• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Can we close the lid now? A64 vs AXP

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
For raw CPU power, you can estimate like this:

Use Sandra CPU Arithemetic Benchmark,
based on Dhrystone ALU Integer MIPS,
AMD Opteron 148 2.2 GHz 1 MB L2 = 9793
AMD Athlon XP 3200+ 2.2 GHz 512 KB L2 = 9142

So at same CPU clock, the A64 is 7% higher than AXP in integer arithemetic performance. Floating point and multimedia benchmarks can be done similarily.

This version of Sandra 2004.10.9.133 does not provide benchmark reference for 754/939 at 2.2 GHz, so Opteron reference is used, until 754/939 references are available.
 
Sisoft sandra is just a crappy way of comparing A64's to AXP's, Sisoft Sandra doesnt fully take advantage of what the A64 can do.

Take the Tbred and Barton for exsample. The Bartons extra 256k cache makes it a little faster then the Tbred, but sandra scores dont change even 1%, you needs to use something to compare this buy cuss the A64 is about 25% faster then the AXP'S.

This was already done like long time ago. The A64 Newcastle is faster then the Barton by around 300-400Mhz, the Clawhammer is around 400-500mhz. A 3ghz Barton is about the same speed of a 2.5-2.6ghz Clawhammer and Newcastle.
 
this is my 2penny's worth
 

Attachments

  • sisoft2.7.JPG
    sisoft2.7.JPG
    53.3 KB · Views: 19
  • 2.70.JPG
    2.70.JPG
    41.1 KB · Views: 18
  • temp.JPG
    temp.JPG
    13.9 KB · Views: 15
I guess I'll post my sandra stuff, but again sandra doesnt show the real 25% diff the A64 has over the AXP. People need to stop relying on sandra to compare cpu's. This was done using 1 stick of my mem to reach my max highest stable oc on the cpu and memory 284x9 4xHTT, my bus speed was 2272 . Btw i have proof this 2555mhz is stable.

Like Gautam said its the on die mem controller which the A64 has to thrash any AXP

2555mhz Cpu-z

284mhz Cpu-z

2555mhz Sandra 1

2555mhz sandra 2

284mhz mem 1
 
Last edited:
eR1k said:
Hi guys,
Here's my contribution to the discussion. The sisoft sandra version is 2004.10.9.133.
- Athlon xp-m 2600+
- Mushkin bh-6 @ 210mhz 11-2-2-2

http://kubnw11.kub.nl/~s304879/sandra.jpg

Time to compare ;)

Cheers!

That's quite an overclock... :drool:

how many stage phase change is that, now?


I agree, Sandra isn't useful here. Mobile bartons will generally win on raw MHz alone. If you want raw MHz, get an Intel. ;)
 
Here is mine:

untitled331.jpg
 
I think that Sandra is such a poor frame of comparison between an A64 and AXP/AXP-M. As was stated earlier, the MAIN difference between an A64 and AXP is the fact that the A64 possess an on-die memory controller, which removes most of the latency and timing issues once plagued by the AXP. Because of this fact, and A64 will typically always beat an AXP clock for clock in most applications. I think that most of us could care less for Sandra CPU benchmarks as P4's usually take the crown here despite the fact that a lower-clocked AXP/A64 would be faster in real world use. Sandra should always be taken with a grain of salt; therefore, I see no point in these moot comparisons.

deception``
 
Gautam said:
That's quite an overclock... :drool:

how many stage phase change is that, now?
Single stage :drool: I'm using a machII R507 to cool this beast :cool: And the best part is, that my cpu is fully stable at this speed. I can game all day long without a single crash :D
I now this baby hasn't reached her limit yet. Its just because i haven't got any volt mods whatsoever. It seems useless to raise the vcore above 2.1. My nb voltage is only 1.7 :(
 
Last edited:
Back