Conroe 64 bit performance?

[Uriel]

I've seen a few Conroe gaming benchmarks around. Look pretty impressive. I'm just wondering how it would do in a 64 bit environment, compared with AMD64 or AM2. Anyone seen any useful info?

 

[fenderbass86]

It'll blow all the AMD's away. Enough said, just gotta wait for them to hit retail now.

 

[slap ed]

i'm sure i read in micro mart in the Russell barnes section that the initail test results comparing conroe to an fx60 turned out to be rigged . intel had let out some crucial drivers for the amd rig thus slowing it down so all may not appear as it seams

 

[Defcon5]

i'm sure i read in micro mart in the Russell barnes section that the initail test results comparing conroe to an fx60 turned out to be rigged . intel had let out some crucial drivers for the amd rig thus slowing it down so all may not appear as it seams

I think toms repeated the tesst with their own gear and got the same results

 

[Socaddict]

It was anandtech, wasn't it, who did all the testing at IDF.

 

[BubbySoup]

Yeah, I think they did all the testing, but iirc they were given the PCs "as is" and were not allowed to mess about with the drivers or config of either PC, just run some benchmarks on them.

 

[Corasik]

Thats why Anandtech built their own overclocked [email protected], with a more uptodate bios, and the latest drivers, and got very similar results to the intel build FX60 system.

The updated bios made some minor improvements to the AMD system. However, they also realised the Intel memory timings were wrong, correcting this made similar improvements to the intel system as changing the bios did for the AMD system.

They did admit to having found an error with the 'Fear' test which reduced the Conroes lead in that test from 40% to 20% (They originally had the AMD at 1280x920, and the intel at 1024x768) OOPS. Even so a 2.66Ghz Conroewas still 20% faster than a 2.8Ghz FX60 when they rebenched the AMD with the correct resolutions..

 

[Defcon5]

However, they also realised the Intel memory timings were wrong, correcting this made similar improvements to the intel system

Intel got the memory timings of their own chip wrong?

 

[Corasik]

Seems so yup. But you know computers.. Probably just had it set to SPD, which in my experience always ends up setting slack timings.

Seems they had set up with 5-5-5-15, and Anandtech managed to get 4-4-4-15 timings, which did improve performance slightly. Eitherway, neither the AMD bios, nor the Intel memory timings made a major difference, 0% to 2% depending on which test they ran.

So the AMD system wasnt 'rigged' to give poor performance, and the Intel wasnt 100% optimized for absolute max performance either. With the exception of the AMD being overclocked to 2.8Ghz, the systems were setup 'ok' but not tuned as a 'fanatic' might have done.

Personally I think the point here is that intel dont need to resort to tuning the last drop of performance from conroe to demonstrate its faster than the fx60. And the fact they overclocked the FX60 from 2.6 to 2.8Ghz should be more than enough to make up for and lazyness in perfecting the configuration.

 

[Iraklis F.C.]

VISTA will show off the performance in 64bit OS