http://www.overclock.net/t/1606224/various-futuremarks-time-spy-directx-12-benchmark-compromised-less-compute-parallelism-than-doom-aots-also
Futuremark exposed as nVidia shills
Time Spy IS NOT a benchmark worth of exemplifying the capabilities DX12 has to offer.
How do I know you're not also an Nvidia shill?
How do I know who to trust?
Everything on this earth needs to die except for the peaceful snip crab
>>55638381
that fucking moronic dev who got the short stick to manage the forums during the weekend showed quite clearly what they think of AMD at Futuremark.
I don't care you your a fanboy, but you reallyshould tone it down when you're commenting on something at a professional capacity.
>>55638381
reminder fps are not a valid benchmarking parameter either.
nvidia cheats with low res textures to boost fps
>>55638646
>Speak professionally
>Second word in retaliating statement is fucking
Go home anon , the trash bin is nice and warm.
>>55638689
go to sleep FMJarnis.
you did a lot already.
> I'm still learning how to read, and interpret this information but I figured I'd share some of the images with you guys and maybe get a better understanding of whats going on.
Oh good, people who don't know how to interpret information they are getting trying to make conclusions based off it.
>>55638816
It was a valid concern when everyone and their mother became electrical engineers privy to closely guarded spec sheets right after the RX480 launch.
That's a good thing. Lot's of e.e.'s graduated from last month are now learning how to code (or debug/reverse engineer) at a very high level.
I bet any amount of money that this will not matter in the long run. You will see a dozen forum threads about it and after that it will die down. The benchmark will be used for reviews and will be taken as what it is presented at: a DX12 benchmark that show the card's capabilities in said API.
>>55638804
Who is fmjarvis
deja vu.
reading those who are defending Nvidia there feel just like when they were defending the GTX 480 as the second coming of Christ... I wonder how xtremesystems is doing these days.
probably the very same folks at a different venue.
>>55639080
Either that or the next driver update from AMD will improve their performance in the benchmark.
>>55638381
So the rumor wasn't just a rumor. Hard evidence coming up pointing fingers at 3dmark/nvidia tampering with the meaning of "async compute" when in reality, this is simply nvidia pre-emption, aka a slightly better serial load than DX11.
>>55639116
It will get swept under the rug like many others debacles over the years. The thing that angers me is not that it favors a party or the other, because that is nothing new (ex: heavy tess).
I hate the fact that is a DX12 benchmark that does not use DX12 features. That is false advertisement but nobody cares. All they talk about is AMD vs nVidia.
>>55639116
I don't think they could do much when the engine itself is developed no not utilize the resources from one specific vendor.
Even if possible the returns from an ex post facto implementation would certainly introduce too much delay/overhead to be worthwhile.
>>55639210
If you think about it, the heavy tesselation was an even worst offender for showing bias. In that they were almost sabotaging one particular set of hardware, but now they are just not making the most out of tha (same) set of hardware.
Besides, they've got plenty of plausible deniability now. They can roll with the fact that there many diffeferent approaches to this new API and they simply chose one that would be perceived as being more crossvendor compatible or friendly. Yes, it's a BS euphemism, but it'll stick anyway.
>>55639225
>I don't think they could do much when the engine itself is developed no not utilize the resources from one specific vendor.
AMD is a partner in the Futuremark development program though.
https://www.futuremark.com/business/benchmark-development-program
It is entirely possible that AMD need to modify their drivers to get the best performance out their cards in the benchmark.
Both AMD and NVidia have to release new drivers for practically every new game as well. It's not really any different.
>>55639322
The whole getting close to the metal was supposed to avoid the "you need driver to play latest game" bullshit.
>>55639340
that will never happen because they have to fix the shitty work made by the game devs
>>55639322
that's the problem.
The engine works under a premise that favors nvidia hardware and software. If AMD is even capable of coming with a fix for that issue, it'll probably introduce some nasty overhead on top.
There's a chance that maybe they can mitigate its underutilization, if the balancing of a reshceduler atop the engine results in a better overall score in the end; but that would literally be sweeping the dirt under the rug, since the underlying problem would still be there.
AMD should start to sponsor Unigine*.
Then, there would be unifags and 3dmfags.
At least it would bring more diversity to fanboy wars.
They are partners already, but I meant to really sponsor them.
ahhahahahahahahahahaha
>>55640591
despicable nvidia mole.
>>55640753
>tfw he admitted that they're lazy and this benchmark should be considered irrelevant.
>>55640591
JUST DX11 MY SHIT UP F@M
Are there any open source GPU benchmarks?
>>55642200
glxspheres
