Jump to content

Featured Replies

I tried 64,55,50,48,40,32,30,28,20,18,14 and I found that the sweetest spot was 18. It's a faildozer family so I don't expect anything fancy for this procs.

Good lord that's a lotta threads.

 

Bulldozer definitely sucks for floating point tests. If you run tests with multiple-core-optimized Integer calculation, that server would be insanely fast. :)

I already did a benchmark (HPL or high performance linpack) in linux and compiled linpack HPL2.0 and AMD's ACML5

with the latest mpi, but still 16x2 cores interlagos is no match for E5 2680 x2 (HT off).

 

Before the release of the faildozer (interlagos and Valencia) I already knew that the procs will not scale that much.

Edited by dhenzjhen

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...