Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

A Cheap 12-Core, 30-Watt Ubuntu Cluster

phoronix.com

Post reply on HN
Print view View original

1–10 of 35 posts

Re: A Cheap 12-Core, 30-Watt Ubuntu Cluster

#4

Post by sp332 »

iSloth wrote: post #2

Would be interesting to know what application it's actually going to be running? what task it's going to be doing?

It's for Phoronix to test new ARM-based software products, mainly benchmarking. The second paragraph:

This cluster will be used for delivering some interesting ARM Linux benchmarks, of course! It's also being used for Phoronix Test Suite purposes for creating more MPI/cluster benchmarks for some of the commercial/enterprise clients, coming up with more ARM Linux benchmarks in general, and this hardware is also going to be part of a much larger (~96+ ARM core cluster - details to be shared at a later time) cluster.

Re: A Cheap 12-Core, 30-Watt Ubuntu Cluster

#5

Post by mtgx »

fiatmoney wrote: post #3

For comparison, you can now get http://ark.intel.com/products/65703/Intel-Core-i5-3470T-Proc... http://ark.intel.com/products/65735 http://ark.intel.com/products/65714/Intel-Core-i7-3517U-Proc...

You're giving Intel graphics as a pro for servers using OpenCL? Either way, I don't know about this set-up, but next year a set-up with Cortex A15 and Mali T658 will have shared cache between CPU and GPU, which should be a lot more efficient than anything Intel or even AMD has today regarding GPU compute. From what I understand cache "latency" between CPU and GPU is a pretty big problem in the desktop space, and this ARM concept should make it a lot better.

http://www.cnx-software.com/wp-content/uploads/2011/11/arm_s...

The "Little CPU" is supposed to be the lower-power Cortex A7. Canonical is very aggressive about supporting ARM chips.

Re: A Cheap 12-Core, 30-Watt Ubuntu Cluster

#6

Post by sliverstorm »

mtgx wrote: post #5
fiatmoney wrote: post #3

For comparison, you can now get http://ark.intel.com/products/65703/Intel-Core-i5-3470T-Proc... http://ark.intel.com/products/65735 http://ark.intel.com/products/65714/Intel-Core-i7-3517U-Proc...

You're giving Intel graphics as a pro for servers using OpenCL? Either way, I don't know about this set-up, but next year a set-up with Cortex A15 and Mali T658 will have shared cache between CPU and GPU, which should be a lot more efficient than anything Intel or even AMD has today regarding GPU compute. From what I understand cache "latency" between CPU and GPU is a pretty big problem in the desktop space, and this…

I'm probably just missing something, but can you explain how the concept is going to fix cache latency CPU GPU?

Re: A Cheap 12-Core, 30-Watt Ubuntu Cluster

#8

Post by pyre »

sliverstorm wrote: post #6
mtgx wrote: post #5

Earlier quoted context omitted.

You're giving Intel graphics as a pro for servers using OpenCL? Either way, I don't know about this set-up, but next year a set-up with Cortex A15 and Mali T658 will have shared cache between CPU and GPU, which should be a lot more efficient than anything Intel or even AMD has today regarding GPU compute. From what I understand cache "latency" between CPU and GPU is a pretty big problem in the desktop space, and this…

I'm probably just missing something, but can you explain how the concept is going to fix cache latency CPU GPU?

I'm assuming this:

  Cortex A15 and Mali T658 will have shared cache
  between CPU and GPU
is what you're looking for.

Re: A Cheap 12-Core, 30-Watt Ubuntu Cluster

#10

Post by madmaze »

fiatmoney wrote: post #3

For comparison, you can now get http://ark.intel.com/products/65703/Intel-Core-i5-3470T-Proc... http://ark.intel.com/products/65735 http://ark.intel.com/products/65714/Intel-Core-i7-3517U-Proc...

the wattage measured here is not just the processor, its also the entire boards.

Ive done some benchmarks testing Intel Atom processors Atom330's and D525's against Tegra3 processors. On average the Terga3 processors outperformed the Atoms 4x-8x in Flop/s per Watt.

Admittedly that is not the only important metric, but gives an initial performance comparison between ARM and x86 performance

Post reply on HN