Hot Chips 20

I’ll be at Stanford for the next few days for Hot Chips 20, a symposium on high performance chips. Sessions I’m particularly interested in:

D.E. Shaw’s specialized ASIC for molecular dynamics which I’ve written about earlier and IBM’s PowerXCell powering Roadrunner.
Upcoming architectures: AMD’s 780G and Intel’s Nehalem (dot products of special interest to me.)
Chips tuned for network or IO (Sun’s Rock, Fujitsu’s SPARC64VII and Intel’s Tukwila.)
Algorithmic content: Roofline models for automatic tuning of kernels (good addition to Demmel’s talk on the future of linear algebra from MMDS.)
Intel’s Larrabee: response to “the can of whoop-ass” (detailed architectural paper from SIGRAPH.)
CUDA: useful for a class of algorithms (based on memory access.)

I’m going to be trying something new this time — live blogging. I’ll try to push constant updates to my twitter stream : gane5h.

I’ll be staying at the Sheraton in Palo Alto. Drop me a line if you want to meetup for a chat.

This entry was posted on Thursday, August 21st, 2008 at 12:56 am and is filed under Activity, Computing. You can follow any responses to this entry through the RSS 2.0 feed. Both comments and pings are currently closed.

2 Responses to “Hot Chips 20”

siddharth says:

August 28, 2008 at 10:08 am

Hi,

The work you are doing, interesting stuff. Now, I’ve been working on amongst other things architectures for acceleration of linear algebra (BLAS3 and BLAS2 on FPGAs). Can’t meet up or chat or anything but feel free to mail me I guess. The work on protien folding – still computationally expensive – the ‘paradox’ still exists. Any ways, the conference how was it – what were the major implications for the GPU ? Unfortunately, your twitter stream ends up becoming rather difficult to follow.
ganesh says:

September 9, 2008 at 12:44 am

That’s cool. One of the showcases for CUDA was BLAS on GPUs. Do you really think it’s worth it in the long run when GPU have achieved commodity status? The biggest gripe I have right now is that the market it too fragmented and I don’t really know who is going to be around in two or three years.

Protein folding is the holy grail of molecular biology for a very good reason. Infact, we don’t even know if the force fields that have been parameterized work well for millisecond and microsecond long simulations. Fortunately, my company Zymeworks doesn’t do protein folding.

Twitter — Ah! I got a lot of private threats to stop spamming it. Let’s just say that it was a failed experiment for live-blogging.

Cheers.