Excellent review of ATI and NVIDIA's dual-core driver

[Unkno]

ATI's dual-core driver

ATI's X1000 Performance difference

Nvidia's dual-core driver


It seems that both ATI and Nvidia are getting a pretty good performance boost with the dual-core drivers.

 

[KeepItRed]

Good review, I hope to see what my AMD 4400+ and X1800XT can really do.

 

[Unkno]

for the x1800 series, i am 100% sure that you would get at least 15% higher with newer upcoming drivers (like in 6.0, 6.1....).

 

[nRollo]

Originally posted by: Unkno
ATI's dual-core driver

Nvidia's dual-core driver


It seems that both ATI and Nvidia are getting a pretty good performance boost with the dual-core drivers.

Thanks!

Good post, I never checked this although I've always wondered about it.

 

[videoclone]

I did upgrade to 5.12's a few days ago good thing to know i am getting those extra frames super.

 

[MDE]

Nice, ATI finally multithreads their drivers right when I buy an ATI card .

 

[CP5670]

I don't think these reviews are very good. They have the same problem as all the other articles I have seen on these drivers; they don't test a single core on the same drivers. It's hard to tell how much of a boost is from the multithreaded aspect of the new drivers and how much from the other optimizations in them.

 

[VirtualLarry]

Would be interesting to compare against a "mature" SMP/multi-threaded video driver set, say the ones for Permedia or Oxygen cards. Perhaps a future workstation benchmark roundup might include those.

 

[Asymptoke]

Originally posted by: CP5670
I don't think these reviews are very good. They have the same problem as all the other articles I have seen on these drivers; they don't test a single core on the same drivers. It's hard to tell how much of a boost is from the multithreaded aspect of the new drivers and how much from the other optimizations in them.

100% Agreed. They need to stick in a comparable single-core CPU and rerun the tests. Heck, even a single game!! I want to see a direct comparison and not look up two different articles on two "similar" systems tested by two different reviewers with previous driver releases.

 

[Unkno]

there'd be absolutely no point in testing that way. BECAUSE with dual core, the processes of the OS and any background programs are proccessed in the second core, leaving the first core to completely run only the game. With single cores, the processes AND the game have to run together.

 

[Asymptoke]

You've just strengthened my argument: using your example, testing both single and dual-core will demonstrate the cost of running those background tasks. I do see what you're getting at though and I could have phrased it better...

 

[Pete]

It'd be interesting to test Q4 SMP performance with pre- and post-dual-core-optimized drivers from both IHVs.

 

[w00t]

just another reason for me to upgrade to dual core

 

[Unkno]

updated, now with link to review of the huge difference it made to the X1000 series.

 

[chinkgai]

they need to have aoe for a testing game more often dammit