« Computing is Beauty; Computing is Truth | Main | Reflections on Tibet »

March 18, 2008


Feed You can follow this conversation by subscribing to the comment feed for this post.

Dave Semeraro

I am excited about multi core and many core platforms. However I have been involved in parallel computing since the mid 1980s and have witnessed the search for the silver bullet that will make parallel computing "easy". In those days it was an automatic parallelizing compiler for distributed memory parallel systems. Expressing my pessimism at the prospect for such a thing at a conference lunch in mixed company, I was set upon by an angry attendee who said, "I suppose you enjoy writing MPI". Well, here we are about 20 years later and software for new parallel computer systems still includes MPI and, to the best of my knowledge, no automatic parallelizing compiler.

There is a fundamental difference in the situation now however. The platform is different and we have 20 years of experience that we, or I should say I, didnt have then. (Other people have more experience) Having said that, I am optomistic that the new partnerships at Illinois and Berkley will produce new tools, libraries, and ways to think about parallel processing.

I do, by the way, still occasionally write the odd bit of MPI code. Now where did I put that CUDA users guide...?

The comments to this entry are closed.

Dan Reed
Dan Reed

December 2023

Sun Mon Tue Wed Thu Fri Sat
          1 2
3 4 5 6 7 8 9
10 11 12 13 14 15 16
17 18 19 20 21 22 23
24 25 26 27 28 29 30

Twitter Updates

    follow me on Twitter
    AddThis Social Bookmark Button


    • Add to Technorati Favorites