QA Consultants: AMD's drivers more stable than nVidia's

[BFG10K]

Tested using Microsoft's Hardware Lab Kit. No games, but I'm still pleasantly surprised by the result. We need AMD to successfully compete like this wherever possible.

 

[DXDiag]

This is hardly convincing considering AMD PAID them to do it, and they used wrong drivers for Quadros. Which is why most of the crashes and problems popped up on Quadros then Firepros. Geforce vs Radeon have comparable results.

Ryan Smith:

I dug into the report a bit between meetings. Overall the methodology is sound, but I find the OS and driver selection odd. Win10 1803 was a bug-riddled mess, and they didn't use NVIDIA's official 1803 drivers for Quadro (R396), which were out at the time.

 

[amenx]

AMD provided all the sample cards in the test and only conducted on one driver from each side.

http://www.legitreviews.com/amd-cla...but-supplied-all-graphics-cards-tested_206705

Jeezuz... if this had been Nvidia who commissioned a similar test under same conditions, there would be a feeding frenzy and probably an [H] article on the scandalous methodology used.

 

[LTC8K6]

They are learning from Intel...

 

[CPUGuy]

Wut? LOL
Companies do this all the time in many different segments of the market. I see nothing wrong with this. It's a common practice. Sometimes the information is made public sometimes it's not. You don't pay an organization like this to create media buzz. You get the Media to create media buzz about your products. You have the idea of Media buzz backwards when comparing it between the media and hiring consultants.

At a guess; this is showing results that AMD already knew. So they hired an independent to conduct this to see if they, independently, come up with this anomaly.

 

[amenx]

Wut? LOL
At a guess; this is showing results that AMD already knew. So they hired an independent to conduct this to see if they, independently, come up with this anomaly.

What conclusions are we to take from the results of one driver tested? That all AMD drivers are more stable than Nvidias throughout the past and future?
Seems they also conducted similar driver audits in 2005 and 2010. What happened in all the other years? No results worthy of publishing? Or no tests at all?

https://radeon.com/worldsmoststabledriver/

 

[CPUGuy]

What conclusions are we to take from the results of one driver tested?

I suggest you cliff "The Art of War". It's all business!
Did you know that AMD's stock price is going up? It's all fresh news too.Coincidence or relative? I don't care.
https://www.investopedia.com/news/amd-jump-over-25-pc-business-pickup-stifel/

 

[VirtualLarry]

Is this the start of the GPU version of the "Cola Wars"?

 

[Batboy88]

I Doubt it...They use to have some of the redacted got so bad some Drivers wouldn't even Install....

Then went back to Team Green after a good while, wen't through I don't know how many drivers and flawless not one issue....




No profanity allowed in the tech forums.


esquared
Anandtech Forum Director

 

[lixlax]

Paid by AMD or not, but the "AMD/ATI drivers are bad" argument have been in the BS category for a long time already. Both companies have their problems and bugs pop up and as far as I know these are getting usually fixed sooner rather than later.

 

[tamz_msc]

To be honest, the fact that you can fail a fundamental WDDM conformance test and that too on professional GPUs with higher error rates is incriminating of both companies.

Having said that, from what I've gathered reading online as well as from personal experience, these days driver problems on NVIDIA basically boil down to nvdkkm.sys going kaput at some point - this means crashes, TDRs, sometimes even BSODs. On AMD's side, driver problems usually mean either stuttering or frame rates plummeting to 1-2 FPS at some completely random time, very rarely a hard crash.

Also, NVIDIA has huge issues with recent Windows feature updates which updated the WDDM, you'll find page after page on the NVIDIA forums where people have reported issues which are ultimately tied to something Microsoft did on their end which broke the drivers and NVIDIA not being able to do anything about it. That is probably why their drivers fail more in this test.

 

[Dribble]

What's more surprising is AMD fails any test. If they commission a test of a known thing (microsoft hardware lab kit) with known cards and drivers (supplied by AMD) they should have been able to pass it with 100% success as they should have fixed every bug before the test happened.

 

[lixlax]

What's more surprising is AMD fails any test. If they commission a test of a known thing (microsoft hardware lab kit) with known cards and drivers (supplied by AMD) they should have been able to pass it with 100% success as they should have fixed every bug before the test happened.

Maybe they actually intended to be objective? I wish the independent rewievers to have more self-initiative on such matters as well. At the moment there is still that fact that AMD paid for this research which can put the results under suspicion.

 

[CPUGuy]

They don't appear to be related at all with AMD nor Nvidia.
https://qaconsultants.com/

Interesting they have the clearance to work for the Air Force. This seem more creditable to me.

 

[EXCellR8]

Personally, I've had good and bad luck on either side... for as long as I can remember. Too many factors to be considered in coming up with a concrete answer.

 

[amenx]

Thing is, even if these tests were truly independent (not paid by AMD or Nvidia), what sort of criteria do you go by? One driver set from each company tested on random GPUs? The results would still only pertain to THAT driver tested. The next set of drivers may be different. Drivers have a short life, typically one month before a new one is released (which often addresses issues relating to previous drivers). To get a good idea of who has the more stable drivers, it may be best done over a lengthy period, where all drivers for each company released within a one year period are tested.

 

[DaveSimmons]

AMD hired QA Consultants, picked the cards, picked the drivers. So clearly a chance for bias there.

I've had more problems with ATI/AMD cards than nvidia in the past, but I've just considered nvidia "better" rather than ATI/AMD drivers being "garbage."

I'm not convinced by this that ATI has reached "better" status but it's some (possibly biased) evidence that their drivers are now comparable to nvidia's.

 

[CPUGuy]

I have to say that the way something is presented shows me that you have interest in it. Although these results can be corrected I find it deplorable that the control panel hasn't seen an update since Windows XP days. Sure, you can tell me otherwise but when I don't see a simple GUI update I simply don't have confidence to believe that the drivers are any better than GUI used for them.

 

[Shmee]

I have to say that the way something is presented shows me that you have interest in it. Although these results can be corrected I find it deplorable that the control panel hasn't seen an update since Windows XP days. Sure, you can tell me otherwise but when I don't see a simple GUI update I simply don't have confidence to believe that the drivers are any better than GUI used for them.

Are you talking about the Nvidia control panel that you get when right clicking on the desktop? If so I see what you mean about it looking the same for a long time. That said, geforce experience has come, and itself changed in look as well. As for AMD, their look certainly has changed over the years with the radeon crimson iterations looking completely different from the old catalyst interface.

 

[Shmee]

As for the actual stability of drivers, I have had no issues from either team outside occasional CFX issues, and I have never gotten the chance to test SLI.

 

[TeknoBug]

I've used both Nvidia and AMD/ATI in the past 20 some odd years after 3Dfx and the magical Voodoo series died out and I don't really remember the last time I've had issues with either the GPU or drivers on the AMD side. Nvidia on the other hand was a handful about 10 some odd years ago with missing artifacts in games, and then recently with the GPU at driver level (like a WHQL'd driver killing cards, lost one of them to one of those dreaded drivers).

I'm not a brand fanboy, I just use whatever GPU is both decently priced and better overall performing, today the two brands are fairly close to each other except for the Vega 56/64 gap from the 1070Ti/1080/1080Ti.

 

[Rifter]

Completely useless and likely biased. Get back to me when this is done by independent testers.

 

[JR88]

I don't think it's that bad now dude...get in there and make them and everything work.
They have my blessings they made something Finally that does surpass the 980ti/1060...


Second accounts under vacation are not allowed.


esquared
Anandtech Forum Director

 

[Organik]

That is a biased statement. Both AMD and nVidia make great drivers imo.

 

[GodisanAtheist]

I had a HD 7950 and upgraded to a 980TI and really I haven't noticed any major issues with stabily involving either for the years I've had either.

However AMD's Crimson driver package update (with the clean look and faster loading) and full incorporation of overclocking tools into their core driver package makes their driver option stand a bit above the NV alternative IMO.