for testing/development and other purposes.
Config is reasonable, with an upgrade later today. 2 x 2.66 GHz (5130) processors, 4 GB ram, nice video card (Quadro FX/4500).
Installed SuSE 10.1 with updates/patches.
My expectations were that this machine would positively blow the doors off of a similarly clocked Opteron (252). Given the massive hype around Woodcrest, this is what one might expect. If you are going to hype like mad, you need to be able to deliver on the hype.
So I ran our Scalable HMMer, GAMESS, and a few other tests. Same binaries on both machines. Made it easy by using the fastest builds we had done previously.
End result? The 2.66 GHz Woodcrest was 10-20% SLOWER than the 2.6 GHz Opteron 252 in a Sun v20z platform. 20% slower on Scalable HMMer, 10-12 % slower on GAMESS.
Going to work on this a bit more in my (copious, heh!) spare time. I suspect that Woodcrest needs to have code rebuilt for it to make effective use of it.
Oh… I do have the 3.0 GHz chips, but I want to do a little more playing with the indentical clock speed so we can see how well it does vs Opteron when you remove clock speed as a consideration.
It is too early to draw conclusions from this. My expectations had been set that this unit would blow the doors off of anything we had. Not sure if I believe that a 15% clock speed delta (2.66->3) is going to make much of a difference. I think it is going to be compiler based if it is anything.
Working with Intel on this, need to understand if we are seeing what we should be seeing. I have had enough machines shipped to us in the past from vendors with mistaken configs. Our customers don’t appreciate this, and I can tell you it isn’t fun chasing down things that I shouldn’t have to. That said, Intel support is excellent.
Update: will do the upgrade later in the week, too much on my plate that is due soon. Boss won’t let me play until I make money for the company.
Viewed 16916 times by 3375 viewers