AMD High Overhead Drivers on DX11

[ObscureAngelPT]

you don't get it it runs like crap on AMD because the AMD cpu is woefully inadequate no optimization on earth is gonna change that not even mantle

really? Then why I get 24fps on mantle and 7fps on dx11. That why I get 60fps rocking stable with mantle and I get 60 with drops to 40 with dx11..

Go away noob .. Please how your wife have patience.. Jesus 'k

 

[Ferrum Master]

i5 faster than same clocked i7 screw those tests, they were drunk.

 

[OneMoar]

really? Then why I get 24fps and 7fps on dx11. That why I get 60fps rocking stable with mantle and I get 60 with drops to 40 with dx11..

Go away.. Please how your wife have patience.. Jesus 'k

because your cpu is so under-powered that mantle is doing what its made to which is shift some load off the cpu because the drawcall limitation on the DX11API
read my edit again and stop trying to argue with people that know more then you
if you ran the same benchmark on a modern i5 or i7 system the results will be no more then a few percent different in-favor of nvidia due to simply better hardware

 

[ObscureAngelPT]

i5 faster than same clocked i7 screw those tests, they were drunk.

HT problems?

 

[ObscureAngelPT]

because your cpu is so under-powered that mantle is doing what its made to which is shift some load off the cpu because the drawcall limitation on the DX11API
read my edit again and stop trying to argue with people that know more then you

You edited it after, noob

I dont care where nvidia have the performance boost, if it is on the final stage or if it is on the DX11, i dont care at all.

What i now is that it deliver more frames with less capable CPU, doesn't matter if it is AMD or Intel.

 

[OneMoar]

You edited it after, noob

I dont care where nvidia have the performance boost, if it is on the final stage or if it is on the DX11, i dont care at all.

What i now is that it deliver more frames with less capable CPU, doesn't matter if it is AMD or Intel.

uh huh ... sure
and people wonder why I go off ... /facepalm

 

[AphexDreamer]

You edited it after, noob

I dont care where nvidia have the performance boost, if it is on the final stage or if it is on the DX11, i dont care at all.

What i now is that it deliver more frames with less capable CPU, doesn't matter if it is AMD or Intel.

I don't see how you can say more frames with less capable CPU when you didn't even change the resolution and test similarly speced cards.

 

[R-T-B]

I'm kinda skeptical at this point as well, but will do a star swarm bench when I can in a few days.

 

[ObscureAngelPT]

I don't see how you can say more frames with less capable CPU when you didn't even change the resolution and test similarly speced cards.

You need to lower the resolution for the test don't be GPU Bound, if you look at my bad shitty quality video on the MSI AB OSD Display, you can notice that on Saints Row IV the GPU usage is at 60%/70%, which means i am already CPU bound by a mile.

Lowering the resolution won't change nothing, it is only going to drop GPU usage from 60/70 to maybe 30/40%.

About the same specced cards, you may be right..
Altough the benchmarks all over the internet (which i dont trully trust) showed us with an i3 and a GTX 960 getting 60FPS
An i3 with a R9 280 getting 48fps, but with a i5 and a R9 280 it gets GPU Bound with 58FPS.

So For the AMD an i5 its needed to not bottleneck, but with nvidia, an i3 is sufficient?

Ho, ok its a Gameworks Game..

And what about this:

Its not that i trust in Digital foundry, but there is too much smoke to not exist fire.

There is people complaining on the first page of the same problem too.
I wish i had a 750TI or a 660 to match the same performance level of the HD 7850.

But i don't have.
So we are here all talking after all, so anybody can benchmark, at least i tried, but people here are more concentraded in hammering down, than trying to get more proofs by yourselfs, except for R.T.B.

I was very skeptical about this, thats why i tried for myself.
I have a friend with a GTX 650, but i can't do it right now, just in the next weeks, for that time this topic will be left for the flies.

 

[R-T-B]

but people here are more concentraded in hammering down, than trying to get more proofs by yourselfs, except for A.T.B

LOL, welcome to TPU forums, btw. It's like this all the time...

 

[ObscureAngelPT]

Well don't worry, Guru3D its the same, and Tom's hardware nobody sayed nothing, and in the best portuguese tecnological fórum "Zwame", nobody even cared.

So until now, this discussion is being great dispite we all have our own ideias, i think we are going to demistify this soon enough.
Its not like we are going to recieve a medal for this, or that AMD is even doing something, because they aren't seeing us here, but at least i want to know the truth about this ^^.
BTW Futuremark will release a overhead benchmark in DX11, DX12 and Mantle, that will be a very nice thing to benchmark too.

EDIT: BTW i will fullfil my profile, i have already done almost 20 posts, at least you can look to a beautiful anime

 

[Ferrum Master]

I googled the pic... look at that AA... it is definetly not the same. Look at the upper pipes.

 

[Tallencor]

Oh my.

 

[ObscureAngelPT]

How the hell can you see AA in so tiny image? ^^.
What is not the same and its noticable its the quantity of cars in scene.

Altough there is no configurable variable for that in that game and judging by the graphic of the i3+R9 280 it was always at ~50

 

[rtwjunkie]

LOL, welcome to TPU forums, btw. It's like this all the time...

Naw, come on, it's not always like that! In fact, most threads are very civil. You're gonna scare Obscure Angel off, LOL!

Thoughts: Star swarm is being used by MS to demonstrate DX12 capabilities. I hardly think they would use a benchmark that was CPU bound to highlight DX12 capabilities. So, it's probably not a good one to use.

Welcome, @ObscureAngelPT

 

[ObscureAngelPT]

But thats what they are doing, because the biggest advantage of the DX12 is really that, showing off that they can do much more with less CPU
And can render more drawcalls without killing all the framerate on the process.

BTW Thanks..
I am liking techpowerup fórum so far, i just don't like the reviews they do in the settings they test.
They force high values of MSAA in higher resolution, it kills completely the results that i want to know

What a way to start in a fórum ^^

 

[AphexDreamer]

You need to lower the resolution for the test don't be GPU Bound, if you look at my bad shitty quality video on the MSI AB OSD Display, you can notice that on Saints Row IV the GPU usage is at 60%/70%, which means i am already CPU bound by a mile.

Lowering the resolution won't change nothing, it is only going to drop GPU usage from 60/70 to maybe 30/40%.

About the same specced cards, you may be right..
Altough the benchmarks all over the internet (which i dont trully trust) showed us with an i3 and a GTX 960 getting 60FPS
An i3 with a R9 280 getting 48fps, but with a i5 and a R9 280 it gets GPU Bound with 58FPS.

So For the AMD an i5 its needed to not bottleneck, but with nvidia, an i3 is sufficient?

Ho, ok its a Gameworks Game..

And what about this:

Its not that i trust in Digital foundry, but there is too much smoke to not exist fire.

There is people complaining on the first page of the same problem too.
I wish i had a 750TI or a 660 to match the same performance level of the HD 7850.

But i don't have.
So we are here all talking after all, so anybody can benchmark, at least i tried, but people here are more concentraded in hammering down, than trying to get more proofs by yourselfs, except for R.T.B.

I was very skeptical about this, thats why i tried for myself.
I have a friend with a GTX 650, but i can't do it right now, just in the next weeks, for that time this topic will be left for the flies.

Just cause the GPU isn't being utilized efficiently doesn't mean its the CPU. It could be many factors other than the CPU like drivers or game optimizations.

I say lower resolutions so that you can see if the CPU even has a bottle neck to begin with. As far as I can tell you are just assuming it is for the sake of your argument. Had you tested the frames for each GPU on the smallest resolution you could get a better idea if the CPU is even bottel necking to begin with .

An i5 shouldn't be a bottle neck. As far the varying frames in the pics I'm not sure but I don't think anyone has the right to make claims just based on an observation such as this. Any ones guess is as good as yours.

 

[Ferrum Master]

That Eurogamer bench is gimped... by the looks of it, the card clearly throttles... yes exactly... It could also be CPU throttle as single core Turbo time limit goes off and thus gear shift occurs... the system is unbalanced. It is a poorly regulated engine, no matter how many pots its has... it lacks air/fuel ratio.

And Son of Rome is Cryengine... We do not why really but R9 290X still reigns there and cryegine likes GCN's raw force...

 

[ObscureAngelPT]

Just cause the GPU isn't being utilized efficiently doesn't mean its the CPU. It could be many factors other than the CPU like drivers or game optimizations.

I say lower resolutions so that you can see if the CPU even has a bottle neck to begin with. As far as I can tell you are just assuming it is for the sake of your argument. Had you tested the frames fore each GPU on the smallest resolution you could get a better idea if the CPU is even bottel necking to begin with .

Nop, i didn't but yes, i am aware that i am being bottleneck, if it is the driver or CPU its clearly that i don't know.
The Saints row IV is a well optimized game, at least i can assure you that, if it was evil within for exemple its another case since it can't maintain 60fps not even with a rocking bitch processor and a great GPU.

I know that something is bottlenecking the GPU based on the GPU used reported by GPU-Z or MSI AB.
So i didn't lowered the resolution, what now? i don't have already the nvidia to benchmark it.

That Eurogamer bench is gimped... by the looks of it, the card clearly throttles... yes exactly... It could also be CPU throttle as single core Turbo time limit goes off and thus gear shift occurs... the system is unbalanced. It is a poorly regulated engine, no matter how pots its has... it lacks air/fuel ratio.

But thats why we are here, to discover the truth.
But what about the all other ones from them, or the GameGPU.
Like i said i don't trust them entirely, but i'm pretty sure that nvidia is not leaving briefcases with Money to them to gimp all their games benchmarks.

 

[Frag_Maniac]

Your video seems nonsensical at the point you start comparing Dying Light on the same CPU but drastically different GPUs, and act surprised that the much less powerful 7850 is getting less than half the frames. I don't care what the GPU usage percentage is, they are drastically different chips you're comparing.

This is no revelation, it's just common sense. I see no reason to make an entire thread pointing out the obvious as if it were strange. Either get yourself a better AMD GPU, or use the 770, but blaming AMD for a much cheaper card not being able to keep up is just silly.

I play Dying Light on a 7970 and it runs just fine. In fact there are many recently released games where equivalently powered Nvidia cards struggle due to driver or other software issues.

On PC there are a myriad of things that can affect performance, both software and hardware, but AMD are doing pretty good on the driver end compared to Nvidia, and I see no DX11 bias, or any proof of it from you.

You said in your intro you were going by benches from sites you don't consider to be all that trustworthy, but if anything it appears you fell for their claims hook, line and sinker off one person's tests, yours, and you've yet to show any proof of your claim.

 

[AphexDreamer]

Nop, i didn't but yes, i am aware that i am being bottleneck, if it is the driver or CPU its clearly that i don't know.
The Saints row IV is a well optimized game, at least i can assure you that, if it was evil within for exemple its another case since it can't maintain 60fps not even with a rocking bitch processor and a great GPU.

I know that something is bottlenecking the GPU based on the GPU used reported by GPU-Z or MSI AB.
So i didn't lowered the resolution, what now? i don't have already the nvidia to benchmark it.



But thats why we are here, to discover the truth.

Then your claims aren't really viable, sorry.

Plus your CPU never reached %100 percent usage, as far as I recall in your videos your CPU usages was always lower than your GPU ussage? Correct? In that case I really don't think its a bottle neck of the CPU.

 

[Ferrum Master]

Then your claims aren't really viable, sorry.

Plus your CPU never reached %100 percent usage, as far as I recall in your videos your CPU usages was always lower than your GPU ussage? Correct? In that case I really don't think its a bottle neck of the CPU.

And we really do not know in what state the CPU is... it the Turbo state time limit is an unknown parameter too, that gives a random factor to each benchmark.

 

[ObscureAngelPT]

your CPU never reached %100 percent usage, as far as I recall in your videos your CPU usages was always lower than your GPU ussage? Correct? In any case I really don't think its a bottle neck of the CPU.

My CPU never reaches 100% maybe the more close i get its with Frostbite 3 on battlefield 4 i get 90-95% of usage on all 4 cores.
So it is a driver problem?

And we really do not know in what state the CPU is... it the Turbo state time limit is an unknown parameter too, that gives a random factor to each benchmark.

My CPU don't have turbo and i had cool n' quiet disable to be always forced at 3.7GHZ

 

[AphexDreamer]

My CPU never reaches 100% maybe the more close i get its with Frostbite 3 on battlefield 4 i get 90-95% of usage on all 4 cores.
So it is a driver problem?



My CPU don't have turbo.

Nvidia, AMD, the game developers, the OS are all done by different people. As hard as some may try to get even frames on all systems with varying hardware and software and OS it just is never going to happen. Your always going to see different results. In a case such as this, I attribute it the following stated and a clear hardware difference.

That isn't to say AMD might not be suffering from some serious issues as to what those are exactly I can't say, but no one really can I think.

 

[R-T-B]

I should clarify that comment I made, there ARE some really smart people here. It's part of why we get so opinionated, we're all smart about tech stuff and think we know more than the next guy.