[hsieh]

Hi,

I am wondering if a dual-core CPU (AMD/intel) helps in OpenFOAM calculation?

How much speed can be gained when comparing a dual CPU system to a single CPU system.

Anyone tried AMD 64 FX 57 CPU? This is a gaming CPU, I will assume it will be good for OpenFOAM too when compared to something like AMD 64 4800+, right?

Any suggestion will be highly appreciated.

Pei

[eugene]

If I had to choose between a FX57 and a 4800 X2, I'd go for the X2. Even with the additional strain on the memory bus, two cores running in parallel will beat one every time.

I would guess the two 2.4GHz cores should be around 1.5-1.8 times faster than the 2.8GHz single core, depending on the case size.

[hsieh]

HI, Eugene,

Thanks for the reply.

Can OpenFOAM handles dual core AMD chips?

Pei

[mprinkey]

I would be careful about making general statements about OpenFOAM and its scalability in SMP environments. Our tests with OpenFOAM on dual xeon systems show very poor performance with two CPUs sharing the same memory bus. This situation is analogous with dual-core Athlon64/Opteron CPUs. Benchmarking is the only real way to be sure.

[henry]

We have also obtained poor performance on dual CPU systems sharing the same memory bus, basically OpenFOAM is often memory access limited. However on dual Opteron machines we get very good performance but on those systems each CPU has it's own memory bus.