Monday September 08, 2008 Vijay Tatkar's Blog
All
|
Benchmarks
|
Business
|
Cloud Computing
|
General
|
Hardware
|
Linux/Unix
|
Performance
|
Software
|
Solaris
|
Sun
|
Sun Studio
Monday September 08, 2008
Interesting Read: Parallel Programming Made Easy?
Michael Wolfe, Senior Compiler engineer/architects at
The Portland Group (PGI Compilers)
recently wrote an article featured in
HPCWire
questioning how Easy? one could claim to make Parallel programming.
Some interesting thoughts and views from one of the leading compiler architects from a compiler that has focussed on high performance computing for decades(?) now.
Whether you agree or not is your own personal view of how this important style of programming should be shaped, but he does represent a very important part of the discussion underway today.
Definitely worth a read!
Posted by tatkar
( Sep 08 2008, 09:11:54 AM PDT )
Permalink
Comments [2]
digg it
slashdot it
Technorati it
Trackback URL: http://blogs.sun.com/tatkar/entry/interesting_read_parallel_programming_made
Post a Comment:
![[Vijay Tatkar]](http://blogs.sun.com/roller/resources/tatkar/vijay_crop.jpg)
Good article, but it misses by a mile my most compelling argument for parallel programming: it makes programming systems simpler. For example, if I am coding a control system that has 8 devices to control, I have two obvious choices: 8 threads running exactly the same code, or one thread multiplexing 8 sets of state and data. The 8 thread solution is far simpler, and I let the OS do the multiplexing. If I use message passing, I don't even have synchronization code or errors. Very easy to debug, and very easy to scale.
Posted by Michael Lyle on September 09, 2008 at 11:18 AM PDT #
Agreed. You are perhaps one of the few that haslearned parallel programming issues.
For most others, its a daunting task. Thinking parallel, rather than sequential is a bit unnatural (if you define natural in terms of high-school math: sequential steps, step 7 follows from step 4 applied to step 2, above, etc kind of learning).
Also, there are no easy tools to help verify that parallel apps are in fact correct, at least at the same level/extent as debuggers today assist in showing you that sequential apps are "correct".
As I said, its a popular view but not necessarily one shared by everyone.
Thanx for your comment, it helps me understand some of the diversity of thinking around this issue
Posted by Vijay Tatkar on September 10, 2008 at 11:50 AM PDT #