Can we close the lid now? A64 vs AXP

[Buhammot]

Well, with all the flaming and debate that some users are having a hard time dealing with, I would like to thank jlin453 for posting their sandra benches at 2.78ghz with their XP-M. Their post can be found at the following address. (Older version of sandra, as suggested by AMD-Me)

http://www.ocforums.com/showpost.php?p=2503132&postcount=96

And on my behalf, I am runnin the speed in my sig, A64 2800+ @ 2.44ghz stock voltage. Being that the video cards are not comparable, no 3dmark scores will be shown here. Strictly CPU bench. If anyone knows of another bench, I will request they do it also, and I will post my results. My bench can be found at the following address.

http://img.photobucket.com/albums/v172/buhammot/Proof.jpg

AMD-Me's Results, using same version of sandra.

http://www.ocforums.com/showpost.php?p=3142737&postcount=16

Edited Picture to show a direct comparison to AMD-Me's Results.

 

[Quailane]

That athlon xp bench is rigged. The fsb bandwidth is so low! I think that if the fsb was more like 260+ the score would be a lot better. Either way, I thought everyone already knew that the A64 beats out the AXP by a lot?

 

[dicecca112]

Either way, I thought everyone already knew that the A64 beats out the AXP by a lot?

Some people don't take well to logic

 

[Buhammot]

I am just using his benches as something to compare it to. If someone would like to post their bench, feel free, and i will prolly even put it at the front too. We can get more accurate figures this way, rather than well so and so said...

 

[Yuriman]

I got about 16k in 3dmark01 with my 9700np with my A64 2800+ at stock, and about 16500 with my barton at 2.5ghz. Perhaps I will post screanies later.

 

[mitro]

were both running the latest version of sandra? An XP @ nearly 2.8 should be 11,000+ at 227x12 (2729) mine gets 11,300. Eh... when I get my DFI NF3 and 3200 DTR this week I'll have to do a shoot out.

 

[Buhammot]

Well I dont want to start a thread about which is overclocked higher, but simply what it takes to equal performance. I dont know if that was with the latest sandra, but if you would like to post a pic, i will reference it in the first post to have another comparison. I will be lookin forward to your shootout too

 

[koss20100]

could it be tht there arent many 64bit softwares tht are supported?

EDIT: software your using for benching?

 

[Buhammot]

I am using Win2k. So the 64 bit portion is neglidgable. Trying to level the playing fields as equal as possible.

 

[Jess1313]

were both running the latest version of sandra? An XP @ nearly 2.8 should be 11,000+ at 227x12 (2729) mine gets 11,300. Eh... when I get my DFI NF3 and 3200 DTR this week I'll have to do a shoot out.

my DTR and DFI get about 12040 at 2550mhz, thats 250mhz less than your A-xp.

another thing, i think that no matter what Sandra Arithmatic shows, the Bandwith of most/all a64 in SC is greater that almost any axp in DC

-Jess-

 

[Yuriman]

I edited my post lol, I was posting 3dmark. I prefer not to use sandra.

 

[CandymanCan]

Cant you just do a search on this, this has been posted over and over again, which is better or how much of a clock diff do they need to have to be equal....

 

[||Console||]

http://www.ocforums.com/showpost.php?p=3106545&postcount=21

and i know i can get more out of this 2

 

[jlin453]

Thanks for PMing me about this thread, I don't mind the comparison. I'd just like to add that I'd rather run an A64 at 2.5ghz than a 2.75ghz Barton ANY day of the week Actually, I think there's no comparison at all. High 3dmark scores are based on a whole range of things; overclocked video card, overclocked processor, the memory timings, etc. I have long since stopped benchmarking because there are more important things to me but that's just my opinion. If you use the same video card to run benchmarks on a A64 and a barton, the A64 would win (obviously, I guess I should add).


btw, fsb was set to 230 or whatever because that's all my motherboard could handle, modded or not.

 

[Buhammot]

Also, this is a direct comparison of cpu speed, not mem speed. Superpi is mem intensive, 3dmark uses video cards, there are some other benches that i dont have at the time that are cpu as well.

 

[AMD-Me]

Well, with all the flaming and debate that some users are having a hard time dealing with, I would like to thank jlin453 for posting their sandra benches at 2.78ghz with their XP-M. Their post can be found at the following address.

http://www.ocforums.com/showpost.php?p=2503132&postcount=96

And on my behalf, I am runnin the speed in my sig, A64 2800+ @ 2.4ghz stock voltage. Being that the video cards are not comparable, no 3dmark scores will be shown here. Strictly CPU bench. If anyone knows of another bench, I will request they do it also, and I will post my results. My bench can be found at the following address.

http://img.photobucket.com/albums/v172/buhammot/Proof.jpg

I hate to burst your bubble but your using different versions of sandra...

 

[Buhammot]

Ah! thank you much AMD-Me, as we have a very close match! 2.44ghz vs 2.7ghz, as i can hope we are both runnin Sandra 2004.10.9.133.

For the not so believers, is this the kind of benches to show what is needed? And 2.4 I dont think is my max oc, as typical varies from 2.4-2.6 on air. Either way, this IS NOT about prices, but sheer performance perspective.

Nice OC from both of you btw.

 

[AMD-Me]

Ah! thank you much AMD-Me, as we have a very close match! 2.44ghz vs 2.7ghz, as i can hope we are both runnin Sandra 2004.10.9.133.

For the not so believers, is this the kind of benches to show what is needed? And 2.4 I dont think is my max oc, as typical varies from 2.4-2.6 on air. Either way, this IS NOT about prices, but sheer performance perspective.

Nice OC from both of you btw.

Danke Sehr! I would be pulling 12k if it werent for crappy a7n8x, that VDD mod is INSANELY HARD! Much love to the fastest A7N8X , wait till i get off my *** and get a better Rad and PSU, then we will see 2.8ghz, i can POST up to 2.85. My poor lil Black ICE Pro is at its limit, thus i keep it at 2.64 for daily use.

 

[c627627]

http://www.ocforums.com/showthread.php?s=&postid=2762781 says

... the part of hitechjb1's post you're looking for is:

A 754 512 KB L2 A64 + 250 GB motherboard ... 15-25% average gain over a Barton (at same frequencies).

A 754 1 MB L2 A64 + 250 GB motherboard ... 20-30% average gain over a Barton (at same frequencies).

(And those are 754, not 939 numbers, they depend on the kind of work you're doing but it paints a good general picture)

 

[Gautam]

The raw processor power of the A64 and AXP are about equal. There are tons of higher-clocked AXP's that can thrash my system in Sandra, given that they're using the correct version.

The A64's main advantage comes from the on-die memory controller, which Sandra doesn't measure, but can easily been seen from running benchmarks that test the system as a whole.