Thursday, January 3rd 2013
![AMD Radeon Graphics](https://tpucdn.com/images/news/amdradeon-v1739475473466.png)
AMD to Fix GCN Latency Issues with Driver Updates
Last month, an investigative report by The TechReport found out that despite being faster, AMD's Radeon HD 7950 graphics card isn't "smoother" than NVIDIA GeForce GTX 660 Ti, in that shows signs of higher frame-delivery latency, a theory proven by high-speed camera recordings. Over the holiday, AMD's David Baumann responded in discussions around the web talking about the issue, in which he put AMD's stand.
Apparently, AMD Catalyst drivers still have refinement left in working perfectly with GPUs based on the Graphics CoreNext (GCN) architecture. Baumann explained that GCN, and AMD's older Very Long Instruction Word (VLIW) architectures feature fundamentally different memory management, and drivers that make the most of it are still a work in progress. Baumann stated "Over the early part of the year you'll see a few driver updates help this across a variety of games."
He continued "Additionally, when we switched from the old VLIW architecture to the GCN core there was a significant updates to all parts of the driver was needed - although not really spoken about the entire memory management on GCN is different to prior GPU's and the initial software management for that was primarily driven by schedule and in the meantime we've been rewriting it again and we have discovered that the new version has also improved frame latency in a number of cases so we are accelerating the QA and implementation of that."
Source:
The TechReport
Apparently, AMD Catalyst drivers still have refinement left in working perfectly with GPUs based on the Graphics CoreNext (GCN) architecture. Baumann explained that GCN, and AMD's older Very Long Instruction Word (VLIW) architectures feature fundamentally different memory management, and drivers that make the most of it are still a work in progress. Baumann stated "Over the early part of the year you'll see a few driver updates help this across a variety of games."
He continued "Additionally, when we switched from the old VLIW architecture to the GCN core there was a significant updates to all parts of the driver was needed - although not really spoken about the entire memory management on GCN is different to prior GPU's and the initial software management for that was primarily driven by schedule and in the meantime we've been rewriting it again and we have discovered that the new version has also improved frame latency in a number of cases so we are accelerating the QA and implementation of that."
95 Comments on AMD to Fix GCN Latency Issues with Driver Updates
But i will happy if any performance boost come again.
GTX680 ass kicked for a while, with new driver NV again fall deeper and deeper...:laugh:
I remember when I compared my then current HD 4870 512MB with an old GTX 8800 I got used three years ago, how much smoother and more responsive the nvidia card generally felt even though performance was lower. I see that nothing's changed and I'm sticking to nvidia.
That's why, I think, this beta is so long in approval...they have to test literally EVERYTHING, since such a change can drastically affect how well things run.
If it's not...well...maybe it WAS them. Funny though, since Microstutter has been complained about for years now, and AMD has finally admitted to it?
That is literally what microstutter is, large variations in render times, which many said they could not detect, so AMD ignored it.
What Tech Report did very well was present this in a way that was very understandable, and with benchmarks. That deserves a pat on the back, for sure.
Since I am noticing large improvements since the 12.8s I'm sticking to AMD.
Was pretty close to going to nVidia though near the end of 2012 despite my last nVidia card was a POS.
Good job to the people who presented the issue and good job to ATi for stepping up. GCN looks like it's only gonna get much better in the future :rockout:
On TR`s site, Amd 6 series were smoother than the nvidia 5 series according to the benches they have, yet we had people saying the nvidia cards were smoother? So which is it? I think it’s all in the head, or people imagining stuff.:laugh:
Even linking to original article induced rage.
My last two nVIDIA cards were a couple of GTX470 (not in SLI). I believe that in most of the games that I've put them through, there was some amount of stutter if PhysX or other Cuda functions (JC 2 for instance) were ON. With 5850 + GT240 for PhysX it was smooth as butter, so I suppose it was a problem regarding resource allocation or something like that. BFBC 2 run with less stutter on my 5850 than on GTX470 although the last one gave higher FPS, especially with AA.
Other "weird" experiences were thanks to Ambient Occlusion, in some games with that option activated, it started to lag or stutter like it was a 20-25FPS or worse even though the actual FPS hit 33-35FPS or more, in others, there was no problem. A powerful card/GPU, in the exact same game, gave a better experience where others lag although the difference shouldn't be that high.
Right now with a 7950@1170/1600MHz I feel some lag or stutter in Sleeping Dogs with Extreme AA, The Witcher 2 with US, Skyrim at times, BF 3 under certain settings/maps (6970 was unplayable on Ultra despite it's 40-45FPS), GTA 4, MP 3 with MSAA, Far Cry 3 under 60FPS, Hitman Absolution with high AA levels etc. I think the problem hangs on AMD/nVIDIA heads but also on the game developers hands and how well they optimize the code. It's not that big of a tragedy as long as the stutter can be avoided by turning of a graphic or other kind of option.
BTW, nVIDIA also had a "stutter problem" with gtx670 and adaptive v-sync I think.
Thankfully AMD is working on a solution for it.
Of course the 7950 had bigger latencies, it was stuck at 850/925 MHz, while the 660 Ti can boost to almost 1100 MHz. The card with higher fps will have the lower latencies...
How can the fps count be higher if the frames comes delayed to the screen? OC Both cards and measure the timings again.
Disappeared in 12.11 BETA 11.