Discussion in 'Videocards - AMD Radeon Drivers Section' started by TwL, May 29, 2010.
Riva Tuner and CCC show me the exact same temperatures, same as with previous drivers.
I wonder, if anyone would care, if we would lose some FPS, support for SC2 AA in cost of absolute clean processing. I found a way to make the processing
very fluid through 1 of the 10.8 builds. partially like on 3.95 was used, but even further pressed up. Sad thing in this is that it loses some 'highest' FPS and I
cannot use DirectX Driver from any of the new releases to do this. Of course DirectX 10-11 would be new, but that's a side matter to this.
What I am driving with this is that this would be really nice bottom line driver to use with anything where we would be able to get near issue free environment on 10.8->10.9 CCC and not continue the v10.7/a-b/10.8/b5/6/b/10.9/a none sense what was started.
Depends on the definition of 'some FPS' I think
well, it's -2-3 FPS on DirectX 9-11 slower with absolute no cuts/jumps or anykind of drawing errors.
Yes, I would love to use that driver! As I've said before, the loss of a few fps is no sacrifice at all if in return we get no jumps and smooth fluid gameplay.
I would also love to see the AA option in-game for SC2.
Well, I am testing one right now using 10.9a baseline because I know it draws correctly. I got 2 FPS increase actually on this now on DirectX 9. Have no clue on DX10-11, but DX10-11 drivers are new as long base is clean the rest will follow.
I'll dump some experimental set here call it v4.20a in few.
Agreed. I'm already forcing 24xaa on games for the fun of it. It would be nice for it to be fluid.
As most already stated, max fps can be sacrificed for better stability between lowest and highest fps.
The closer those 2 are, the more fluid game-play can be expected.
I even think current 4.20 goes far beyond my 60Hz V-Sync, since the card seems to get 5C hotter when it is under heavy load condition.
I wouldn't have anything against a decrease in max fps if it can smooth up the game-play.
Dude your awesome :banana:
Can't wait to give that one a go! I grow weary of strange hitches, though few they are.
Ok, public vote wins..
I build 3 builds under same condition now and here's the builds descriptions:
Experimental (Zero error tolerance builds) v4.21a-c:
None of the builds are bad they are all very usable. Selection of which driver to use depends on usage. Ideally BF:BC2 would most likely work fastest on build B while game could run cleaner with Build A and for example DirectX 9 games (disregarding SC2) all DirectX 9 games would run fastest and cleanest with Build A while SC2 would need either Build B or C and probably run smoother with Build B, if you run an CrossfireX best idea would be the Build C over any case Build B might work, but none of the builds are meant to be run with anykind of CrossFireX setups.
* All Builds baseline driver is v10.9a.
* All builds are fixed with new cleaner/Installer v3.6.1 so Application profiles are applied now correctly.
-Build A: This is the fastest of 3 with most fluid processing & over all very good driver using NATIVE 10.9a DirectX 10-11 & 10.8b6 DX + Processing. StarCraft 2 Anti-Aliasing is NOT support.
-Build B: This is an very close attempt of creating same effect of Build A slower driver DirectX and processing is based on 10.8 WHQL with NATIVE faster DirectX 10-11. With StarCraft 2 Anti-Aliasing support kept.
-Build C: This is same as Build B with exception this build uses closest possible DirectX 10-11 driver TO the 10.8 WHQL creating very good result in processing through in high FPS and on extreme CPU/Memory stress. With StarCraft 2 Anti-Aliasing support kept.
Experimental Build v4.21a - 32-bit / Experimental Build v4.21a - 64-bit
Experimental Build v4.21b - 32-bit / Experimental Build v4.21b - 64-bit
Experimental Build v4.21c - 32-bit / Experimental Build v4.21c - 64-bit
These experimental builds will not be uploaded to another server do not ask. They may be added on first page, if people prefers cleaner drivers instead of v4.20 before I get the
'Final Build' back in the clean order and processing truly faster to attempt to create something even better. Ideally I am still 1o5 with 10.8-10.9 driver series I have some
understanding on them but since v10.7a I have more and more lose track which I am gaining, atm. You need to understand what you 'have' before you can build clean-up ATI
Driver Team didn't make it very easy with 10.8 WHQL and onward.
I'll try Build B with Starcraft 2 whenever this download finishes.
I would like to test out the v4.21a due to it's superior performance.
I am not using AA for my SC2. So, it ain't a problem for me.
gonna try c out seen as though i play alot of bc2 also checked out dragons age and i might as well wait for the ultimate edition to come out which includes all dlc and expancion packs,
Downloading Build A.
Will test tomorrow when i get my new mobo and memos
Thanks for the excellent work man!
PS: As a side note. If ppl want speed but AA on SC2, just use RadeonPro.
PS2: Also, i just saw you are some kind of "legend" at NFORCERSHQ.
Dind't noticed till i saw (again after 2 years) your name there in the nForce IDE Driver cleaner xD
Whic set recommend to play Battlefield BC2 and Dirt 2, i only play this tow games right now
Bye and thanks for all the effor u put in every new build
Well, I did there so many months of benchmarking on nforce 430 and multiple other nforce boards 'Fernando 1' most of kept me in books(wow, seems his moderator these days there) and later on added
SoundMAX ADI and RealTek sound driver sections and did a minor testing on NVIDIA network adapter drivers on nforcersHQ, but they changed the forum
style and I really dislike the new forum styles which made reading and writing clean really hard. So, didn't do much after forums changed.
Just wanted to give you guys a bit options which to try and while at it some clean builds where there's no attempt to create an miracles, but simply use an 'good driver'. Ideally when I have an issue with an game here these are the kind of stable builds I use for testing new games. Also to show in real sense these 3 packs that there's an huge difference between 2 of the driver modules we are using there. That is the turn point when 'maybe working' turns to Jimmy Hoffa.
hmm, if someone missed the post on experimental builds. It's RIGHT HERE from last page. Let me know how they go on.. Shouldn't give much of an trouble.
these experimental drivers are meant for single card users though right? not crossfire + physx users like myself?...
Yep, they are for single card users only. They are missing crossfirex mods, but I tested today one HD4 CFX I think I can improve the 4.70a quite a bit.
I'm running 4.21a's and they are still excellent for all they have no crossfire mods in place. Crossfire has been very good with official drivers since 10.8 on both Dx9 and 10 so I thought I'd give these a blast as they are the most stable.