Or it means NVIDIA architecture is better than AMD at fully utilizing the GPU in that particular application. If AMD could fix their utilization issue suddenly async compute wouldn't look as cool.
Or you can continue making ill-informed comments that clearly demonstrate your lack of understanding of the subject matter overall while still thinking you are correct.
I don't really believe that he thinks he's correct. He's putting too much effort into discrediting it. If there was true confusion, I don't think he would be bringing AMD into it.
Or it means NVIDIA architecture is better than AMD at fully utilizing the GPU in that particular application. If AMD could fix their utilization issue suddenly async compute wouldn't look as cool.
1, 2, 3..waiting for the irate replies of those that still don't understand how async compute works/what it does and doesn't do.
God, for the thousand time, how can you gain performance if your resource is already fully utilized ? (ie nvidia). You can't go 110% !It s tailored toward async compute, either your hardware support it and there s some gain or it doesnt support it and at best there s no losses...
I guess that you didnt get that async compute is about gaining something...
Or it means NVIDIA architecture is better than AMD at fully utilizing the GPU in that particular application. If AMD could fix their utilization issue suddenly async compute wouldn't look as cool.
1, 2, 3..waiting for the informed replies of those that still don't understand how async compute works/what it does and doesn't do.
That is why AMD spent so much effort to lay down the software groundwork for PC gaming as well as capturing the console chip deals. To sow the seed for the architecture that they designed to bear fruits.God, for the thousand time, how can you gain performance if your resource is already fully utilized ? (ie nvidia). You can't go 110% !
You can only gain performance with DX12 async if you have idling ALUs under DX11. Thats the whole story of CGN poor utilization under DX11 and the reason why AMD much higher FLOPs never materialize in performance lead (under DX11).
what is so difficult to understand ?
Or it means NVIDIA architecture is better than AMD at fully utilizing the GPU in that particular application. If AMD could fix their utilization issue suddenly async compute wouldn't look as cool.
1, 2, 3..waiting for the irate replies of those that still don't understand how async compute works/what it does and doesn't do.
Oh boy. When you ended up proving your opponents point, but you didn't even realize it. So wait Nvidia implemented a feature because they already have perfect or near perfect utilization? If you have perfect utilization you would just say async compute is BS here is our already perfect utilization because our hardware and software rocks. They didn't though. They tasked their driver team to implement dynamic load balancing and then demo'd it . Therefore they acknowledged they have gaps in their utilization or they are straight up lying for marketing purposes. Not only that, but scene to scene never takes the exact same resources. To claim otherwise is just plain silly. I think we can safely assume Nvidia is attempting to tackle an issue that is very much real.
Thing I don't understand is this:
If concurrent graphics+compute didn't matter then why would Sony have specifically requested for an additional 6 ACEs in GCN? Why spend money on the hardware if it isn't important?
Why have so many prominent devs post about the benefits of asynchronous compute? Is it common for them to tweet about trivial hardware features?
About Pascal, I'm just glad there isn't a performance regression in dx12, going to be very good for all owners of GCN cards.
Cerny expects developers to run middleware -- such as physics, for example -- on the GPU. Using the system he describes above, you can run at peak efficiency, he said.
"If you look at the portion of the GPU available to compute throughout the frame, it varies dramatically from instant to instant. For example, something like opaque shadow map rendering doesn't even use a pixel shader, it’s entirely done by vertex shaders and the rasterization hardware -- so graphics aren't using most of the 1.8 teraflops of ALU available in the CUs. Times like that during the game frame are an opportunity to say, 'Okay, all that compute you wanted to do, turn it up to 11 now.'"
Thing I don't understand is this:
If concurrent graphics+compute didn't matter then why would Sony have specifically requested for an additional 6 ACEs in GCN? Why spend money on the hardware if it isn't important?
Why have so many prominent devs post about the benefits of asynchronous compute? Is it common for them to tweet about trivial hardware features?
About Pascal, I'm just glad there isn't a performance regression in dx12, going to be very good for all owners of GCN cards.
Who said async compute doesn't matter?
If AMD could fix their utilization issue suddenly async compute wouldn't look as cool.
Your rate of flip flopping is higher than a quantum qubit D:Who said async compute doesn't matter?
Your rate of flip flopping is higher than a quantum qubit D:
Having the ability to do proper Async Compute doesnt say anything about performance improvements.
Ashes is tailored towards AMD. So why would anyone expect the same or any gain on nVidia hardware?
Add to this, Copy queues can directly work on DMAs engines (GCN has 2 active, Kepler/Maxwell has 2, but 1 is disabled in DX, only 2 accessible via CUDA for some reason), to get transfers going concurrently.
Or it means NVIDIA architecture is better than AMD at fully utilizing the GPU in that particular application. If AMD could fix their utilization issue suddenly async compute wouldn't look as cool.
1, 2, 3..waiting for the irate replies of those that still don't understand how async compute works/what it does and doesn't do.
You seem to not understand that without async compute certain tasks have to wait until another task is completed before it can run. That creates a stall. Async allows these tasks to run simultaneously. Without it, they simply can't do that. nVidia can't possibly do without it what AMD, or any hardware, does with it. nVidia is the one who has something to fix. Not AMD.
Your rate of flip flopping is higher than a quantum qubit D: