AMD Catalyst 10.12 Performance Analysis

[W1zzard]

To read this review go to: http://www.techpowerup.com/reviews/A...2_Performance/

[bear jesus]

Thank you so much for this, hopefully it will cause a few less comments about what driver version is used in reviews i know it wont but at least this is the perfect thing to link to when people do start complaining.

[Dj-ElectriC]

Wow w1zz this is brilliant!

[HXL492]

Found this article absolutely excellent and helpful!

[Benetanegia]

Excelent review. It finally puts an end to the driver improvements myth.

[Sasqui]

Really interesting that the 5870 saw a 3% gain from 10.11 to 10.12, but the 6870 didn't see squat.

Hope we see more of these comparisons in the future! Must have been a lot of time spent to do all the tests and compile results. Thanks W1zzard!

[20mmrain]

I am also glade someone did this so possibly AMD will get off their high horse and actually see that there is really no benefits coming from certain drivers.

That might be away to take this review that was never intended. But to me it is a perk.

Besides that thought pattern I like this review for the fact that it gives people info on how good a driver is before they upgrade. If these types of reviews were put into effect with every CCC.... it would take allot of guess work out of the whole situation.

Thnax Wizz

[bear jesus]

Originally Posted by 20mmrain

AMD will get off their high horse and actually see that there is really no benefits coming from certain drivers.

The only problem i have with that comment is that all the improvements are normally listed as game specific and many improvements are not games that are bench marked here/most sites or even played as much anymore in some cases.

From this article what i take is that there are improvements between driver versions but on average the improvements are low enough and also game specific enough for it not to be worth anyone complaining about what driver version is used in a review.

[human_error]

Thank you for doing this review w1zz. It's very interesting to see how small the differences are with new drivers over a long period.

I would love to see this become an annual feature where we can see the changes in performance of the top cards for the past couple of generations from each vendor with the different drivers, to help prove or disprove the myths about magical performance boosts people claim cards will get with new drivers.

I wouldn't mind seeing some sli/cfx comparisons as well so we can see if scaling is getting better or not over time (maybe it will help encourage the amd and nvidia to actually make some improvements over time).

[bear jesus]

Originally Posted by human_error

I would love to see this become an annual feature where we can see the changes in performance of the top cards for the past couple of generations from each vendor with the different drivers, to help prove or disprove the myths about magical performance boosts people claim cards will get with new drivers.

It would be nice to see how performance differed from the drivers given for the release review compared to the drivers available when the next generation is released as i think that would give a great picture of real improvements over the time high end cards really are current high end cards.

[yogurt_21]

Originally Posted by Benetanegia

Excelent review. It finally puts an end to the driver improvements myth.

yeah 5-6% from launch drivers isn't really that much considering how many drivers were inbetween and the length of time.

it's certainly not enough to change card performance placements.

though i will say if you've got a gtx480 with 197.17 and you're playing bad company 2.... you need to udate your drivers lol.

none of the others were that dynamic.

[mrg666]

Thank you! I always check Techpowerup for the GPU benchmarks and this one is a classic.

[lucassp]

XVID = DivX (old version) = MPEG-4 SP/ASP, so both can be decoded by newer ATi and nVidia chips.

[Black Panther]

Great review! I was a tad disappointed not seeing the 5970 though...

[Wyverex]

This is one of the most awesome reviews I ever saw. Thank you, W1zz!
Very useful

[63jax]

very good article. thank you!

[Super XP]

Very interesting indead. I didn't think Drivers made that much of a difference. It does seem as though CCC 9.x had better performance than the CCC 9.12 for the HD 4870's. The CCC 10.12a seemed to have picked up some performance for the HD 6970 card(s) and brings it on par with the GTX 580 in some games.
Like always, Great Review Wiz once again.

[dir_d]

Thank You w1zz this review is great.

[Dice]

Very interesting!
Bit furstrtrating for me as all the recent drivers for my 5870 - the ones that actually increase performance 10.10&e .11 & .12, cause terrble stuttering when watching tv in wmc (live or recorded) anyone else had this issue?

[T3RM1N4L D0GM4]

W1z is magic and this review is what all we ever want!

no more myth plz...

[razaron]

Interesting review. It amuses me how the results for less popular games are so erratic.

[avatar_raq]

A much needed review. Thanks W1zz! Hope you can add CF and SLI scaling in the future as well as eyefinity/surround resolutions.

[HellasVagabond]

Originally Posted by Benetanegia

Excelent review. It finally puts an end to the driver improvements myth.

How do you figure ? ATI still has their Low Quality setting as Quality and their Quality as High Quality....And i really doubt they are changing this any time soon.

[Mussels]

Originally Posted by HellasVagabond

How do you figure ? ATI still has their Low Quality setting as Quality and their Quality as High Quality....And i really doubt they are changing this any time soon.

and on the latest nvidia drivers, their HQ setting doesnt even work. whats your point?

[HellasVagabond]

Looking at your specs Mussels i really don't know how you can tell what the new NVIDIA drivers do or don't but i am not going to start an stupid flame. The proof is out there for anyone who wants to know the truth behind the performance numbers ATI has been getting lately, that's all that matters to me.