[Beowulf] x86-64 NUMA vs SMP kernel: appl. performance?

[Greg Lindahl]

On Fri, Sep 24, 2004 at 03:04:05PM -0400, Robert G. Brown wrote:

> What compilers have you tried, and what improvements do they
> produce?

Robert,

As you might recall, I do work for a compiler company, so obviously that
should be kept in mind. The 3 apps mentioned by the original poster
are Gaussian, Gamess, and NWChem.

For NWChem we have a customer who reported this number for one dataset:

| Opteron-2.2+portland group:
| Total times  cpu:      635.6s     wall:      636.6
| 
| Opteron-2.2+pathscale-1.1:
| Total times  cpu:      514.4s     wall:      515.5s
| 
| Opteron-2.2+pathscale-1.2:
| Total times  cpu:      480.8s     wall:      482.3s

So that's 32% better wall time for this dataset.

Here's a customer-reported result for GAMESS:

>	path     CPU=   1  CPUTime=  124.84s  WallTime=  126.10s
>	ifort    CPU=   1  CPUTime=  148.10s  WallTime=  148.20s
>	pgi      CPU=   1  CPUTime=  224.17s  WallTime=  225.60s

I can't seem to find a customer-reported result for Gaussian, although
several of our customers are running it.

-- greg