the course-taker. -- just an idea :)
The examples were all in Fortran. This was also a good thing for me though, because referencing the solution still allowed the chance for me to finish my implementation.
Silly example of array search. Some examples are in FORTRAN only.
not enough details about specific commands (for example there is no explanation about message queues in MPI, the way MPI_Isend works and so on... ). No mentioning of MPI_Cancel -which is extremely useful.
the code samples, they were non-descriptive at most
too many things put inpages just describing usage of command not very useful
There should be document with the complete document in PDF format
 
Additional Comments:
Thank you for providing this course. I got enough information to start working in our MPI project.
good job
Excellent course. I am currently developing a project withMPI at university, and this course was very useful. Thanks!
Thanks a lot for preparing such a good tutorial.
NA
Some of the examples in C (eg: parallel I/O gave me segmentation fault). In some examples fortran 90 compiler didn't like "C" for comment. I changed that to "!". Also my compiler didn't like "MPIF.H". I changed that "mpif.h". Otherwise, all examples were fine. I don't know why some examples were segfaulting. I will look at them and see if I can fix it.
The best web course I have ever attended, TAs are very very helpful. If it is possible please prepare the Pdf documents of the course as it is.
Very well thought though, and a useful guide to come back too
There is a memory management bug in the supplied C example for the parallel SOR example (unneccesary and incorrect use of the replicate function). Also, for the mpi_module.c module in the same suite, the Open-MPI will fail, since it cannot handle a negative rank being passed in the MPI_Sendrecv command. This module must be rewritten with protective casing for the degenerate end-cases.
No
Excellant introduction into MPI in general. Clear concise explanations.
I need some kind of varification that I took the course.
Certain definitions, have little information on how they are defined, and/or how they do what they do... if nothing else an analogy section... such as this code, acts like a struct definition...
thanks!
great job to the MPI course developer.
I like this course very much. This course expresses any concept and idea in a such a way that I can easily understand and accept it.
I think it would be a good idea to also show the solutions to the course problems as C-programs and not just as Fortran programs.
section 5.12: meaning of "It should then calculate the average of the target value and its index" could be clarified with a formula; mostly because it's a strange sort of average to take. I had to read through the example code to understand exactly what the intention was. section 2.10: these pictures are not very clear. Something more like what's in section 6.3 would make more sense. section 7.6: spelling error in the very first bit of C code: 'int MPI_Cart_create' should be 'int MPI_Comm_create' section 7.7: the function of colour and key could be more clearly stated. Essentially colour and key are ways of creating new groups and ranks within those new groups, and assigning communicators to the appropriate processors contained in similar group, without having to use the MPI_GROUP commands. Processors with the same colour are grouped together, and the key is (almost) the processor's rank within that group. This is explained in the course, but the first time I read through it I didn't immediately understand it. Only after tinkering with the group and communicator commands while working on the chapter assignment did I realize exactly what they do. program code in section 7.8.2: in the integer declarations, variable ndim appears twice on lines 2 and 5 section 8.1.4: what happens if one of the cartesian dimensions is periodic, does the sub-group inherit the periodicity? This is indicated in the Notes at the end of the section, but not explicitly stated. section 8.1.6, fortran example: is the belongs() array correct for creating row subgrids? Compare it to the array given in section 8.1.4. The variable 'remain' called by MPI_CART_SUB should be 'belongs' section 9: this section is kind of weak. It would be nice to have more details on how etypes, filetypes and views all go together, with some examples of how this is actually used. From this section I get the feeling that parallel IO files are just like regular files interlaced together. What happens if there are huge empty junks in the files? Does space get wasted? I imagine parallel IO files have capabilities more interesting that just interlacing. But maybe they don't. The course problem is also pretty weak. It only works when each node returns only one data piece. A more elaborate example for multiple data pieces, and some examples of the abilities of parallel IO which make it worthy of being included in the MPI-2.0 specification, would be nice. Then again, perhaps something like this would be considerably more complicated. --------------------------- Thank you for making this course freely available. It seems as though a lot of effort has been put into creating it, and I learned a lot from it.
Should add more technical topics like on programming subjects,networking,Linux/Unix,and other computer science subjects
Thanks
It's very nice to have this available and I really appreciate it being offered free of charge.
That was the first WEB course I did. And I am very positive astonished about that course. C Examples not always correct (e.g. MPI_INTEGER did not exist ... should be MPI_INT)
Chapter about derived types coud be improved.
make it easier to get access to other courses
Thank you very much for high performance computing community! That's the best introduction coure I have ever met. Thank you all!
I would like to thank every one worked on this course and wish you all luck.
I expect to begin...
Errors/typos/suggestions for the course: ** 1.4.2 idle time (repeated words): “with with” ** Problem solution 1 spelling: “parelled” on line 1 ** Course problem 2 (Serial Search Program (Fortran 90)) solution: The two “do i=1,300” lines should be replaced by “do i=1,N” to take advantage of the “parameter (N=300)” statement, otherwise errors/incorrect solutions can occur for N values other than 300 ** Self test question 1: The Fortran should be {PRINT *, "Processor", myrank, "of", size, ": Hello World!"} instead of {PRINT *, "Processor", myrank, "of", rank, ": Hello World!"}, i.e., should use “size” instead of “rank” which is not defined in the example ** too many windows for self-test... have to keep shifting the scroll bar from left to right and back to see all the text. There needs to be a way that I can manually resize any display window. ** Course problem 3.17 solution typos: “...300 elements long, .” Need to delete the third- and second-to-last characters (the comma and space). ** 4.3.3 Notes, second point: "count" should be capitalized to "COUNT" ** 4.3.4.1 Definition line for request needs closing ")" ** 5.2 Needs the corresponding Fortran example ** 5.4 Needs the corresponding C example ** 5.5 Needs the corresponding C example ** 5.6 Needs the corresponding Fortran example ** Should 7.4 graphic not include the even part? ** 7.10 “processesing” spelling ** 9.1: need to define “MPP” as done in section 10.1 ** 9.6 Missing word: “Concepts [analogous?] to those used to define derived datatypes are used to define how data is partitioned in the file” ** Move 10.2 acronym descriptions to section 10.0 ** 10.3 Checklist item 4: repeated word “the the” ** 10.6 Missing word for sentence under Figure 10.6: “The syntax that is shown below is that which is used when we ____ to multiply the entire array A with the entire array b.”
It's a wonderful site to learning! I hope to take other courses.
Generally I appreciate your work alot. It really helped me to understand more about parallel programing. Thanks alot
You are doing a good job!
High quality, freely available courses like this are just wonderful! Thank you very much to everyone who made it possible.
Thanks alot for such a nice resource over the internet. Way to go and best of luck.
Thank you for posting this course. I believe I lost my first registration, and had to re-register. I think I am both read and read1. The e-mail address should be the same: read@boulder.nist.gov
I think that derived datatypes would have be a more detailed text ... But, I know that the this is in the Intermediate MPI course ...
I think this course is quite excellent, or maybe because this is my first visit, I dont know, anyway Thanks.
I took this course to gain information for a research project I am working on. It was tremendously helpful at filling in gaps and explaining the rationale behind information I found in a lot of other places. Thanks!
Very usefull overall. A good presentation.
An all-around excellent course. I look forward to the intermediate course. Many thanks, and kudos on a job well done!
This course is excellent, maybe you can provide more parallel computing courses on the web.
Keep up the good work!!!
Thanks for keep it free.
It would be nice to get some certificate of completion of this course.
Good Course. I will refer this to my friends.
Thanks a lot! This was a good experience.
GREAT JOB!
I cant believe something so good and useful is FREE!! Perhaps the world is a much better place than I thought. Thank you!
CAN YOU PLEEEASSSE PROVIDE C SAMPLES IN ADDITION TO FORTRAN? Is there ANY way to get this??? dgurgul@partners.org 617.724.3169
Would be nice to have a table which summarize all MPI subroutes by their's usage etc.
More graphs and pictures to help beginners to understand the course. Thank you.
I think that derived types is little bit earlier for beginner at least for me, I was very scared when I start derived types immediately after basic message commands,
I enjoyed this and feel it is a good supplement for the Parallel Computing model at my university. Thank you.
I really appreciate this resource.
Thanks for the course. Hope you'll improve its quality.
source code should include more in-code comments, also, variable names should be more self-explaining, it's easy to get lost in the code.
Thanks for the good work! I refreshed my MPI knowledge and learned new things.
I am missing some information about how to install/compile e.g. Open MPI. I am trying to run fortran code on windows and it is just not working.
 
Introduction to Multi-core Performance
Total number of evaluations submitted for this course is: 2
This web-based course as a whole was:
ResponseCountPercent
Poor 1 50%
Excellent 1 50%
The effectiveness of the web-based format was:
ResponseCountPercent
Fair 1 50%
Very Good 1 50%
The quality of the course content was
ResponseCountPercent
Fair 1 50%
Excellent 1 50%
The organization of the course was
ResponseCountPercent
Good 2 100%
Usefulness of self-tests was
ResponseCountPercent
Good 2 100%
Usefulness of references provided was:
ResponseCountPercent
Fair 2 100%
Usefulness of audio/video resources was:
ResponseCountPercent
Not Applicable 2 100%
What was your knowledge of this topic before this course?
ResponseCountPercent
Basic knowledge 2 100%
What is your knowledge of this topic now?
ResponseCountPercent
Basic knowledge 1 50%
Advanced knowledge 1 50%
Do you think the knowledge you gained will improve the quality of your work?
ResponseCountPercent
No 1 50%
Yes 1 50%
Would you take another web-based course?
ResponseCountPercent
Yes 2 100%
Why or why not?
It is very effective.
 
What did you like most about this course and why?
Running example code segments demonstrating each concept.
 
What did you like least about this course and why?
 
Additional Comments:
How do we come to know that the course content is updated, I follow the course content regularly. However, I was surprised to find that there is a new section added on the GPU multi-core performance. (MPI and GPU)
 
Introduction to Multi-core Performance - Version 1
Total number of evaluations submitted for this course is: 27
This web-based course as a whole was:
ResponseCountPercent
Fair 1 3.7%
Good 3 11.1%
Very Good 21 77.8%
Excellent 2 7.4%
The effectiveness of the web-based format was:
ResponseCountPercent
Good 6 22.2%
Very Good 12 44.4%
Excellent 9 33.3%
The quality of the course content was
ResponseCountPercent
Very Poor 1 3.7%
Good 5 18.5%
Very Good 14 51.9%
Excellent 7 25.9%
The organization of the course was
ResponseCountPercent
Good 6 22.2%
Very Good 11 40.7%
Excellent 10 37%
Usefulness of self-tests was
ResponseCountPercent
Not Applicable 1 3.7%
Very Poor 1 3.7%
Good 7 25.9%
Very Good 12 44.4%
Excellent 6 22.2%
Usefulness of references provided was:
ResponseCountPercent
Not Applicable 5 18.5%
Good 7 25.9%
Very Good 10 37%
Excellent 4 14.8%
No response 1 3.7%
Usefulness of audio/video resources was:
ResponseCountPercent
Not Applicable 21 77.8%
Fair 2 7.4%
Good 1 3.7%
Very Good 2 7.4%
No response 1 3.7%
What was your knowledge of this topic before this course?
ResponseCountPercent
No knowledge at all 7 25.9%
Basic knowledge 10 37%
Intermediate knowledge 6 22.2%
Advanced knowledge 4 14.8%
What is your knowledge of this topic now?
ResponseCountPercent
No knowledge at all 1 3.7%
Basic knowledge 7 25.9%
Intermediate knowledge 13 48.1%
Advanced knowledge 6 22.2%
Do you think the knowledge you gained will improve the quality of your work?
ResponseCountPercent
Yes 27 100%
Would you take another web-based course?
ResponseCountPercent
Yes 27 100%
Why or why not?
Direct confrontation with problems and auto evaluation is a good technique for memorization and understanding
It result intersting
I intend to take at least 3 other related courses from this organization.
You can take the curse anytime as per your convenience.
I'd like to take another course. Because I want to improve my knowledge on supercomputer and in turn improve the efficiency of my work.
i just don't know any better
is usec microseconds? More questions in self test would be useful.
I need to do parallel programing
convenience of being able to take course from home or work
Yes. I enjoy free web-based tutorials and courses. I semi-frequently participate in online labs via Microsoft's website, or other places. It a good way to test run new software or learn about something in a short time span.
There are new insights about multicore computation with new architeture of processors are growing
You have an introduction to MPI, which I don't currently understand and might need to.
I would like to learn distributed computing for CFD applications.
Learn more
because at this moment i'm involved in a issue of multicore server based project and our software isn't able to scale on that machine.
 
What did you like most about this course and why?
Course progressively going from L2 to OpenMP to OpenMPIGreat as an introduction.
The examples of benchmarks programs.
It assumed a reasonable level of prior knowledge and didn't try to over-simplify the topic. At the same time, it offered very clear explanations and wonderful examples (with a repeatable command sequence).
Simple and appropriate examples and explanation
The content is well organized and language is precise to the core. The example is well selected.
The high frequency use of the word "multi" was multi-impressive.
This web course went along with a 2-day workshop. I like that I can refer back to this course to review what was covered in the workshop. What I found most useful about the workshop was the introduction to a few profiling tools, and the one-on-one interaction available in a small setting such as this.
It's practical
Free, available, pertinent.
It's a relevant topic. I am looking at poor scaling with an application on our dual quad-core InfinBand cluster.
The main virtue in this format is that it is presented in condensed form. Some of the finer details might be missed for viewers who don't already have a basic understanding of multiple core hardware, cache usage, and threading. Providing that many who venture to enroll are at least familiar, this is a simple and quickly digestible material. I went through in less than half an hour, and performed well on the short self evaluations. Also, your simple diagrams are great- very easy to understand what is being described in text!
This course give me understood about some things that I was observe on my multicore machines like less preformance in aplications that share large amount of memory, in quad core and dual core processors , than old dual processors machines , now I know how mensure this problem
Fairly simple examples to explain its points.
Clearly explained topics, and provided useful examples without a requirement to run the code oneself.
Giving practical examples to explain the theory
the subject was explored deeply and the reports showing executions are very ilustrative
 
What did you like least about this course and why?
anything
If there are more explanation on basic knowledge of supercomputer, it would be better.
I never want to use the modifier "multi" again.
This isn't really applicable since this course wasn't stand alone.
The contents is little bit simple.
Not much meat to the course. It seemed that the course could be expanded to go into more detail. I was left feeling hungry for more.
Vocabulary familiarity is assumed, not defined. Each section would benefit from an initial list of terms and concepts before getting into explanations. Given the existing format, I would recommend this be the first page of each section. Specific Example: Section 3 Title is "MPI Performance on Multi-core Systems". Nowhere therein did I see MPI defined. The first line of the article talks about evaluating performance of MPI. Even if one can glean its meaning from context, or guess what is being evaluated by examples, why make a beginner go through the guesswork? You cut the potential educational value of this article in half by not defining all the terms.
there is less self questions , I think that more questions are necessary
It was very Linux-centric, and the stream examples didn't explain whether big numbers were good or bad.
I don't know enough of the topic to give a qualified answer, as an introduction I don't know what was missing or poorly explained.
nothing
 
Additional Comments:
are we done now?Bring on the lemmings!
Graphs have very thin bars ( slightly thicker would be better for presentation).
(I said something similar to the folks at TACC back in the spring) The next time NCSA hosts a small computing workshop like this one, I would suggest setting up some dedicated nodes beforehand, accessible from each participants login for the 2-3 day duration of the course. Having to rely on the queue to test things out hindered productivity. The machine is large, and with such a small group of participants, it shouldn't be impacted too greatly were everyone to be given one or two dedicated nodes. Anyway, that's just a suggestion. Nice workshop overall - thanks again!
Good job
After writing an article or course, read over it and then put every acronym, specialized term, and other important concept on the first page of each section. Explanations can be brief; keeping the course short is, after all, still of virtue in my mind. But including this page will contribute greatly to understanding the following material. The merit of reading such and article will then be more accessible, especially for beginners.
 
Introduction to OpenMP
Total number of evaluations submitted for this course is: 88
This web-based course as a whole was:
ResponseCountPercent
Fair 2 2.3%
Good 20 22.7%
Very Good 39 44.3%
Excellent 27 30.7%
The effectiveness of the web-based format was:
ResponseCountPercent
Fair 3 3.4%
Good 22 25%
Very Good 35 39.8%
Excellent 28 31.8%
The quality of the course content was
ResponseCountPercent
Fair 6 6.8%
Good 22 25%
Very Good 33 37.5%
Excellent 27 30.7%
The organization of the course was
ResponseCountPercent
Fair 3 3.4%
Good 17 19.3%
Very Good 38 43.2%
Excellent 30 34.1%
Usefulness of self-tests was
ResponseCountPercent
Not Applicable 1 1.1%
Poor 1 1.1%
Fair 8 9.1%
Good 24 27.3%
Very Good 29 33%
Excellent 25 28.4%
Usefulness of references provided was:
ResponseCountPercent
Not Applicable 33 37.5%
Poor 3 3.4%
Fair 6 6.8%
Good 21 23.9%
Very Good 13 14.8%
Excellent 9 10.2%
No response 3 3.4%
Usefulness of audio/video resources was:
ResponseCountPercent
Not Applicable 64 72.7%
Very Poor 1 1.1%
Poor 5 5.7%
Fair 3 3.4%
Good 6 6.8%
Very Good 3 3.4%
Excellent 1 1.1%
No response 5 5.7%
What was your knowledge of this topic before this course?
ResponseCountPercent
No knowledge at all 41 46.6%
Basic knowledge 38 43.2%
Intermediate knowledge 8 9.1%
Advanced knowledge 1 1.1%
What is your knowledge of this topic now?
ResponseCountPercent
Basic knowledge 39 44.3%
Intermediate knowledge 43 48.9%
Advanced knowledge 4 4.5%
No response 2 2.3%
Do you think the knowledge you gained will improve the quality of your work?
ResponseCountPercent
No 1 1.1%
Yes 87 98.9%
Would you take another web-based course?
ResponseCountPercent
No response 4 4.5%
No 1 1.1%
Yes 83 94.3%
Why or why not?
Very effective and efficient way to learn.
Good possibility to get basic knowledge of a topic
Informative and easy to use.
yes as soon as possible
Yes, because the topics discussed here are quite relevant nowadays and will continue to be so for years.
goog to learn no things
Well, it successfully gives me information I need, so if I need anything I can find on your site, of course I would use it.
There are other interesting courses and this was fine
Easy resource to find, can poke through topics at my own pace
Goood quality. The concepts of the topics are organized very well. Easy to grasp the idea of the topic.
The course information was presented well
efficient way to find basic necessary information.
The web-based course is convenient, free, and interfaces well with programming. As I read each page I try writing and running a little code.
Fast and concise introduction into the topic.
It was quite useful, easy and interesting.
To improve my knowlege in this and related topics. And it's the only form of courses available for me on this topic now.
It was a quick and effective means of adding to my knowledge of OpenMP.
It is precise and gives a good intro.
This course material will enable me to immediately begin parallelizing a code for my job. I intend to take the parallel debugging course next.
The course is very effective.
Excellent tutorials to get basic understanding of a subject
I am just interested in parallel programming
self paced.
This course does help me learn a bit more that I would like to known.
Good way of learning
Because this one was really good and I need other courses (MPI ...) !
They are a good way to have a quick start.
I found that this to be an effective way to introduce oneself to a previously unknown topic.
i was not yet compleated the course
it's convinient on my own time
This was a great source for basic information for future classes I am taking.
Because they are free, and I can take them at my own pace as time becomes available.
It is interactive compared to just reading the same topic from a book. But nowadays, some books are even complete with CD animation.
It is covenient.
Good to refresh knowledge
I would take another if there are topics of interest to me
It is effective.
I think that I would take any free course. However, I have not had exposure to programming, so this was difficult because i was not familiar with C or Fortran.
It was nice to be able to come and go as I had time to work on this. The information was also very well presented.
Because it is more accessible
Self-paced approach is a winner for me.
 
What did you like most about this course and why?
Broad and deep.
I liked the self tests most because I they providedthe possibility to apply/test the knowledge.
Very complete description with many examples. Despite being dated it covered many topics.
The course is a good introduction to Open MPThe concepts were explained very well with small examples. The material is short, clear with good quizzes.
freedom with quality
The fact that it was concise and the explanations were clear and simple.
code examples
Well, I actually didn't quite like the course (see next question for more comments). What I liked was perhaps that the course is quite concise.
The organization of the course is good
The examples. They help to understand.
*FREE!*; just a brute-force, basic introduction
It tells what is the purpose of certain command or clauses. This helps me a lot on when and how to use openMP in my applications.
It gives a great introduction to the topic and provided the key information without being too brief or overly wordy.
It was good to have many examples in several languages.
Self tests are really good.
The course included sample code so I could see just how and why I might use each topic.
Conciseness because of lack of time. Examples in both Fortran and C/C++.
It was quite useful, easy and interesting.
Clear language, selftests and usage of more than one programming language. Easy to follow. Well organized.
Plenty of well-explained examples
Examples in Fortran and C are available.
The topics were presented in an easy-to-read format with simple examples. This seemed much faster than reading a textbook would have been.
The subject is very well explained.
Very convenient format
It was written in a simple-to-understand way.
free access and self paced.
Reasonably comprehensive
A good introduction to MP. It taught me very quickly what MP is and isn't.
The numerous examples
Dificcult to say
The format proved a good step-by-step lead-in to the concept, gradually expanding on the details (syntax/implementation), and introducing the more advanced features.
the examples
Step by step illustration of concepts by examples.
Plenty of examples without intimidation. I also liked the C and Fortran code. Since I don't know fortran it wouldn't do much good to only have those examples.
It was short and concise. The subject matter is well suited to a short web based course such as this.
free and self pace
It has colored graphics. But there should be more graphics.
Simplicity
I was able to run it offline
Not too long format but covered the essentials.
The concepts and content were presented very clearly.
I think that Open MP was a very straightforward course. I liked that the commands and instructions were pretty simple and laid out clearly. I feel that the material was very well organized, although it was still difficult to fully understand
The code examples were very helpful and the content was consistent throughout.
I liked the selftest because it offers justification for wrong answers which facilitates learning.
Everything was explained clearly. working of various functions along with examples were really good.
Primarily an exposure as to what is possible with OpenMP but with enough samples/pseudocode to also show me how.
 
What did you like least about this course and why?
The more advanced concepts such as "More Directives" could have used a little more explanation.
The examples were clearly written in FORTRAN then converted to C. There were clear errors in a number of C examples.
I would have liked more information about the implicit information associated with each directive.
Probably more drilling excercises may be included.
not yet
The fact that some examples didn't seem illustrative for the discussed concept and that a few sections contained mistakes.
It feels more like a reference for openMP, not like a course on openMP. Perhaps this is due to the very "official", terse, dry language. It gives all the information, but the feeling I get is that "here's the information, do whatever you want with it", whereas the course would give you information AND at the same time show how to use it to achieve your goals. This "goal-achievement" thing (i.e., how to apply all that) is missing altogether.
too few examples
It could be longer and include more examples, for instance, about the use of flush.
No good examples - the course provides the basic framework for OpenMP, but gives no concrete examples good places to use it.
Should provide certain implementation dependent performance topic. This would help me to get on track more quickly.
The section on CRITICAL can be improved. Some of the code sample probably can be improved as well to illustrate the topic.
It may have been nice to have some type of pop-up definitions for some terms. There were a few times I had to go back and look up a term or two.
I did not fully understand private variables. I guess they must be declared before being listed. Should they also be initialized? Are those initial values passed to the threads? It would have helped me to see slightly longer sample codes, showing code above the private list, in the parallel region (as the samples do) and some use of the private variables after the private region.
Older parts of course and data (implementation of nested parallelism at 2001.), some typos, code examples were not highlighted in some places, I guess there was a error in one of selftests (omp_get_thread_num doesn't set a max number of threads, does it?).
There were formatting problems on some of the pages- i.e. text that did not wrap.
There were some typos in the slides.The answers on the first self-test I don't think were right or the information they were based on was wrong.
One of the page has sentences spread over 3 ft.
does not apply
The quantity of examples was rather poor.
No programming exercises.
Course content is not updated, especially nested parallelism.
n/a
Dificult to say
The provided code examples were very good at enforcing what was being described in the text. Perhaps the self tests could have also included some snipets of code for which the correct operation needed to be interpreted (multiple choice), or the odd simple coding problem (with a model answer that could be revealed).
It would be nicer if complete C/Fortran program examples are presented. Or rather, a section on "Put everything together" that shows the structure of a typical complete openmp program.
No coding exercises were given.
There are not enough graphics to help a beginner to grasp the concepts better.
The navigation method. Missing next/prev buttons.The glossary was not really helpful.
I did not care as much for this course because there were no recorded lectures available to watch. I would have like to have even a recording of a lecture on this material to watch.
No complaints.
It is very theoretical.A greater amount of practical exercises could further facilitate learning
few examples of working of recent real time projects would come in handy.
Towards the end of the course the C++ examples were not provided.
 
Additional Comments:
Thanks for providing such an excellent resource. The CI Tutor is really excellent.
keep it up ............
With a few qualitative changes and some useful examples, this course could turn into a great one.
NA
Need more real code to show the concepts
The content navigation menu does not work properly in Safari or Chrome. Had to find one of the hyperlinks which actually did work (think it was somewhere around chapter 9) and then go back to the beginning using the navigation buttons on the upper right.
Thanks
There are a few typos in the formulation of some tests, e.g., "omp_get_num_threads() returns what has been set with omp_get_num_threads()" -- should be "_set_".
I haven't finished the coding exercises. After that, I think I will be more sufficient to say I am good at openMP. And I think to take advantage of these knowledge, one needs solid coding work with it.
Thank you.
I would have had benefit from programming exercises.
Nice work! Keep it up! ;) I'm looking forward on more advanced courses and hands-on experience now.
Need to include a sample program at the end that combines many of the concepts in a single piece of code.
I wish you had more extensive self-tests.
Typo in Chapter 10 (More Functions) Self TestQuestion 3:returns a value set by call to omp_get_num_threads.Maybe it should be (see $ mark)returns a value set by call to omp_set_num_threads. $
hope the free access continues.
The self-tests are much easier than actually programming.. The best way to learn to program is to program, so the examples help...For the self-tests, could there be a free response? Entering the actual line of code "#pragma omp critical" (for example) would help me remember better than selecting something like that from a list.
Thank you for providing the course. I am on my way to the MPI course.
Thanks!
It would be a good idea to include examples and exercises (small programs to write) - then the student would better learn how to use OpenMP himself/herself.
Overall, it is a very concise and instructive tutorial.
Addition of programming exercises would be useful. The discussion of locks was not clear to me, especially as they pertain to the threads performing small parallel tasks. Maybe this topic could be discussed more thoroughly.
Animation will definitely help.
get an english major up in here to proofread your content. then get a test-writer to proofread and clarify the questions in the self-tests.
I would like to see assignments
Really liked the course content.Thank You!
Very pleased with the presentation. I've bookmarked the site for future online courses. Thanks so much for making this available.
 
Introduction to Visualization
Total number of evaluations submitted for this course is: 25
This web-based course as a whole was:
ResponseCountPercent
Very Poor 1 4%
Poor 1 4%
Fair 7 28%
Good 5 20%
Very Good 6 24%
Excellent 5 20%
The effectiveness of the web-based format was:
ResponseCountPercent
Very Poor 1 4%
Fair 5 20%
Good 7 28%
Very Good 6 24%
Excellent 6 24%
The quality of the course content was
ResponseCountPercent
Very Poor 2 8%
Fair 5 20%
Good 8 32%
Very Good 6 24%
Excellent 4 16%
The organization of the course was
ResponseCountPercent
Very Poor 1 4%
Fair 5 20%
Good 4 16%
Very Good 9 36%
Excellent 6 24%
Usefulness of self-tests was
ResponseCountPercent
Not Applicable 17 68%
Very Poor 1 4%
Poor 1 4%
Good 2 8%
Excellent 2 8%
No response 2 8%
Usefulness of references provided was:
ResponseCountPercent
Not Applicable 3 12%
Poor 1 4%
Fair 6 24%
Good 6 24%
Very Good 6 24%
Excellent 2 8%
No response 1 4%
Usefulness of audio/video resources was:
ResponseCountPercent
Not Applicable 12 48%
Very Poor 1 4%
Poor 1 4%
Fair 4 16%
Good 4 16%
Very Good 1 4%
Excellent 1 4%
No response 1 4%
What was your knowledge of this topic before this course?
ResponseCountPercent
No knowledge at all 3 12%
Basic knowledge 16 64%
Intermediate knowledge 4 16%
Advanced knowledge 2 8%
What is your knowledge of this topic now?
ResponseCountPercent
Basic knowledge 10 40%
Intermediate knowledge 13 52%
Advanced knowledge 2 8%
Do you think the knowledge you gained will improve the quality of your work?
ResponseCountPercent
No 10 40%
Yes 15 60%
Would you take another web-based course?
ResponseCountPercent
No response 1 4%
No 1 4%
Yes 23 92%
Why or why not?
Ease of use even during working a very tight schedule. Saves tons of resources (take cost of fuel for example) while effectively transferring knowledge.
I think it was very well structured and explainedIt's very easy to follow since it's web based and you can study at your own pace.
The other CI-Tutor courses are good although this course is really bad.
It seems useful
Web-based is not a deterent. As long as I know the content is acceptable and interesting to me, there is no reason to not take the course. I feel that the way this course is set up on teh web is fairly easy to use, after I figure out what those browsing icons are. Those icons are terrible because I can't actually see what's on them - just green on white and grey and black, or something. Are there arrows? They're not even in an order that makes any sense to me, or that I guessed in the first place. Some text when moused over would fix this problem (using firefox). Once I figured out the one to go forward, I just used it. Eventually I figured out backwards.
Very convenient way to learn new things without going through a lot of red tape
i find them a help in my research
Web based courses give me the flexibility to take them on my schedule.
Great opportunity to expand my knowledge of a very technical subject without going to a university course (and finding a place to park)!
 
What did you like most about this course and why?
I liked seeing the visualizations from different areas
3d graphics...I've always had a magnetic attraction to 3d graphics.
It's a basic level, but even so I got some knowledge I didn't have.
Nothing
Short and to the point, yet covers sufficient material.
I feel it was sufficiently comprehensive. I feel that this is a very useful reference tool. I also like that the sections are concise and short, but this leads to problems (discussed next).
Very informative, excellent explanations
It provided a good overview of the subject and vocabulary. This will make it easier for me to read more detailed articles about visualization.
The organization was excellent and it made navigating this topic easy, so I could focus on the information being provided
Summarized a lot of information on a subject I am very interested in. As a research consultant, I'm always looking for new ways to portray data, and I work wit researchers in a number of medical, scientific, and business fields.
Gave some ideas for visualization methods.
 
What did you like least about this course and why?
I was hoping for a discussion of advanced visualization techniques, such as various ways to render vectors and tensor fields in 4-D. What combinations of glyphs, coloring, etc are most effective?I was also looking for a discussion of current tools and their strengths and weaknesses. Only a few tools were mentioned and most were not discussed in any detail.
I regularly do CFD analysis and found the course to be very introductory. It would be difficult to write a course that covered every type of data set I'm sure.
It could have used some animated examples.
You could have added a section on hardware used typically.
The objective of this course. The course does not provide what the readers (who wanted to take this course) wanted.
Though I like how short the sections are, I frequently feel that certain things are not fully addressed - but I am no writer of courses or textbooks, so this may not matter. There are two main parts to this. 1. Special considerations and Research Issues and 2. their placement in the course. I believe that they should be placed within the relevent section, not at the end. 1. Considerations such as the fact that symbols used on a weather map may be great (considering their abstract nature versus being realistic), but the standard HCI or Social Computing researcher will ask "is there a way to do it better?" While I don't enjoy asking this question about everything, it was glaring when I read this section of the course (4.1.1. I believe). For one thing, abstract doesn't automatically mean harder/longer to learn. The course notes that it *can* be the case, not that it always is, but the fact that this particular example does take time to learn means that it could possibly improve to the point of little learning time required. This is a human factors type of question, but I feel like it could be mentioned in the course. This was not mentioned, even, in the research issues section at the end. 2. Some of the research issues (if not all) could/should be discussed in earlier sections (where they belong). This is a course, not a research paper. While text books will save such research issues for the end of chapters, they don't save them for the end of the entire book. This web-course could fit into one chapter of a larger book, but because this stands alone, I believe readers will treat it as a book in itself, so taht each section and/or subsection may be considered a chapter. Special considerations and research issues should be discussed WITH the topic so that they make immediate sense with the topic. You're not presenting and defending your research here.
Unfortunately, what I'm really interested in is not what you are calling "scientific" visualization. I am working with text data rather than numerical. However, on a general level I found this quite interesting.
It was just an overview. I had hoped that the web based nature of the course would allow more direct experience with some of the visualization tools that OSC users have access to.
I did not see the Print Course button, and proceeded to select and copy each module so I could print it later.
The course is outdated.
Outdated. Moderated wiki format would probably allow for updates from experts in the field.
 
Additional Comments:
I plan to put this knowledge to use immediately!Phil Elliott, GS-13Sr. Information Technology SpecialistSME for Applied Technology Supporting Genomics Research Office of the Surgeon General 2510 Kennedy Circle, Suite 208, Brooks City-Base, Texas 78235-5115 Phone 210-536-4058, DSN 240-4058, Fax, ext. 2625 "Integrate, extend and enhance. Repeat or die. The Quest for Relentless Innovation."
The whole course is completely off the point. Other courses in the CI-Tutor provide working knowledge of the relevant topics to the readers. However, this course just provide some basic information on scientific visualization. The readers (including myself) want to gain a working knowledge so that they can really able to produce, say, a 3-D animation. I have wasted my time to stduy this course.
thanks!
I will look for more courses like this!
 
Multilevel Parallel Programming
Total number of evaluations submitted for this course is: 14
This web-based course as a whole was:
ResponseCountPercent
Not Applicable 1 7.1%
Good 3 21.4%
Very Good 4 28.6%
Excellent 6 42.9%
The effectiveness of the web-based format was:
ResponseCountPercent
Fair 2 14.3%
Good 4 28.6%
Very Good 3 21.4%
Excellent 5 35.7%
The quality of the course content was
ResponseCountPercent
Good 4 28.6%
Very Good 5 35.7%
Excellent 5 35.7%
The organization of the course was
ResponseCountPercent
Good 4 28.6%
Very Good 4 28.6%
Excellent 6 42.9%
Usefulness of self-tests was
ResponseCountPercent
Fair 3 21.4%
Good 4 28.6%
Very Good 3 21.4%
Excellent 4 28.6%
Usefulness of references provided was:
ResponseCountPercent
Not Applicable 3 21.4%
Poor 1 7.1%
Fair 1 7.1%
Good 4 28.6%
Very Good 3 21.4%
Excellent 2 14.3%
Usefulness of audio/video resources was:
ResponseCountPercent
Not Applicable 8 57.1%
Fair 1 7.1%
Good 1 7.1%
Very Good 2 14.3%
Excellent 1 7.1%
No response 1 7.1%
What was your knowledge of this topic before this course?
ResponseCountPercent
No knowledge at all 2 14.3%
Basic knowledge 4 28.6%
Intermediate knowledge 5 35.7%
Advanced knowledge 3 21.4%
What is your knowledge of this topic now?
ResponseCountPercent
Basic knowledge 3 21.4%
Intermediate knowledge 5 35.7%
Advanced knowledge 6 42.9%
Do you think the knowledge you gained will improve the quality of your work?
ResponseCountPercent
Yes 14 100%
Would you take another web-based course?
ResponseCountPercent
No response 1 7.1%
Yes 13 92.9%
Why or why not?
Advantage of taking course anytime.
I found this medium encourages you to work through the material in a disciplined, structured way - whilst you can still allow yourself time to pause and think about the reasons for/implications of what is being stated.
I just think that you guys nailed it correctly with the format of the course
Th course was excllent. So I want to participate again in some new courses.
I think this is a greta way to learn things.Please put more courses online
 
What did you like most about this course and why?
bright and brief.
Pros/cons of using MPI vs OpenMP was wonderful. I would definitely like to have seen more in this section.
Examples and explaination
Thorough discussion of the pro's and con's of the three possible programming strategies (pure MPI; pure OpenMP; MLP) through consistently presented case studies combined with real-world performance comparisons.
Taking the Laplace problem all along the course was an excellent choice, it really exposes clearly where MLP can fit (actually I'm working in the thermal and wave equations so Laplace example fitted just perfectly)
I think the example based approach of explaining a concept is superb.
 
What did you like least about this course and why?
Please embed the sample code in the exercises. The CI Tutor layout makes it difficult to refer back to the code during the sample questions.
A minor point - but as someone with a preference for C, some code examples were provided in C as well as Fortran, but not all.
Nothing in particular
The laplace was slightly complicated.A simpler example would be easier to understand.
 
Additional Comments:
The MPI and OpenMP review sections were reasonably complete in their own right - (re-)introducing some aspects of MPI functionality I have not had the opportunity to use, and in a couple of cases expanding on what I have recently learned in the "Introduction to OpenMP" course.
I think it would be handy to expose also the multi-core vs SMP, that adds another granularity problem to the MLP code. In this scenario, my first approach would be to run as many MPI process as Physical CPUs are in a node, then as many openMP threads as cores are in each CPU. The whole point is to maximize the access to same cache. In OpenMPI it is quite easy to specify the number of process per node, but what I don't know yet is how to be totally sure that each process in the node is going to start in a separated processor and how tell openMP to effectively spawn its threads in the cores of the same processor. Some expose on this would be more than welcomeThank you again to made these courses freely available.
 
Pablo I/O Characterization Optimization Tools
Total number of evaluations submitted for this course is: 1
This web-based course as a whole was:
ResponseCountPercent
Excellent 1 100%
The effectiveness of the web-based format was:
ResponseCountPercent
Excellent 1 100%
The quality of the course content was
ResponseCountPercent
Excellent 1 100%
The organization of the course was
ResponseCountPercent
Excellent 1 100%
Usefulness of self-tests was
ResponseCountPercent
Excellent 1 100%
Usefulness of references provided was:
ResponseCountPercent
Excellent 1 100%
Usefulness of audio/video resources was:
ResponseCountPercent
Excellent 1 100%
What was your knowledge of this topic before this course?
ResponseCountPercent
Intermediate knowledge 1 100%
What is your knowledge of this topic now?
ResponseCountPercent
Advanced knowledge 1 100%
Do you think the knowledge you gained will improve the quality of your work?
ResponseCountPercent
Yes 1 100%
Would you take another web-based course?
ResponseCountPercent
Yes 1 100%
Why or why not?
 
What did you like most about this course and why?
 
What did you like least about this course and why?
 
Additional Comments:
 
Parallel Computing Explained
Total number of evaluations submitted for this course is: 80
This web-based course as a whole was:
ResponseCountPercent
Fair 2 2.5%
Good 21 26.3%
Very Good 33 41.3%
Excellent 24 30%
The effectiveness of the web-based format was:
ResponseCountPercent
Fair 5 6.3%
Good 20 25%
Very Good 31 38.8%
Excellent 24 30%
The quality of the course content was
ResponseCountPercent
Fair 6 7.5%
Good 26 32.5%
Very Good 29 36.3%
Excellent 19 23.8%
The organization of the course was
ResponseCountPercent
Poor 1 1.3%
Fair 2 2.5%
Good 17 21.3%
Very Good 37 46.3%
Excellent 22 27.5%
No response 1 1.3%
Usefulness of self-tests was
ResponseCountPercent
Not Applicable 7 8.8%
Very Poor 4 5%
Poor 8 10%
Fair 23 28.8%
Good 20 25%
Very Good 6 7.5%
Excellent 10 12.5%
No response 2 2.5%
Usefulness of references provided was:
ResponseCountPercent
Not Applicable 17 21.3%
Fair 3 3.8%
Good 26 32.5%
Very Good 16 20%
Excellent 13 16.3%
No response 5 6.3%
Usefulness of audio/video resources was:
ResponseCountPercent
Not Applicable 62 77.5%
Very Poor 2 2.5%
Poor 2 2.5%
Fair 3 3.8%
Good 3 3.8%
Very Good 2 2.5%
Excellent 1 1.3%
No response 5 6.3%
What was your knowledge of this topic before this course?
ResponseCountPercent
No knowledge at all 18 22.5%
Basic knowledge 41 51.3%
Intermediate knowledge 13 16.3%
Advanced knowledge 7 8.8%
No response 1 1.3%
What is your knowledge of this topic now?
ResponseCountPercent
Basic knowledge 33 41.3%
Intermediate knowledge 34 42.5%
Advanced knowledge 11 13.8%
No response 2 2.5%
Do you think the knowledge you gained will improve the quality of your work?
ResponseCountPercent
No response 2 2.5%
No 3 3.8%
Yes 75 93.8%
Would you take another web-based course?
ResponseCountPercent
Yes 80 100%
Why or why not?
Excellent and efficient way to learn.
because I am able to get the information I need successfully
I like the way the course goes through various concepts with a bird's eye view. I would like to take the other MPI based courses for this.
convenient, free
convenient location, scheduling and pace
Self-paced, I can leave and return easily
I have to lean more than I know for accuplish current work.
interest on the subject
Because I liked !
1)self paced learing.2)free access to course.
Good refresher for something I haven't worked with in a while.
Interesting informations and show some topics that I woudlnt ahve given the correct attention
Cost effective, possibility of communicating with others interested in the area, and self testing with active feedback is attractive.
I found the information concise and helpful.
Ease of use and the ability to access coursework from work, home, etc.
convenience
convenience
Good way to learn information.
This is very convenient and gives me a glossary to be able to start to understand the more advanced references. Thank you for your efforts. satwood@hsph.harvard.edu
The subject matter is relavent to my profession and interests me.
Because it is a readilly available, high quality, and tutorial type source of information, on subjects regarding High Performance Computing.
Is convinient to do things n web based course
I still need to know more about Parallel Computing. This course although it was good did not cover the topic in depth. I feel more examples were needed and a more indepth study would have given me the feeling of having learned more.
Very interesting topic. Would like to have hands-on experience in HPC, mostly on Linux Clusters.
The MPI tutorial was good.
I would take more web based courses in the series to help understand further how best to utilise the computing resources I help manage here at Manchester Metropolitan University
Some videos can be incorporated.Some one or two practical mathematical problems can be explained how they can be parallelised.
I like the tutorial provided
It allows greater flexibility with my schedule.
I need more knowledge for my work.
Very useful for work.
effective and fast learning in YOUR OWN PACE
This course seems like an excellent introduction to a broad base of information.
Presents good access to needd subject matter.
Ramping up on basics of parallel processing
I enjoyed the experience of this course.
 
What did you like most about this course and why?
Broad overview combined with explicit examples.
Plain-text explanations at the beginning of the course
The organization of the material
its brevity
Code examples helped me estimate what I can do.References to particular computers helped me judge relevance to the allocation I have.
When general categories of parallel computing were discussed and explained. An example would be the discussion of cache and general methodologies of optimizing cache use.
Well orgnized materials clearly explained.
I can lean overall view.
introduction
I like the language used that is wonderful to those that don't understand English very well.
Free access to course. Thank you very much and hope that you can keep it free.
simplicity and easy to understand
I like the brevity...I could get through it in a couple of hours of free time
The tutorial contains a good deal of collected wisdom that will help novice users, e.g., a) rules of thumb, such as loops needing about 1000 fp-ops to benefit from parallelization b) examples of why code might perform poorly, such as false sharing
The examples
Jump started getting 'my head into the application space', I am new to parallel computing. Describing the trends over the last few decades with a bit of light theory I believe provides a good introduction to parallel computing.
The terse nature of the course notes gave relevant information and concrete examples in a relatively short amount of space.
Gave a good introduction to terms used in parallel computing.
Good focus on defining terminology and concepts.
freely available
convenience! being able to learn when, where, and how i am able to is fantastic.
Great basic information- just what I needed. I'm not a computer expert, so the course started with the basics.
simple to understand
I started with basic understanding and a couple hours later I had a greater understanding and was not bored by the presentation.
The clear and accurate language used in the text. Because, this way, apparently complex concepts can be well understood, and there is a bigger motivation in, eventually, studying them in more detail.
The way its formatted.
The course is very easy to understand. The split up in a large number of short chapters is very good and in each chapter the small paragraph size makes it easy to read. The language is also very good and easy to understand.
Clear and precise explications. To the point.
Certain parts were stuctured rather well.
Cache relatedDebugging details
simplicity of the tutorial......
index layout to provide current where-am-i and out of sequence progression
It gave me more insight into the process of parallel computing.
Ability to study with a flexible schedule
- Lots of insights, which are based on experience.- Technical details are held at a minimum, which is good for an introduction.
good explanations, focus on relevant information
The depth of the material is well matched to the appropriateness of the examples.
Good overview
info content was relevant but may be dated for 2012
concepts were explained well
being simple and understandable. also the shortness of the course is a great feature, it does not make the reader tired.
Free and easy access
The degree of knowledge.
 
What did you like least about this course and why?
Some sections could use self-tests (I think there was only one) and I would have liked some more worked examples which I found very useful (e.g. loop blocking)
All these technical details about what is the size of L1 cache of different processors; why would you include such detailed listings in the course for total beginners? Also the material seems to be somewhat out of date, Intel Itanium is 5 years old now.
focus on a few specific systems
NA
self-test not very useful, didn't seem polished, some information may be outdated, not enough c/c++ examples
I was not entirely sure of the memory-sharing model at the NCSA Cobalt.
Specifics were often used when general ideas was what I was after. I feel that going into how to program for a specific computer architecture is not necessarily the best use of teaching space.
too many details on specific tools for an introductory course
This course uses on the most part the Fortran language. I would like to see more examples c/c++ codes.
Nothing I didn't like.
The quiz includes interconnection terms that are not defined in the tutorial material. It also seems odd that interconnection is the sole topic of the only quiz. Why not a quiz on cache terms or on one of the other parts of the tutorial? Thus, this tutorial seems incompletely developed.
-
The self-test included several things that weren't covered by the course content. It was useful, but further explanation of some of the things on that self test would have been helpful.
Perhaps another test or two would enhance the introduction, albeit it may make the introduction 'too heavy'?
I would have preferred fewer references to specific CPU and system architectures, since I personally don't have access to any of the example systems used. Also, more effective quizes (and more of them) would be appreciated. The interconnection networks quiz is the only one in the course, and it deals with networks that aren't even described previously.
The self test. It did not cover many parts of the text and did cover things not in the text (like the mesh and torus networks). Answer "a" seems to indicate that a bus based network must be based on a coaxial cable.
No reference to latest Intel I7 processor.
Not enough detail, although, being an overview course, it's probably not a fair criticism. The quiz focused only on the interconnects - some of which weren't mentioned.
no discussion of how to design parallel software Details would not be expected, but a section of the differences between linear and parallel programming, with simple examples, would have helped
some of them do not work on linux cluster
i had different expectations for the course material based on the title of the course. there were very few (one) self tests/comprehension questions/problems.
History. When I am working I do not think about "IBM,HP,Sun,etc, or Billy,Steve,Linus etc", I'm more concerned with the task at hand.
The lack of focus on distributed memory (Linux Cluster) systems. Because that is the kind of systems I currently use, and are the ones potentially available to a large number of students and investigators.
The Quiz
More material should have been provided. Possibly some demo software that simulates parallel programming environment on a PC so I can actually do some programming and gain first hand experiance.
Would add pictures of the systems involved. Useless, but interesting :)
The tutorial as a whole was too much disconnected. The parts did not seem to follow well. I had to choose which section I wanted to do instead of being led through.
Some of the references to particular computers were quite old and are now (March 2009) out of date, so some updating would be good.
lack of examples and its explanation
Some of the references are out of print.
The self-test takes too much time. Some sketches could be more useful at this point.
diagrams and visualizations could be a little more intuitive
Some content is dated, although still mostly useful for illustrative purpose.
constantly clicking to go to next page
the menu was annoying to use as i have to keep scrolling down to get to the next section
some examples were out of date. for example the parallel computers mentioned in the text were very old. also less attention was paid to linux clusters which I think are more important as all home users use linux for clustering.
some links are broken...processor info is dated
Updating of the hardware cited.
 
Additional Comments:
Thanks for hosting this excellent resource!
N/A
None
Some data is out of date, such as the use of Pentium III computers for the standard 32-bit PC.
Sorry about my English... =P
hope that you can keep it free.
1) The crossbar diagram is incorrect; the memories should be located on the vertical lines rather than on the same horizontal lines as the processors.2) The tutorial could stand some other minor cleanups: a) inconsistent presentation style regarding command lines with "unixprompt>" used in one example and "Prompt%" later (the use of ">" as a prompt character could be confusing since the same example uses I/O redirection via ">"); a style sheet for authors and revisions by a style editor might be helpful. b) spelling errors in section titles, e.g., "Scaler" and "0ptimization" (i.e., a leading zero rather than a capital O)
Very satisfied with the course, plan to take the introduction to MPI.
Overall very informative and effective.
Good overview without too much unnecessary history.
need more technical details about linux clusters.
thank you for making these and making them accessible.
Self-taught and easy to access- great. Can learn whenever I want to.
You ask questions in the self-test about material not in the course.
Thank you to the whole group involved in these online training!
Overall this is a good introduction to the topic of parallel computing. As a technician who helps maintain some small parallel/supercomputing resources at our university this has helped understand what we have and how to best use it effectively.
Some more assignments can make the course more interesting,
The definition-look up upon hovering the mouse cursor over the definitions is very useful. Maybe it is possible to export the definitions and sketches of a course for use in flashcards?That would really be great!
nice work! tahnks
I especially liked the references to published materials, so that a person could read further in any area of particular interest.
could use an update
Good job.
 
Performance Tuning for Clusters
Total number of evaluations submitted for this course is: 5
This web-based course as a whole was:
ResponseCountPercent
Good 1 20%
Very Good 2 40%
Excellent 2 40%
The effectiveness of the web-based format was:
ResponseCountPercent
Good 1 20%
Very Good 2 40%
Excellent 2 40%
The quality of the course content was
ResponseCountPercent
Good 1 20%
Very Good 1 20%
Excellent 3 60%
The organization of the course was
ResponseCountPercent
Very Good 1 20%
Excellent 4 80%
Usefulness of self-tests was
ResponseCountPercent
Not Applicable 1 20%
Good 3 60%
Very Good 1 20%
Usefulness of references provided was:
ResponseCountPercent
Not Applicable 1 20%
Fair 1 20%
Good 2 40%
Very Good 1 20%
Usefulness of audio/video resources was:
ResponseCountPercent
Not Applicable 4 80%
Good 1 20%
What was your knowledge of this topic before this course?
ResponseCountPercent
No knowledge at all 2 40%
Basic knowledge 2 40%
Intermediate knowledge 1 20%
What is your knowledge of this topic now?
ResponseCountPercent
Intermediate knowledge 5 100%
Do you think the knowledge you gained will improve the quality of your work?
ResponseCountPercent
Yes 5 100%
Would you take another web-based course?
ResponseCountPercent
Yes 5 100%
Why or why not?
Well Structures and self tests are very useful
Yes, the format makes it easy to follow. Most of us need guidance. These web-based courses give it. The also allow study at any time.
The courses provided here are very easy to learn
 
What did you like most about this course and why?
presentation, flow
The content of the course was detailed and has made me look at my programming in more detail. Certainly with respect to cache ram. The course content was spot on.
The course is well focused on the topic and short
Practical.
 
What did you like least about this course and why?
june 2003, It would be great if these courses are updated at regular intervals
Nothing
It's already a bit out-dated. There was no info on Infiniband for example.
none.
 
Additional Comments:
I have to say that it is wonderful that these courses are FREE. For courses of this quality you would normally expect to pay big bucks. for an independent like myself this is a bonus
Good job!
 
Tuning Applications for High Performance Networks
Total number of evaluations submitted for this course is: 6
This web-based course as a whole was:
ResponseCountPercent
Fair 1 16.7%
Good 1 16.7%
Very Good 2 33.3%
Excellent 2 33.3%
The effectiveness of the web-based format was:
ResponseCountPercent
Very Good 4 66.7%
Excellent 2 33.3%
The quality of the course content was
ResponseCountPercent
Fair 1 16.7%
Good 2 33.3%
Very Good 1 16.7%
Excellent 2 33.3%
The organization of the course was
ResponseCountPercent
Good 2 33.3%
Very Good 2 33.3%
Excellent 2 33.3%
Usefulness of self-tests was
ResponseCountPercent
Not Applicable 2 33.3%
Good 3 50%
Very Good 1 16.7%
Usefulness of references provided was:
ResponseCountPercent
Not Applicable 1 16.7%
Good 2 33.3%
Very Good 1 16.7%
Excellent 2 33.3%
Usefulness of audio/video resources was:
ResponseCountPercent
Not Applicable 4 66.7%
Good 1 16.7%
Excellent 1 16.7%
What was your knowledge of this topic before this course?
ResponseCountPercent
Basic knowledge 3 50%
Intermediate knowledge 2 33.3%
Advanced knowledge 1 16.7%
What is your knowledge of this topic now?
ResponseCountPercent
Basic knowledge 1 16.7%
Intermediate knowledge 4 66.7%
Advanced knowledge 1 16.7%
Do you think the knowledge you gained will improve the quality of your work?
ResponseCountPercent
Yes 6 100%
Would you take another web-based course?
ResponseCountPercent
No response 1 16.7%
Yes 5 83.3%
Why or why not?
effective way
it was very well written and organized
nice format, take at my own pace. Which is slower than most.
They can increase my knowledge on many topics.
 
What did you like most about this course and why?
small, to the point
easy access
that's it was written precisely. it was explained very well
Some tunning tips, testing aplications and links are very useful.
 
What did you like least about this course and why?
not enough low-levvel details (like OS related issues)
could have more info
None.
 
Additional Comments:
This course might be more comprehensive.
 
Using the Lustre File System
Total number of evaluations submitted for this course is: 6
This web-based course as a whole was:
ResponseCountPercent
Fair 1 16.7%
Very Good 3 50%
Excellent 2 33.3%
The effectiveness of the web-based format was:
ResponseCountPercent
Good 2 33.3%
Very Good 1 16.7%
Excellent 3 50%
The quality of the course content was
ResponseCountPercent
Good 1 16.7%
Very Good 3 50%
Excellent 2 33.3%
The organization of the course was
ResponseCountPercent
Fair 1 16.7%
Very Good 1 16.7%
Excellent 4 66.7%
Usefulness of self-tests was
ResponseCountPercent
Fair 1 16.7%
Good 3 50%
Very Good 1 16.7%
Excellent 1 16.7%
Usefulness of references provided was:
ResponseCountPercent
Not Applicable 2 33.3%
Fair 1 16.7%
Good 1 16.7%
Very Good 1 16.7%
Excellent 1 16.7%
Usefulness of audio/video resources was:
ResponseCountPercent
Not Applicable 5 83.3%
Good 1 16.7%
What was your knowledge of this topic before this course?
ResponseCountPercent
No knowledge at all 1 16.7%
Basic knowledge 3 50%
Intermediate knowledge 1 16.7%
Advanced knowledge 1 16.7%
What is your knowledge of this topic now?
ResponseCountPercent
Basic knowledge 1 16.7%
Intermediate knowledge 3 50%
Advanced knowledge 2 33.3%
Do you think the knowledge you gained will improve the quality of your work?
ResponseCountPercent
Yes 6 100%
Would you take another web-based course?
ResponseCountPercent
Yes 6 100%
Why or why not?
test feedback 1
Fast method of learning information.
I find these courses pretty helpful.
I find this form of self-paced education the best for both my schedule and my learning process. Once I understand something I am happy to move through it quickly, while I can back up if I have not gotten a concept completely.
Its faster ans easier to upgrade oneself.
 
What did you like most about this course and why?
test feedback 2
Short. Relevant.
-
Clarity of explanations.
Availability, comprehensive content, easily digestible size.
practical examples and diagrams
 
What did you like least about this course and why?
test feedback 3
Need to log in. It would be helpful to have the material in a format that is easy to reference so that I do not need to log in every time I need to look something up.
-
Nothing stands out.
coverage of topics is limited.
 
Additional Comments:
test feedback 4
Would be helpful if other centers could also use the material and modify it so that it was applicable to their lustre file systems -- not all of them are setup the same.
-
Apparently the content has not been updated since 12/2011. It would be great if it could be updated to the latest version of Lustre FS, 2.5.
it would be beneficial to include topics like lustre recovery, quota, client IO, Distributed Namespace, HSM, Lustre backend etc