PDA

View Full Version : Who else thinks modified drivers don't add crap to performance?


glObalist
03-23-06, 01:00 AM
Based on a suggestion from another forum I tested the new XG modified drivers against some nVidia ones. Here is the results, the difference in performance between the various drivers being well within what I call "benchstatistical error":

81.94
AquaMark03: 63019
3DMark05: 4166

XG driver
AquaMark03: 62877
3DMark05: 4171

84.24 (the "Oblivion" driver)
AquaMark03: 62569
3DMark05: 4161

I don't know how about others, but in my book this does not translate to a measly one frame per second difference.


Yes I did use DriverCleaner and all the necessary procedures in between the various driver installs.

And yes, based on these and other results from the past I think this driver-tweaking business is a hoax no matter how many people claim OMGZ MY FRAMERATE WENT UP A GAZZILION FPS AND TEH IQ IS UBAR!!!!111 BST DRIVER YET1111!!!!ONEELEVEN

Discuss.

Peoples-Agent
03-23-06, 01:30 AM
The idea isn't to just add Performance, if you want that... just set the control panel to High Performance mode.

They make a stack load of difference to IQ, and I get more fluent gameplay using SLI.

Tbh, they make the Vanilla IQ look very washed at times, and they iron out a kinda judder you get with some SLI modes.

And yes, based on these and other results from the past I think this driver-tweaking business is a hoax no matter how many people claim OMGZ MY FRAMERATE WENT UP A GAZZILION FPS AND TEH IQ IS UBAR!!!!111 BST DRIVER YET1111!!!!ONEELEVEN

You aint got a clue if that's what your benching with lol

Run some ingame benchmark tests that use the game engines, then compare the performance and the IQ.
Using Aquamark and 3dMark don't cut the mustard i'm afraid, try say the built in Fear test.

RejZoR
03-23-06, 02:26 AM
I think the same...

The most funny was the so called boost of Omega Drivers, but then i see the charts in the same review. 1fps gain in tests and some even with no boost at all. I mean lol? I have nothing against Omega drivers guy and i respect his effort but these modded drivers are manly joke. Wanna bench? Rew up the LOD bias and mipmap, enable optimisations and tweaks via lets say RivaTuner and voila. I tried few of the most known and i haven't seen any boost at all.
Now i simply use official drivers delivered by the vendor (let it be NVIDIA ForceWare or ATi Catalyst). Always worked like a charm :)

RanCorX2
03-23-06, 06:01 AM
I agree, i've tried many modded drivers but have never noticed any increase in fps or better iq, most of the time worse iq and poor fps.

Shamrock
03-23-06, 07:13 AM
I had a set of XG drivers (81.98) mess up my entire Windows so bad, I had to re-format, I couldnt even use driver cleaner. I will now only use drivers if they are off of NV's website.

j0j081
03-23-06, 08:00 AM
Well if you are complaining about the XG drivers not adding little if any performance the point of them is to increase IQ with no performance drop and sometimes they do have a slight benchmark advantage so you are barking up the wrong tree.

GOZ
03-23-06, 08:13 AM
Well if you are complaining about the XG drivers not adding little if any performance the point of them is to increase IQ with no performance drop and sometimes they do have a slight benchmark advantage so you are barking up the wrong tree.

Well said:D

saturnotaku
03-23-06, 08:27 AM
I've always found modified drivers don't provide any difference in performance or IQ compared to the ones released by the chipset makers--NVIDIA or ATI. What really bugs me about 3rd party sets is that they copy the driver files to a folder and create a start menu entry. Just package them as a self-extracting executable (with WinRAR or 7Zip) and let the user choose what to do from there.

GOZ
03-23-06, 09:21 AM
I've always found modified drivers don't provide any difference in performance or IQ compared to the ones released by the chipset makers--NVIDIA or ATI. What really bugs me about 3rd party sets is that they copy the driver files to a folder and create a start menu entry. Just package them as a self-extracting executable (with WinRAR or 7Zip) and let the user choose what to do from there.

you can uncheck the start menu entry on install, and the stock drivers extract to a folder too, and you can choose were you want it to go with the xg installer.

glObalist
03-23-06, 09:26 AM
Well if you are complaining about the XG drivers not adding little if any performance the point of them is to increase IQ with no performance drop and sometimes they do have a slight benchmark advantage so you are barking up the wrong tree.

Hehe, maybe I should have mentioned the xg drivers was advertised as follows:

"If you have been using the 7x.xx series until now, thinking that most of the 80 series drivers just havn't stood out much, this is the time to upgrade.

The 84.12 is the best driver for Geforce 6 and 7 cards available. Image quality is brilliant, performance has increased, it feels smoother in heavy games such as COD2, Quake 4, and BF2, it is stable, it has no known problems, and is all around the very best driver of the 80 series.

Well, now Daishi and JRd1st from TweaksRUs have released an Xtreme-G modified version of it, and its even better than it was to start with.

Image quality is brilliant. Performance has gone up, and all around it is the best set to date that I have used on a Geforce 6/7.

Very Highly Recommended."

And although this was not written by the authors themselves (I presume), the person who posted it does have a post count high enough to influence a lot of noobs on that forum. And it's not like it is an unknown forum either.

And this happens almost everywhere, everyday. Guru3d has an entire section dedicated to modified drivers.

Besides that, unless the authors post a screenshot comparison of how great the IQ is compared to my vanilla nV drivers, I'm not able to see any remarkable difference whatsoever myself. Haven't seen that type of proof posted anywhere yet, maybe you could correct me.

Peoples-Agent
03-23-06, 09:40 AM
You do get alot of nOOBs and general dumbasses who don't follow the instructions and use Driver Cleaner properly...end up .infs and allsorts get a mish mash.

Any boost in IQ is a bonus, but if your an SLI user fluency of frame rates or the average frame rate is more important than simply high frame rates.

To put it into perspective, using the XG drivers I find I can use VSync to eliminate tearing but avoid the fps hit ..ala the VSync's function.

Colours look more vibrant with them, all they do is modify what's already within the drivers, but nVidia have to be quite conservative with how it's done for compatiability and blasted WHQL reasons.

Peoples-Agent
03-23-06, 09:44 AM
Besides that, unless the authors post a screenshot comparison of how great the IQ is compared to my vanilla nV drivers, I'm not able to see any remarkable difference whatsoever myself. Haven't seen that type of proof posted anywhere yet, maybe you could correct me.

You obviously haven't been using them long, that's plain to see.

I always install the Vanilla first, use them for awhile.... and using a newer method I apply the tweaks afterwards, reboot and voila. You can switch between the two easy.

You don't get a huge difference in every single set of drivers, some more than others.... but in my vast studies between the two, what I have generally found is you get the IQ of the High Quality setting using the regular Quality mode, with Q mode performance. That's where the bonus comes in.

SLippe
03-23-06, 07:08 PM
I haven't used Omega drivers since I had my Ti4400 and driver version 45.23 or some sh*t. I used to think they made a difference back then, but I don't like using 3rd party drivers anymore.

gstanford
03-23-06, 07:40 PM
I like XG's modded drivers, they have great IQ compared to stock drivers.

Performance wise, what I look for is not necessarily fast frame rates but "smooth" frame rates. It's quite possible to have drivers with very fast frame rates that also jerk like crazy and I find that quite off-putting.

RanCorX2
03-24-06, 12:57 PM
Well if you are complaining about the XG drivers not adding little if any performance the point of them is to increase IQ with no performance drop and sometimes they do have a slight benchmark advantage so you are barking up the wrong tree.

if thats the only advantage of modified drivers, only benchmark freaks should use them then ;) for us gamers they are useless and just crash our games.

also i'd like to add imo the coolbits extras are pretty useless unless you have a card that supports the various cpl options, maybe the modified drivers should have the option to add/remove only the registry keys you need, otherwise you're just filling up your registry with junk, true a small amount but still, it'll make things easier.

bigjohns97
03-24-06, 01:08 PM
I never noticed any diffrences but i'm sure they aren't moding the driver code but really just setting driver options in the registry.

LOL i can do that.

Q
03-24-06, 02:00 PM
There was only one time in history that I can think of that modified drivers ever did crap, and that was when 3dfx died. The mod community pieced together drivers and INIs to get decent XP support and there was even some work on creating an independent driver. Of course, it was all for naught really, but the modded drivers did at least mean something then.

I also remember some of the early, nasty ATI drivers that benefited greatly from some INI tweaks. Other than that....what's the use really? Adjust the LOD using RivaTuner if you're that anal.

Peoples-Agent
03-24-06, 05:45 PM
They are just modifications in the .inf to the registry, correct.

nVidia drivers as Vanilla are far from optimised due to this WHQL crap.

Why it's more often than not you tend to get better performance from the Betas over the final releases.

Seraphim
03-25-06, 01:28 AM
I tried out some Omega(I think) drivers for my old Radeon 8500 back in early 2002. Supposedly they increased AF quality and had this whole long list of improvements.

The results? I noticed my performance wasn't what I thought it should be, then I ran 3DMark and I was getting like 5500 marks when at the time Radeon 8500s were getting like 8000-10000.

Also the Ego on the Omega guy is nuts, saying he was going to get out of making modified drivers simply because other people dared to release modified drivers as well.

Sorry man, didn't realize people on the internet needed to ask your permission to do things. :screwy:

Darkfalz
03-25-06, 01:31 AM
Correct, just registry tweaks and things. It's not like they have access to the source code to fix any bugs. With the 3dfx stuff it's different, they are writing new cores to work on XP and fix bugs. Quite interesting stuff, if you still have your Voodoo5 lying around (like me).

Total waste of time.

Seraphim
03-25-06, 01:45 AM
Correct, just registry tweaks and things. It's not like they have access to the source code to fix any bugs. With the 3dfx stuff it's different, they are writing new cores to work on XP and fix bugs. Quite interesting stuff, if you still have your Voodoo5 lying around (like me).

Total waste of time.

LOL its funny you mention that because I was thinking about 3dfx drivers but I didn't bring it up.

There have been some good driver releases but the 3dfx driver modding community (from what I have heard) but many ridiculous ones as well.

On another forum I post on someone came out with these "new" drivers they had been "tweaking for months". Turns out they were just a mix and match of components from the official drivers with a few hex edits and ini tweaks. Gains of like 1% and no bug fixes or anything.

I really wonder about the 3dfx "community" at this point. Forums that were pretty vibrant up till about 2002 are now ghosttowns besides the off-topic sections. I can't understand why anyone would still use one of those cards when most integrated gfx card these days is probably better.

Darkfalz
03-25-06, 03:44 AM
Many of the early 3dfx driver packs were just mix and match, ie. complete rubbish. None of the new drivers are really any better on a 9x system, but for an XP system yeah. Although MesaFX is slow compared to 3dfx OGL, it certainly has more features support. On a high CPU (2 GHz minimum) it can run newish games okay. Frankly though, the Voodoo5 is so long on the tooth that it is a bit silly. I keep it in my P3 system just for Q3A/UT and Glide games and emulators.

j0j081
03-25-06, 12:22 PM
if thats the only advantage of modified drivers, only benchmark freaks should use them then ;) for us gamers they are useless and just crash our games.

that doesn't make any sense because I said they increase IQ, not performance and I've never had a crash from them not that I use them all the time though.

Peoples-Agent
03-27-06, 08:44 PM
They don't fundementally change the driver structure, so them crashing your games is down to the drivers themselves, not the tweaks .....I'd most likely say "User Problems".