Frustrated AMD Owner

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

thilanliyan

Lifer
Jun 21, 2005
12,040
2,255
126
Skyrim is just borken in CF...the problem is Skyrim, don't blame the cards.

If CF is broken in Skyrim (and SLI works fine), AMD should fix the issue pronto as it is a very popular game. Don't blame the cards...just blame the driver team I guess.

I'm so glad I am not using a dual GPU setup.
 

DrBoss

Senior member
Feb 23, 2011
415
1
81
Poor crossfire drivers and micro-stuttering are the reasons i got ride of my two MSI R6950 Twin Frozr III's. It was always annoying when new games were released and i have to wait around a week for new crossfire profiles before the gpu's ran their best.

I ended up purchasing an MSI N670 power edition and love it. I will stick with single GPU solutions from here on out.
 

sze5003

Lifer
Aug 18, 2012
14,304
675
126
Sell both cards and get the 7970 GHz, an after market one not reference cooled and you can overclock it. Newer beta drivers for games from AMD will be out soon, ex black ops 2 and assassins creed/ other performance increases.
 

Siberian

Senior member
Jul 10, 2012
258
0
0
i've never really had any problems in skyrim with a 9800 GTX+, GTX 560ti, or the GTX 680. But if I turn V-sync off I get this horrible flicker in areas that have water.
Are you using Adaptive Vsync? I really like that feature.
 

blackened23

Diamond Member
Jul 26, 2011
8,548
2
0
Over the past year, I have become increasingly frustrated with AMD due to their driver support. I'm running two 6950's, and am at the point where i just want to get rid of them and change over to Nvidia.

My question is, is it worth making the switch to nvidia right now or would it make more sense to wait for the next gen of their cards to come out?

My motherboard does not support SLI, so i am just looking for a solid single gpu card. Would something like a 670 be an upgrade or not?

Mods kill skyrim performance, and you can't run skyrim with vsync off. The physics engine is tied to the vsync.

Skyrim may run better on nvidia (shrug) but I can say 100% that no matter which direction you go, there are mods that will destroy skyrim performance. A lot of the mods really aren't bothering with. My rig (below) destroys skyrim in terms of performance yet I know there are mods that will completely offset the framerate - that is something to keep in mind.
 
Last edited:

badb0y

Diamond Member
Feb 22, 2010
4,015
30
91
nVidia has better multi-GPU drivers from what I gather from around the web but when it comes to single card setups the driver quality is pretty much the same. I would suggest selling the 6950s and grabbing a 7970 or a GTX 680.
 

Dark Shroud

Golden Member
Mar 26, 2010
1,576
1
0
nVidia has better multi-GPU drivers from what I gather from around the web.

When you take time to read the Nvidia forums you'll fine this isn't true.

Nvidia works with a fair amount of devs so they'll have SLI profiles out for those games right away. There are games that Nvidia will leave broken for awhile.

AMD recently stepped up their Gaming Evolved and they're drivers big jumps in performance show this.

The OP's problem is he's running mods in Skyrim. Skyrim uses a cobbled together engine from a company known for having lower tier programming.

Selling those cards to Bitcoin miners and buying a 7970 OC or Ghz Ed would solve his problem.

Adding in the card's wide memory buss and big chuck of Vram will help with the modding.
 

BoFox

Senior member
May 10, 2008
689
0
0
Until microstuttering is a thing of the past, just stick with a single GPU. It's usually better to buy only one GPU (slightly more powerful) and with the $$$ saved from not buying a second one, use it towards buying a better GPU 1-2 generations later.

skyrim-99th.gif

http://techreport.com/review/23150/amd-radeon-hd-7970-ghz-edition/7

17.1 ms is pretty darn close to 16.7ms for a perfectly fluid 60fps gameplay with 99+% of all the frames.
 

exia989

Member
Sep 7, 2011
28
0
0
I've come to the realization that my crossfire setup is not worth ditching for one game being dumb even though skyrim is one of my favorite games. that said, I'm not going to run dual gpus in the future when its time for me the upgrade


Posted from Anandtech.com App for Android
 

Aikouka

Lifer
Nov 27, 2001
30,383
912
126
One of the larger reasons why I dropped AMD was because I use Splashtop a lot, and there's a bug in AMD's drivers that will eventually force Windows down to a basic theme when using Splashtop to access your desktop. It's not really bad for your system -- at least that I know of -- but I just got tired of fixing it. I was in the market for a new graphics card, and I bought a GTX 680 instead of a 7970. No more dropping to a basic theme on my main desktop, but my server (5450) and second desktop (6950) are another story. :p
 

taltamir

Lifer
Mar 21, 2004
13,576
6
76
Microstuttering is a dual GPU issue. I've heard that it tends to be worse on dual AMD setups.

This is incorrect.
Microstuttering has a plethora of causes just like regular stuttering has a plethora of causes.

Dual GPU is indeed a cause of micro stuttering, but it is one of many causes and it could be that his issue is not related to dual GPU.

Personally I don't think Dual GPU is worth it though, too many issues. Better get a higher quality single GPU.
 

BrightCandle

Diamond Member
Mar 15, 2007
4,762
0
76
I've come to the realization that my crossfire setup is not worth ditching for one game being dumb even though skyrim is one of my favorite games. that said, I'm not going to run dual gpus in the future when its time for me the upgrade

Just bare in mind that is throwing the baby out with the bathwater. AMDs crossfire has a severe issue of stuttering where NVidia doesn't. That isn't to say SLI is perfect but it is vastly superior to CF at this time. It wasn't always true, the 5000 series crossfired well, but then they changed the drivers a few times and the stuttering started.

Now bare in mind I have also compare my 7970's against my 680's in single card setups and the 680 is a little smoother. That is my eyes are fooled by motion down to a lower FPS than on the 7970's. The techreport data doesn't show that out in data and I also couldn't capture it with Fraps frame times so I suspect its not part of the cycle that fraps measures but I keep coming to the same conclusion that 30 fps is smooth on the 680 but I need 40 to be happy on AMD.

I understand your frustration, I was there until I switched to NVidia's solution. Overall NVidia's drivers have been much better, no stupid problems with power saving, no negative scaling in SLI (so far) and well I don't have any problems whereas I had about 25 bugs raised with AMD. So NVidia has been dramatically better for me. But you pay more for similar performance, but most review sites measure the performance on FPS instead of frame times so aren't really worth the pixels they consume.
 

Creig

Diamond Member
Oct 9, 1999
5,170
13
81
So you're claiming that you can't directly compare FPS between single AMD cards and single Nvidia cards because you need higher framerates on the AMD card than the Nvidia card? I don't think I've ever heard anyone say that before.

And zero bugs for SLI while experiencing 25 bugs for Crossfire also seems a bit hard to believe.
 

BoFox

Senior member
May 10, 2008
689
0
0
Just bare in mind that is throwing the baby out with the bathwater. AMDs crossfire has a severe issue of stuttering where NVidia doesn't. That isn't to say SLI is perfect but it is vastly superior to CF at this time. It wasn't always true, the 5000 series crossfired well, but then they changed the drivers a few times and the stuttering started.

Now bare in mind I have also compare my 7970's against my 680's in single card setups and the 680 is a little smoother. That is my eyes are fooled by motion down to a lower FPS than on the 7970's. The techreport data doesn't show that out in data and I also couldn't capture it with Fraps frame times so I suspect its not part of the cycle that fraps measures but I keep coming to the same conclusion that 30 fps is smooth on the 680 but I need 40 to be happy on AMD.

I understand your frustration, I was there until I switched to NVidia's solution. Overall NVidia's drivers have been much better, no stupid problems with power saving, no negative scaling in SLI (so far) and well I don't have any problems whereas I had about 25 bugs raised with AMD. So NVidia has been dramatically better for me. But you pay more for similar performance, but most review sites measure the performance on FPS instead of frame times so aren't really worth the pixels they consume.

Interesting, thanks for your input. Microstuttering just needs to go already.. the last SLI setup I had was with 7900GTX'es - I had 2x 8800GTX and 2x GTX 460's but I used them in separate rigs, refusing to pair them together.
 

taltamir

Lifer
Mar 21, 2004
13,576
6
76
So you're claiming that you can't directly compare FPS between single AMD cards and single Nvidia cards because you need higher framerates on the AMD card than the Nvidia card? I don't think I've ever heard anyone say that before.

Its actually entirely possible if the average FPS is not indicative of the min FPS. But I would not say its an nVidia vs AMD thing but a card A vs card B thing (which also varies by game). It could very well be that in that particular game that particular nvidia card has better min FPS than that particular AMD card. Heck, it could even be in most games for those two specific cards (I didn't check)

BrightCandle, based on what you said I suggest in the future you focus on reviews that compare min FPS rather then averaged FPS.
 

moonbogg

Lifer
Jan 8, 2011
10,731
3,440
136
Well I'm pretty sure my xfire 6950 is faster than Your 7970 in majority of cases!

Yeah it probably is, but do dual GPU setups ever get you a little depressed? I mean its like this. I buy two beautiful GTX570's and I can feel the power emanating from the case. My mind is totally at ease knowing that I have the fastest possible setup within reason, meaning buying two 580's would have been not worth the cost to me.
I have confidence knowing that no game could challenge my rig. Instead, game after game buckled at the knees and fell to worship my god-like rig for months and months. Then it happened. BF3 came out and I couldn't max it due to a lack of Vram. I was crushed. My self esteem plummeted and my heart began to fracture right down the center, and out of it poured my love and adoration for my eleet dual card setup.
I took notice of the newer, cheaper single card solutions on the market and how they ran BF3 better on max settings, despite lacking the same raw GPU power of my 570's which had suddenly fell from grace. I looked at my bank account and came to realize that maybe, just maybe I could have been alright with a single high end card until the next gen came around. But no, I had to splurge on the dual card setup to satisfy my unquenchable thirst for overkill computing power, and now I was paying the price in terms of lost money and a frayed ego.
There was only one solution for me. I had to have two GTX 670's to make up for the mistake I had just made. I had lacked foresight when purchasing the two 570's, but things would be different this time. Nvidia had learned their lessons about not including enough Vram on their upper end cards, and I had learned my lesson too, or so I thought.
Now I sit here, pondering my addiction and consequent actions as they continue to spin their sticky, inescapable web of deceit and prideful trickery; a trap for those like me. Maybe next gen I will learn. Maybe I can start anew and do things right this time. Maybe, just maybe I can settle for a single GTX780...that is unless Crysis 3 doesn't pull 60fps. Or maybe Far Cry 3 has dips in the 50's....what then? What then are my options? Perhaps two 770's. That should do it. I'll probably be OK with dual 770's next time, because next time will be different.
 
Last edited:

BrightCandle

Diamond Member
Mar 15, 2007
4,762
0
76
It isn't the minimum frame rates, its inconsistent frame times, aka microstutter. You can see some difference by just running fraps in frame time mode. But I also know what that frame time is measuring and it isn't able to record when the GPU delivered a frame, only when DX was done with it. What we really need is the ability to know when a frame gets delivered to the screen, and there is no software that can do it yet.

I perceive the stutter as loss of appearance of motion and anything above about a 10% swing breaks motion for me. So I know its different between the two brands and I much prefer Nvidia, but I don't like not being able to show a graph with the difference.
 
Jul 29, 2012
100
0
0
For stutter in Skyrim run it in borderless windowed mode. This also improves Sleeping Dogs stuttering, and apparently for many other PC games. Search online for the mods

Crossfire has worse micro stuttering vs SLI, but the RadeonPro tool is said to do away with that. It is on a thread on this very forum. AMD really does need to hire some drivers teams if a hobbyist has surpassed them, though their personnel are stretched pretty thin. If 12.11 had come out earlier than almost a year after the 7000 series they would have been far more competitive.
 

amenx

Diamond Member
Dec 17, 2004
4,408
2,729
136
Minimum fps can be misleading. It could be a very small dip in a very brief part of the game where its almost unnoticeable. Average fps, although preferable to me, is also not an entirely reliable measure if the minimums are just enough to cause some stutter in some places.
 

BrightCandle

Diamond Member
Mar 15, 2007
4,762
0
76
For stutter in Skyrim run it in borderless windowed mode. This also improves Sleeping Dogs stuttering, and apparently for many other PC games. Search online for the mods

Running windowed disables crossfire....which is likely why it stops the stuttering.
 

taltamir

Lifer
Mar 21, 2004
13,576
6
76
Minimum fps can be misleading. It could be a very small dip in a very brief part of the game where its almost unnoticeable. Average fps, although preferable to me, is also not an entirely reliable measure if the minimums are just enough to cause some stutter in some places.

Minimum FPS is only misleading in such a manner if you look at a single number rather then a chart plotting min FPS over time