AMD Project ReSX is an eSports Gamer Outreach by Making Radeon More Responsive

[kruk]

AMD's drivers have historically been more CPU intensive than Nvidia's when it comes to DX (and OpenGL). How can a developer work around that? What's the job of a driver programmer if not squeezing everything that can be squeezed from the hardware?

My point is that some devs obviously can make games run faster on DX11 that others (just look at the benchmarks here on TPU from like 2016). If you can't optimize like they do in DX11, why not go Vulkan/DX12? Several engines already support it. Why should AMD resort to driver hacks, if there is a much better native alternative for both, red and green team?

 

[bug]

My point is that some devs obviously can make games run faster on DX11 that others (just look at the benchmarks here on TPU from like 2016). If you can't optimize like they do in DX11, why not go Vulkan/DX12? Several engines already support it. Why should AMD resort to driver hacks, if there is a much better native alternative for both, red and green team?

So your argument against AMD doing something about their higher driver overhead is that devs should use a different API. Neat.

 

[renz496]

You getting on topic at some point??

You brought up compute not I, i just corrected some of your miss direction.

I find it really funny when trolls work tangential arguments.

And i initially did not talk to you either. I was replying to HD64G about what he says about GCN compute over nvidia. Now you complain me talking about compute and not you who brought the topic? Before you call me a troll maybe you should look yourself at the mirror first.

 

[BiggieShady]

I was replying to HD64G about what he says about GCN compute over nvidia.

... and it was a good point you made. However I would add that mining algorithms are not written for GCN, rather suitable for GCN - the whole class of algorithms favors GCN as it lets it operate closer to architecture's peak efficiency (perf/watt). All kinds of cuda mining software have great gcn counterparts, can't say the same for vice versa. As for pointing out Fermi as a milestone, I agree - all that nvidia has been doing last 7 years was refining fermi and look where it got them perf/watt wise.

 

[kruk]

So your argument against AMD doing something about their higher driver overhead is that devs should use a different API. Neat.

Great selective reading skills and nice bait, but I'll pass ...

 

[bug]

Great selective reading skills and nice bait, but I'll pass ...

Thanks. It means a lot coming from you.