Phenom II vs Zambezi core speed

realdasein

Distinguished
Aug 20, 2011
76
0
18,630
Hi Guys,
After spending two days researching the following topics, I’m going to lay them out here and hope that some nice veteran will feel like giving specific answers.

So here’s what happened--I built my first system six months ago:
AMD Phenom II X6 1100T BE Thuban 3.3GHz, Socket AM3, 125W, Six-Core CPU
Gigabyte GA-990XA-UD3 motherboard
8 GB RAM

When tested on whohasthefastestcomputer.com, my CPU consistently gets a score of 19.52 gigaflops. There is no overclocking going on.

I recently built another computer for my dad:
AMD FX-6100 Zambezi 3.3GHz Socket AM3+ 95W Six-Core CPU
ASUS M5A88-M AM3+ Micro-ATX AMD motherboard
16 GB RAM

His is much slower—getting a score of 8.5 gigaflops on the same site. Whohasthefastestcomputer.com uses a "flopsmeter" to test a CPU by giving it millions of floating point operations to solve. As such it is only supposed to measure CPU “brain” speed, and cuts out peripherals such as graphics, internet speed, memory amount, etc.

So here are my questions:

1) I’ve read a lot about motherboard BIOS versions messing things up with FX processors when FX first came out. But does the BIOS version matter for CPU speed, or just for whether the thing boots up at all? Also, the ASUS board allows for “core-unlocking”: all six cores are unlocked. There is no overclocking going on.

2) Does the fact that it’s a micro-ATX instead of a full ATX motherboard matter?

3) Is the Flopsmeter meaningless? If so, is there a benchmark that measures gigaflops that you can recommend, so I’m comparing apples to apples?

Thanks everybody for any concrete thoughts!
-Heather

 
1) It might with the added functionality of the FX series. I doubt it's the source of your "problem" though. I'd update the BIOS simply because it's so easy to do on modern mobos.

2) Nope.

3) Difficult question. Floating point operations is perhaps the weakest link of the FX series. The most common CPU tasks are integer operations though.
 

loneninja

Distinguished
Solution
As has been said, the BD arch is crippled when it comes to FP performance. BD was designed for servers, first and foremost, where Integer math dominates. AMD was hoping OpenCL/OpenMP would take off, and all FP work would be offloaded to the GPU. This has not happened, hence the poor FP performance.

Its accepted that at a given clockspeed, a PII X4 would be about 10% faster then BD.