I think I might see the problem here.Then i guess i miss your point, unless you are trying to say that AMD is intentionally gimping compute performance like Nvidia did for the titan compute drivers?
If AMD could flip a switch and unleash this performance like Nvidia did why would they have not done so for launch?
Nvidia could have done this at any time, they just had to flip a switch in the drivers, i dont see how this compares to the current vega situation unless AMD is truly trolling all of us.
Lol, this is painful to watch, So now Vega is good only at low settings, 1080p and when NVIDIA IS CPU limited in one instance! But when Vega is destroyed in all other benchmarks it has to be a bottleneck! The level of contradictions in your post is alarming to the amount of mental gymnastics and straw grasping you had to do.
I think I might see the problem here.
When I and others say an instant performance increase with a driver release, we do not mean that its already here. We mean that the hardware is there but not enabled because AMD has not completed the software. When its finished and the enabling driver released, the result should be 'like flipping a switch' in your words.
I realize this, but the other poster compared it to the Nvidia titan compute driver release, that was just like flipping a switch as the titan had always had that performance but it was being artificially disabled in drivers , which was why i was confused as to how that related at all to the vega situation, unless he was trying to say vega was in the same boat as the titan and was just waiting for a switch to be flipped.
infinity fabric and many of the new features some of which still need software enablement explain most of the (Vega performance) delta
They're still building the switch, and nobody know if it will work, or how well.
AMD fans are trying hard to prove a point they have no evidence or citation for, they use edge and abnormal cases, and even bad examples, they are willing to twist a case of simple pro drivers installation on a Titan or running low settings with CPU limit to prove a point so hopelessly vague and fantastical, AMD themselves never promised or even hinted at. AMD was very clear on where Vega 64 stands, they gave not even a single hint about it's competitiveness against a higher tier card such as 1080Ti. They barely even compete with the normal 1080 with liquid cooling let alone the 1080Ti. AMD was even very clear in being cautious with the whole shanenigans of Primitive Shaders and DSRB.. Etc, yet people still ignore all of these signs, still chose to hype stuff up, still chose not to learn lessons from before and after Vega launch, they saw the signs and leaks of Vega being underwhelming yet they ignored it and continued their hype.
This proves people are willing to hype themselves to the moon and beyond, with little to no incentive. Just their blind bias. I now fully believe AMD had only very small part in hyping up Vega, people had the biggest role in this mess. They are the ones responsible for the whole disappointment and frustration.
How about a discussion about that actual disabled features and what they will bring performance wise?
How about a discussion about that actual disabled features and what they will bring performance wise?
It's a far better option than some posters "you're all idiots" and "AMD sucks" rhetoric. I'd much rather read about the possibility of features and what they could potentially bring as opposed to the "AMD bad, people stupid" talk we've had since forever."No one knows"
End discussion.
Well, with respect to primitive shaders that question amounts to asking what would be the performance implications for Vega of a significant increase in polygon throughput? If primitive shaders manage to achieve even half of the gains claimed in the whitepaper, you are talking about doubling Vega's culled polygon throughput. What are the implications for game performance of a new reality where Vega outperforms Pascal in tessellation?How about a discussion about that actual disabled features and what they will bring performance wise?
Enabled is one thing and performance enhancer is another thing entirely. I posted this like a 100 times. AMD told Anandtech to expect very limited performance uplift from DSBR except in resourse starved GPUs, they told Anandtech that primitive shaders are very hard to code and optimize for, that it requires going through inefficient paths in the driver, meaning potential negative repercussions. AMD is very very cautious when talking about these features. Which is the case when you don't expect significant changes to the current situation with said features. AMD is not even promising a single 1% of performance, yet people are ignoring all of that for the sake of a hype that will come vehemently crashing down like the ones before it.Yep, it's been confirmed from many sources that new uarch features are not yet enabled in current public drivers
I guess that we can take this as confirmation from AMD that most new features of Vega are not functioning in release drivers as many here have been saying.I'd lean towards the software not being finished yet because it doesn't make a lot of sense to disable that functionality when it makes your product look bad compared to the competition. Earlier today Raja tweeted the following:
"infinity fabric and many of the new features some of which still need software enablement explain most of the (Vega performance) delta"
That was in reply to someone asking about why there seemed to be a per mm^2 performance degradation between Vega and previous (presumably Fury) Radeon architectures. That doesn't really tell us what is meant by "enablement" as you could argue it means flipping some switch just as much as actually writing the software.
I suppose its good news in that Vega can expect some performance bumps (the exact amount is anyone's guess) but it still means that we're waiting for Vega.
This is the issue, all the disabled features are pretty much all new, so no one knows how much if any performance they will add.
This is the root issue with this whole discussion. Some people are expecting as much as 50% performance, some are thinking it will be minimal. Only time will tell.
No one has any proof of how much performance these features will/could/may add. Only option at this time, wait few weeks and see who was right.
Well, I can think of two games off the top of my head where Vega relatively underperforms which are known to use lots of tessellation: GTA V and Watch Dogs 2. Gameworks games in general have often been accused of deliberately using large amounts of tessellation to tank performance on AMD cards. It's a good question to ask which games currently used frequently to benchmark GPUs use lots of tessellation?I don't think many games nowadays are constrained by tessellation performance, are they?
Strange thing is that we're talking mainly about the primitive shaders and features enabled by activating it. NGG fast path, IWD and improving occupancy of shaders in CU, a noted deficiency of GCN.Enabled is one thing and performance enhancer is another thing entirely. I posted this like a 100 times. AMD told Anandtech to expect very limited performance uplift from DSBR except in resourse starved GPUs, they told Anandtech that primitive shaders are very hard to code and optimize for, that it requires going through inefficient paths in the driver, meaning potential negative repercussions. AMD is very very cautious when talking about these features. Which is the case when you don't expect significant changes to the current situation with said features. AMD is not even promising a single 1% of performance, yet people are ignoring all of that for the sake of a hype that will come vehemently crashing down like the ones before it.
One than has to wonder why the features are not enabled or why the cards were not held back for a quarter to get them enabled.I guess that we can take this as confirmation from AMD that most new features of Vega are not functioning in release drivers as many here have been saying.
That above statement was to the posters misleadingly claiming otherwise.
AMD fans are trying hard to prove a point they have no evidence or citation for, they use edge and abnormal cases, and even bad examples, they are willing to twist a case of simple pro drivers installation on a Titan or running low settings with CPU limit to prove a point so hopelessly vague and fantastical, AMD themselves never promised or even hinted at. AMD was very clear on where Vega 64 stands, they gave not even a single hint about it's competitiveness against a higher tier card such as 1080Ti. They barely even compete with the normal 1080 with liquid cooling let alone the 1080Ti. AMD was even very clear in being cautious with the whole shanenigans of Primitive Shaders and DSRB.. Etc, yet people still ignore all of these signs, still chose to hype stuff up, still chose not to learn lessons from before and after Vega launch, they saw the signs and leaks of Vega being underwhelming yet they ignored it and continued their hype.
This proves people are willing to hype themselves to the moon and beyond, with little to no incentive. Just their blind bias. I now fully believe AMD had only very small part in hyping up Vega, people had the biggest role in this mess. They are the ones responsible for the whole disappointment and frustration.
Strange thing is that we're talking mainly about the primitive shaders and features enabled by activating it. NGG fast path, IWD and improving occupancy of shaders in CU, a noted deficiency of GCN.
In addition to the above, you are outright lying in the bolded portion. There, I've said it openly.
Vega whitepaper quote:
"In the case of “Vega” 10, we observed up to 10% higher
frame rates and memory bandwidth reductions of up to
33% when the DSBR is enabled for existing game
applications, with no increase in power consumption"
Now you can argue if you believe them or otherwise, but do not make claims as you are doing.
Well, I can think of two games off the top of my head where Vega relatively underperforms which are known to use lots of tessellation: GTA V and Watch Dogs 2. Gameworks games in general have often been accused of deliberately using large amounts of tessellation to tank performance on AMD cards. It's a good question to ask which games currently used frequently to benchmark GPUs use lots of tessellation?