AMD Running DOOM Vulkan Benchmarks On Intel Hardware

That is the exact opposite of what DX12 is supposed to accomplish. DX12 allows performance improvements for lower end cpu's to help alleviate that bottleneck. In fact the faster the processor the less improvement you'll see, although it does tend to help alleviate xfire and sli bottlenecks as well.
We all know that. Just because DX12 benefits slower processors doesn't mean that a faster processor won't benefit also. The slower the processor, the higher the chance of a CPU bottleneck. Faster CPU with more cores/threads will ensure there is no chance of that happening and eliminate the CPU as a variable in testing. You have to remember, this discussion is about AMD's use of a super high end Intel chip for benchmarking and marketing purposes of its GPU's.
 
We all know that. Just because DX12 benefits slower processors doesn't mean that a faster processor won't benefit also. The slower the processor, the higher the chance of a CPU bottleneck. Faster CPU with more cores/threads will ensure there is no chance of that happening and eliminate the CPU as a variable in testing. You have to remember, this discussion is about AMD's use of a super high end Intel chip for benchmarking and marketing purposes of its GPU's.

I never said it didn't. You are arguing against yourself here. I just pointed out where you were incorrect the way you stated it in the previous post. It will benefit slower processors the most.

Maybe its just the way you posted the comment?
 
I don't think we'll see any DX12-only games any time soon, at least not from non-MS studios - it makes no financial sense for them.

Nor Vulkan only titles probably, not any major ones. As you point out it's about money. Windows 7 as a reason to use Vulkan will become less and less of reason over the coming years. I'm not saying Vulkan doesn't make more financial sense. But if it did then I'm guessing no game dev would have ever used DX anyway and gone with OGL. All that's certain over the next few years is that Windows 10 will make up more and more of the new game market on the PC side and typically areas of growth are where devs focus attention.
 
I never said it didn't. You are arguing against yourself here. I just pointed out where you were incorrect the way you stated it in the previous post. It will benefit slower processors the most.

Maybe its just the way you posted the comment?
I didn't discuss anything about which processors would benefit most. However, I think I see where we got off on the wrong foot. I highlighted the sentence in question below in our discussion.

I don't think its about "real world pairing" per se. I think it has to do with eliminating any chance of a CPU bottleneck at all cost to show the true potential of the GPU for marketing reasons. Throwing the fastest CPU with the most physical cores/threads at the problem would be ideal especially for DX12 multithreading.

That is the exact opposite of what DX12 is supposed to accomplish. DX12 allows performance improvements for lower end cpu's to help alleviate that bottleneck. In fact the faster the processor the less improvement you'll see, although it does tend to help alleviate xfire and sli bottlenecks as well.
~~~~~~~~~~~~~~~~~~~~~~~~~~
I believe you either misread or took my last sentence out of context. That highlighted sentence is directly referencing the previous sentence. ;)
When I said "Throwing the fastest CPU...at the problem..." the key word is "problem" which is the possible chance of a CPU bottleneck during benchmarking a GPU. It doesn't matter how miniscule that chance is, you want to eliminate it to set a baseline. These benchmarks are for GPU performance, they aren't referencing how the CPU benefits. If they had wanted to show how DX12 benefits lower end CPU's they would have chosen a different CPU while keeping the GPU static. I mentioned this in a previous post in this same thread.
 
It's all about reducing variables. A comparison review or even in this case a single benchmark of GPU's using different CPU's would be quite idiotic.
 
Back
Top