• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Async compute gets 30% increase in performance. Maxwell doesn't support async.

Iklos said:
Well, perhaps the developer did not want to add more debugging for a feature - when not getting paid.

So it is a possibility that there was no TressFX on Nvidia and everything else, due to them not wanting to spend the time? and money?

No proof that AMD did this to hinder nVidia, especially since they did not do it with tomb raider.. just sayin' as an SLI 980ti owner
Click to expand...

How can you interpret "as part of our partnership with AMD" as anything but AMD's decision?

When a NVIDIA partnership does it, it's evil and bad for gaming, but when an AMD partnership does it it's not a bad thing.

Would you like some salt? You seem to need some.
 
trick0502 said:
You sound like your on a gamework witch hunt. Oh I mean oxide witch hunt. Nv has the source code for the game, what more do they need?

Where is your proof? Source? Link?
Click to expand...

/sarcasm on

ATI/AMD has had the source code to the Linux kernel for 20 years and their drivers still suck!

/sarcasm off

Just because people/companies have source code and time with it doesn't mean things are automatically going to get better.
 
StormClaw said:
the code is open source, dumbass
Click to expand...


where is the source then? If it was open source we should all be able to see it right now right? We should be able to compile the engine and test the shaders right? Come one, I don't know where you get your information from but the nitrous engine is not open source.

Who is the the dumb ass? I suggest you go back to your closet and stay in there for a long time.

http://www.oxidegames.com/nitrous/
 
razor1 said:
where is the source then? If it was open source we should all be able to see it right now right? We should be able to compile the engine and test the shaders right? Come one, I don't know where you get your information from but the nitrous engine is not open source.

Who is the the dumb ass? I suggest you go back to your closet and stay in there for a long time.

http://www.oxidegames.com/nitrous/
Click to expand...

http://www.oxidegames.com/2015/08/16/the-birth-of-a-new-api/

Our code has been reviewed by Nvidia, Microsoft, AMD and Intel. It has passed the very thorough D3D12 validation system provided by Microsoft specifically designed to validate against incorrect usages. All IHVs have had access to our source code for over year, and we can confirm that both Nvidia and AMD compile our very latest changes on a daily basis and have been running our application in their labs for months.
Click to expand...

YOU = OWNED.
 
StormClaw said:
http://www.oxidegames.com/2015/08/16/the-birth-of-a-new-api/



YOU = OWNED.
Click to expand...


I suggest you read everything, they stated they will only put code in if it doesn't hurt current optimizations already done. So this is why that a shader that was previous created was hurting nV's hardware they decided to put in another path for nV, what that shader is doing we don't know, and they stated this on an open forum, so why can't they show the shader profiler information on a open forum, they went far enough to say it, now show it.

Don't be blind to what they did and try to sway this at something else.
 
leeleatherwood said:
razor1 ultimate fanboy level 9000 achieved.
Click to expand...


IS this the level that this forum has become? I think you guys need to go out and get some psychedelic drugs and clear you minds or actually just kill all your brain cells it would be better for the rest of us not to hear crap like this.

I even stated when this first started off, the benchmark that Oxide provided might be valid but without knowing what is going on its to early to say what it was and how it was going about.

http://hardforum.com/showpost.php?p=1041818052&postcount=29
 
razor1 said:
IS this the level that this forum has become? I think you guys need to go out and get some psychedelic drugs and clear you minds or actually just kill all your brain cells it would be better for the rest of us not to hear crap like this.

I even stated when this first started off, the benchmark that Oxide provided might be valid but without knowing what is going on its to early to say what it was and how it was going about.
Click to expand...

re-read your posts, you are a fanboy, a desperate one at that. You disregard facts and convolute the truth because of your clear bias towards nvidia.

Take a break, your fanboyism is embarrassing.
 
leeleatherwood said:
re-read your posts, you are a fanboy, a desperate one at that. You disregard facts and convolute the truth because of your clear bias towards nvidia.

Take a break, your fanboyism is embarrassing.
Click to expand...


I suggest trying talking about the topic at hand, instead of accusations, because that takes away from the conversation.

Dispute what I have stated with hard facts and go from there, you won't find any outside of she/he said and that isn't facts, I want data that shows it.
 
razor1 said:
I suggest you read everything, they stated they will only put code in if it doesn't hurt current optimizations already done. So this is why that a shader that was previous created was hurting nV's hardware they decided to put in another path for nV, what that shader is doing we don't know, and they stated this on an open forum, so why can't they show the shader profiler information on a open forum, they went far enough to say it, now show it.

Don't be blind to what they did and try to sway this at something else.
Click to expand...

why are you such a Novidia apologist? are they paying you to shill for them?
 
StormClaw said:
why are you such a Novidia apologist? are they paying you to shill for them?
Click to expand...

err? yeah ok, and what are you doing?

Open source I know how you got that idea, you googled oxide engine and it came up with github right? What you think google is a cure all for being an idiot lol.
 
razor1 said:
Dispute what I have stated with hard facts and go from there, you won't find any outside of she/he said and that isn't facts, I want data that shows it.
Click to expand...

FACT: Novidia had access to the game's code for over a year.

You claiming that AMD or Oxide is somehow gimping NV cards in the code on purpose is utterly moronic.
 
StormClaw said:
Novidia had access to the game's code for over a year. You claiming that AMD or Oxide is gimping NV cards with the code on purpose is utterly moronic.
Click to expand...


Read their comments in full man not just a piece here and there, if you want me to post the entire thing, I will! Its easy use google, you will find it, because you seem to like to use it.

This is what you are referring too correct?

http://www.oxidegames.com/2015/08/16/the-birth-of-a-new-api/


now lets go to here, link will follow shortly, this is the open forum post of an oxide dev.

http://www.overclock.net/t/1569897/...ingularity-dx12-benchmarks/1200#post_24356995

http://www.overclock.net/t/1569897/...ingularity-dx12-benchmarks/1210#post_24357053
AFAIK, Maxwell doesn't support Async Compute, at least not natively. We disabled it at the request of Nvidia, as it was much slower to try to use it then to not.

Weather or not Async Compute is better or not is subjective, but it definitely does buy some performance on AMD's hardware. Whether it is the right architectural decision for Maxwell, or is even relevant to it's scheduler is hard to say.
Click to expand...

Personally, I think one could just as easily make the claim that we were biased toward Nvidia as the only 'vendor' specific code is for Nvidia where we had to shutdown async compute. By vendor specific, I mean a case where we look at the Vendor ID and make changes to our rendering path. Curiously, their driver reported this feature was functional but attempting to use it was an unmitigated disaster in terms of performance and conformance so we shut it down on their hardware. As far as I know, Maxwell doesn't really have Async Compute so I don't know why their driver was trying to expose that. The only other thing that is different between them is that Nvidia does fall into Tier 2 class binding hardware instead of Tier 3 like AMD which requires a little bit more CPU overhead in D3D12, but I don't think it ended up being very significant. This isn't a vendor specific path, as it's responding to capabilities the driver reports.
Click to expand...
Why the hell does it seem to be working on the small program that B3D created then, if Oxide is saying

Curiously, their driver reported this feature was functional but attempting to use it was an unmitigated disaster in terms of performance and conformance so we shut it down on their hardware. As far as I know, Maxwell doesn't really have Async Compute so I don't know why their driver was trying to expose that
Click to expand...
How are they trying to expose the async on Maxwell is my question, and why are those results happening in the way they are, what ever those results are, we don't even know. Without those two questions answered it doesn't really have any merrit, because of the B3D app, prior to that app, it did had some weight.
 
Last edited:
trandoanhung1991 said:
How can you interpret "as part of our partnership with AMD" as anything but AMD's decision?

When a NVIDIA partnership does it, it's evil and bad for gaming, but when an AMD partnership does it it's not a bad thing.

Would you like some salt? You seem to need some.
Click to expand...

Well, cuz i don't see a conspiracy in everything, and yes its annoyingly stupid evil when nVidia does it, i just dont see why they would do it, but i can see why the developer would choose not to, the game taking into consideration, they made a still buggy (but fun imho) game on another studios engine - so i choose to interpret as this - we did not include the tressFX for nvidia because money, plain and simple - less things to debug.

and nope, no salt needed, my popcorn has plenty - now on to the discussion :) but thank you for asking :)
 
razor1 said:
Because Oxide did something in code that hurt nV's performance, what ever it is, without a shader profiler of the offending shader no one will know outside of Oxide's word for it, and they are not willing to share that.

Who ever has this game and has both an nV and AMD card PM me, I'll walk you through how to set up a shader profiler on your system, and lets see if we can figure it out? This won't be easy either unless the game has built in shader profiler and we can find the console codes to bring it up.
Click to expand...

He doesn't know what Async compute is or his fanboy level is above PR1ME.

latest
 
razor1 said:
I suggest you read everything, they stated they will only put code in if it doesn't hurt current optimizations already done. So this is why that a shader that was previous created was hurting nV's hardware they decided to put in another path for nV, what that shader is doing we don't know, and they stated this on an open forum, so why can't they show the shader profiler information on a open forum, they went far enough to say it, now show it.

Don't be blind to what they did and try to sway this at something else.
Click to expand...

go back to your nvidia sponsored green cave are you an ihv please show us your cpu and gpu its open source to hardware vendors aka AMD, INTEL and NVIDIA. GTFOH or STHU.
 
Getting past the unless noise I will post again.

Read their comments in full man not just a piece here and there, if you want me to post the entire thing, I will! Its easy use google, you will find it, because you seem to like to use it.

This is what you are referring too correct?

http://www.oxidegames.com/2015/08/16/the-birth-of-a-new-api/


now lets go to here, link will follow shortly, this is the open forum post of an oxide dev.

http://www.overclock.net/t/1569897/...ingularity-dx12-benchmarks/1200#post_24356995

http://www.overclock.net/t/1569897/...ingularity-dx12-benchmarks/1210#post_24357053


Why the hell does it seem to be working on the small program that B3D created then, if Oxide is saying

How are they trying to expose the async on Maxwell is my question, and why are those results happening in the way they are, what ever those results are, we don't even know. Without those two questions answered it doesn't really have any merrit, because of the B3D app, prior to that app, it did had some weight.
Click to expand...
Which Maxwell architecture is he talking about is another question that has to be asked, I or 2.

Its not about the end performance which is better GCN or Maxwell 2, its about why couldn't they get it to work on Maxwell 2 with acceptable performance even if its lower it shouldn't be lower than doing things in serial which is the shader that nV provided (guessing by what they have posted), and why can't they elaborate on it by showing us the profiler data? Its not going to show any source code they can give use psuedo code if they like for reference points to the profiler data.
 
razor1 said:
Why the hell does it seem to be working on the small program that B3D created then
Click to expand...

It doesn't. It times-out or crashes.

Go read the B3D thread directly, instead of getting your info from random news blogs.
 
StormClaw said:
It doesn't. It times-out or crashes.

Go read the B3D thread directly, instead of getting your info from random news blogs.
Click to expand...


go read it yourself, what crashes is a forced synchronized path, a path that is not doing async code, its telling the driver to push everything through one after the other.

The reason that it was put in was to see how the driver and architecture would react to serial code when the architecture and drivers have no say.
 
It doesn't seem to be working, they had a few mixed up graphs and someone with a Titan X SLi setup caused some confusion. Maxwell's results are additive, compute+graphics, which means everything is being handled serially. At this point I can't find any conclusive evidence that Maxwell supports async whatsoever, aside from a false indication by the driver itself. It "pretends" to do handle the data asynchronously but it doesn't actually do it.

Nvidia is intentionally staying quiet. If they tell the truth; refunds start happening. If they lie; they get busted. The best course of action for them is silence.
Gonna give Newegg a ring later. Maybe I can overwhelm the CSR with tech babble and get my money back. :)

RItaa39.png
 
TaintedSquirrel said:
It doesn't seem to be working, they had a few mixed up graphs and someone with a Titan X SLi setup caused some confusion. Maxwell's results are additive, compute+graphics, which means everything is being handled serially. At this point I can't find any conclusive evidence that Maxwell supports async whatsoever, aside from a false indication by the driver itself. It "pretends" to do handle the data asynchronously but it doesn't actually do it.

Nvidia is intentionally staying quiet. If they tell the truth; refunds start happening. If they lie; they get busted. The best course of action for them is silence.
Gonna give Newegg a ring later. Maybe I can overwhelm the CSR with tech babble and get my money back. :)

RItaa39.png
Click to expand...


The second program shows it more than the first program Tainted, it shows that when doing compute + graphics async its around 10 ms faster than doing them separately.

If you guys are looking at the single command path which is the final test in that program that is not async as explained above.

The fist test is compute, the second test is graphics, the third test is async compute+ graphics, and the single command path is the final test.
 
TaintedSquirrel said:
Two 980 Ti tests with the new bench, both additive.

https://forum.beyond3d.com/threads/dx12-performance-thread.57188/page-15#post-1869464
https://forum.beyond3d.com/threads/dx12-performance-thread.57188/page-15#post-1869480

The only results I've seen that show any amount of parallelism were the mixed up graphs and the Titan X SLi user.
Click to expand...


Its always lower, even though its close, its still lower, if it was serial it would be more considerably. And that was what that last test was for to see what happens when both are done in serial. If you want to see if the hardware is a capable of something you need to take out or view what ever the driver is doing to see what is going on.
 
I'm not going to believe anything one way or another until NVIDIA explains it themselves, which I expect they will do eventually now that this is blowing up in the tech media.

I don't think Oxide is "wrong", and I hope no one is faulting them for any of this since they were more than transparent about what they were doing and they stated very clearly the situation. Don't shoot the messenger.
 
I would expect Nvidia's legal department to find someway to force them to rescind their statement, not much unlike they did to futuremark with the cheating fiasco when they were busted twice for adding profiles to intentionally decrease visual fidelity to increase benchmark scores.
 
yeah I agree, I just want more information from them to see how they reached their point of view. If nV talks about it even better.
 
The Mac said:
I would expect Nvidia's legal department to find someway to force them to rescind their statement, not much unlike they did to futuremark with the cheating fiasco when they were busted twice for adding profiles to intentionally decrease visual fidelity to increase benchmark scores.
Click to expand...
IANAL but I don't see anything in Oxide's posts that would expose them to any liability.
 
i wouldnt have thought futuremark's was either, but they managed it.
 
Wow, seems like this benchmark is the only glimpse of hope for AMD product users.
Meanwhile, enjoying Mad Max and MGS V in full glory.
 
You must log in or register to reply here.
Back
Top