Nvidia drivers are handicapping Ryzens performance, says new analysis by AdoredTV

[BenchZowner]

I do not understand you.
You don't agree with the word "handicapped" you claim me and others are ignoring "variables" and when I explain myself you come with this vague story about everyone on youtube being wrong?

Please point out where the reviewer is wrong, or where my logic is wrong. I really am open to other opinions but with all the info I see, best case scenario is that nvidia failed to implement true cpu scalability in their drivers, worst case scenario is that nvidia is sabotaging AMD.

I lean towards the first, but history teaches us that nvidia is capable of orchestrating the second possibility as well.

HTC's answer above covers a few of the aspects you are ignoring.

You are also ignoring that it isn't "affecting" the AMD setup IT IS ALSO HAPPENING on the Intel setup ( although at a lesser degree ).

You are also missing another point, why would nVIDIA do that ?
Do they lose sales from AMD's x86 processors ? Nope, nVIDIA does not sell x86 CPUs.

[m411b]

So time really does repeat itself.

AMD releases a new chip and as per usual; it sucks. So everything else gets blamed.

[vario]

So time really does repeat itself.

AMD releases a new chip and as per usual; it sucks. So everything else gets blamed.

[Ursus]

So time really does repeat itself.

AMD releases a new chip and as per usual; it sucks. So everything else gets blamed.

Only a genuine troll would would say "it sucks" about a brand new architecture from an underdog tech company that beats the competition in a part of the benchmarks, and offers superior performance per dollar and per watt in most of them.

[vario]

Similar but different test in the division, now its 470 vs 1060 in single card mode (from what i gather).

Similar results

https://www.youtube.com/watch?v=QBf2...ature=youtu.be

[Ursus]

Similar but different test in the division, now its 470 vs 1060 in single card mode (from what i gather).

Similar results

https://www.youtube.com/watch?v=QBf2...ature=youtu.be

The plot thickens.
One conclusion could be that nvidia just has very bad DX12 drivers, and Ryzen really likes DX12?

[HTC]

The plot thickens.
One conclusion could be that nvidia just has very bad DX12 drivers, and Ryzen really likes DX12?

Another conclusion is that if / when nVidia "fixes" their drivers, their card's performance will rise ...

On a more serious note, it's a bit early to tell because these may be the only games "affected" for all we know, but it definitely warrants further games testing, IMO.

Somewhat related:

https://www.youtube.com/watch?v=nIoZB-cnjc0&t=0s

If this is true, then nVidia will not be able to "fix" this, the same way AMD can't "fix" their lower DX11 performance

[zanzabar]

Another conclusion is that if / when nVidia "fixes" their drivers, their card's performance will rise ...

On a more serious note, it's a bit early to tell because these may be the only games "affected" for all we know, but it definitely warrants further games testing, IMO.

Somewhat related:

https://www.youtube.com/watch?v=nIoZB-cnjc0&t=0s

If this is true, then nVidia will not be able to "fix" this, the same way AMD can't "fix" their lower DX11 performance

i dont think it is not unfixable for pascal and it 100% will be fixed in volta as they are adding even more hardware threading. pascal has 3 threads for a hard ware scheduler to use, but the question is if they can feed that from different cpu threads with graphics work. we know GPGPU can do it, and that mixed use can do it, the question is if graphics only can or should do it. there might be a maxwell like problem where it supports two hardware threads that can be fed from cpu threads, but the latency was too high so it was never used with graphics api since it lowered performance (or so they said when pascal launched and async questions had to be answered.)

[HTC]

i dont think it is not unfixable for pascal and it 100% will be fixed in volta as they are adding even more hardware threading. pascal has 3 threads for a hard ware scheduler to use, but the question is if they can feed that from different cpu threads with graphics work. we know GPGPU can do it, and that mixed use can do it, the question is if graphics only can or should do it. there might be a maxwell like problem where it supports two hardware threads that can be fed from cpu threads, but the latency was too high so it was never used with graphics api since it lowered performance (or so they said when pascal launched and async questions had to be answered.)

I didn't explain myself properly: i meant they can't fix it unless they make changes to the way the cards work, like a new architecture.

AMD seems to be in the same situation because unless Vega's architecture changes something, it will be slower in DX11 then nVidia's cards, regardless of how fast in DX12 it is.

[Ace123]

Ryzen absolutly does not suck. Only a fool would say that.

[kromosto]

Ryzen might not satisfied all of us but saying it sucks is very extreme.

I can give a possibility why nvidia does something like this. Maybe they dont want AMD to make high income because they can control it with it's low budget.

And on the dx12 side i dont like how dx 12 is going. Threads performance ok but i really want to see the shared memory along with cards and seeing multi gpu as one aspects of it. But rather than that nvidia seems to preparing itself to left sli behind.

[zanzabar]

Ryzen might not satisfied all of us but saying it sucks is very extreme.

I can give a possibility why nvidia does something like this. Maybe they dont want AMD to make high income because they can control it with it's low budget.

And on the dx12 side i dont like how dx 12 is going. Threads performance ok but i really want to see the shared memory along with cards and seeing multi gpu as one aspects of it. But rather than that nvidia seems to preparing itself to left sli behind.

it has nothing to do with that since it is an intel problem on 2011 too. nvidia has a single thread feed the gpu most of the time and amd can chop it up.

[bill_d]

i dont think it is not unfixable for pascal and it 100% will be fixed in volta as they are adding even more hardware threading. pascal has 3 threads for a hard ware scheduler to use, but the question is if they can feed that from different cpu threads with graphics work. we know GPGPU can do it, and that mixed use can do it, the question is if graphics only can or should do it. there might be a maxwell like problem where it supports two hardware threads that can be fed from cpu threads, but the latency was too high so it was never used with graphics api since it lowered performance (or so they said when pascal launched and async questions had to be answered.)

I seem to remember my two asus gtx 980 Poseidon's say dx12 all over the box's and when it didn't work they said they would fix it with pascal, i'll believe they will fix dx12 in volta only when I see it work

now I just got a 1080 ti strix but this time I know it is sh*t at dx12 and will be sh*t for it's whole life span no matter what it says on the box
but at 4k I need it till Vega hits 980 sli is dead at 4k and my fury x cfx is running out of gas also
now hoping I can live with one as NVidia seems to want to kill sli also now

but it is going on a new C6H 1800x and if NVidia does not suck it up and work out any and all problems with their drivers with Ryzen I'll stop buying anything NVidia .

[StyM]

ryzen is a beast for crunching period...

[zanzabar]

ryzen is a beast for crunching period...

you guys are just drooling over the SR right, having g34 with viable IPC.