Oldest Driver 8600m gs (released 2007.12.21) best one of all!?

[Santori]

Hey Folks,

i recognized a kind of weird thing about the driver 177.35 for my GF 8600M GS.

To begin with here the most important thing about my laptop:

17"

Asus A7SV Series

Core2Duo T9300@2x2.50 GHz

3070 MB DDR 2 800 MHz

GF 8600M GS 702/500 256 Dedicated via 1024 Turbocache (whatever that is...)

With the OEMs driver 2007/12/21 I had 3671 Credits @ 3DMark06

With some of the 1.66 i had between 2991-3321 Credits @ 3DMark06

The (i guess) newest CUDA 177.35 (Physx support?) one i had exactly 3000 (and after cooling down) 3002 Credits @ 3DMark06

One of the 1.66s Driver even changed my Index of 4.5 Vista Points to 4.4?!

I installed the oldest ones i know and it again runs best in Crysis, Bioshock (+7FPS maxed 1024x768 compared to the CUDA one)!

Why is that, anyone has a clue?

BTW the Asus Driver (guess the OEM Inf isn't modified or restricted to Asus , well maybe not ) can be found on Asus.com typing in A7SV in the Searchfield (upper left) Drivers->Display

Its called V7.15.11.6749 and for Vista 32!

Anyway which one is the FASTEST 8600M GS Driver, if someone already tested them all, or what do you use?

Greetings from Germany (we're in the European Championship finals yeaaaaaaah)

Thanks for replys byebye

Edited June 27, 2008 by Santori

[Michael Marley]

I use 177.39 for my 8600m GT and it performs best for I everything I do. I do not play either of those games. however.

[Guest CyGnus]

I already tested a few drivers (All 17x series and 169) the best driver for my 8600M GS is 169.25 WQOL

Asus Z53SV T7300.

The newest driver aren't always the better ones, Why is that?

Easy answer Newer GPU'S come out to the market and they have to had drivers for them as well and the 8xxx series cards already pass their time. Not that they aren't god but yes passed way. So CONCLUSION the Best performence drivers for 8600MGS in MY case are 169.25. Give them a try and compare 3Dmark06 and some FPS in Crysis or COD4 with the last ones and you will understand what i mean.

Chers from Portugal aka [CyGnus] www.Techzonept.com

[Covert]

I slightly disagree, the general trend for newer drivers is that they are better. 3DMark tests on a large range of driver shows a steady, increasing, trend line going up in the version number. All nVidia cards use the same unified architecture too so as drivers continue to be released all cards usually benefit, nVidia doesn't write completely different drivers for the different cards. It's mostly all in your head!

[zipper]

All nVidia cards use the same unified architecture too so as drivers continue to be released all cards usually benefit, nVidia doesn't write completely different drivers for the different cards. It's mostly all in your head!

No. The older cards generally work slower on brand new drivers when working on older software. Newer software usually benefits of new drivers. Tendency can be seen here:

http://forums.laptopvideo2go.com/index.php?showtopic=8245

[Covert]

No. The older cards generally work slower on brand new drivers when working on older software. Newer software usually benefits of new drivers. Tendency can be seen here:

http://forums.laptopvideo2go.com/index.php?showtopic=8245

Read between the lines :)

The link you cited only shows a fluctuation of about 100 3DMark points between the drivers on 3DMark03, not huge leap there! (In 3DMark 03 that is!). The 3DMark06 tests however show a steady increase over time. Retests usually show a variation of 5 points under the same testing environment and over time PCs tend bloat their registry a bit - I doubt the testing environment between driver benchmarks was exactly the same. What do you mean by older and newer software? Do you mean newer and older cards? Do you mean operating systems? Newer drivers = better performance or better stability at a very slight cost of performance.

With the latest (177.66) I got the highest 3DMark score I have ever gotten on 8600M GT (3871), which is higher than both my dell OEM drivers specific for my card and the oldest drivers I could find for my card.

I've benched a fair range of drivers done across 2 days without altering my system in any way except for the change in drivers and newer is (usually) always better in my case. Chipset drivers, DirectX version, drive fragmentation, operating system, startup files, and running processes can all affect your computers performance so drivers aren't always to blame if the benches aren't done on exactly the same software-set.

[zipper]

a bit - I doubt the testing environment between driver benchmarks was exactly the same.

No, but pretty standardized - extraneous unneeded processes stopped so just the same configuration was running. I've done this almost 4 years and tried to get comparable results.

[Covert]

No, but pretty standardized - extraneous unneeded processes stopped so just the same configuration was running. I've done this almost 4 years and tried to get comparable results.

So you ran the benchmarks with a fresh TinyXP installation? and ran the newer drivers on the directX version that came out 4 years ago? or ran the 4 year old drivers with the latest directX?

haha, anyway I take your word for it. :)

[zipper]

So, the dangers of tinkering experienced. I ran defrag at boot time (Diskeeper) and after reset my HD bombed out (famous missing bootmgr error). It´s quite difficult to fix a hd when it doesn't boot and the HP rescue didn't work either. Luckily had another lappie to get a repair disk from the net. After finally having a working chkdsk I finally could use my backup, just lost a weeks worth of emails. Took 4 years until a defrag operation bombed, but it's always possible.