RDNA can join two compute units to compute a single block of workitems in OpenCL. Will they increase this to add more hardware available per GPU thread?
But also like to have some hybrid of RTS + RPG + FPS like Planetside-2 with a touch of Diablo(or Gothic 2).
Planetside 2 + Gothic 2 mix would be awesome. Dragons + orcs vs APC+tanks+helicopters, with style.
Anything like System Shock 2 and a bit more depth on skill tree: https://cdn.mos.cms.futurecdn.net/uXLdVS2ZP6RmLGvGRyhrnB.png
If BioShock is like that, I may try.
Does battlefield add my char critical-hit ability? Or hit-point regeneration skill?
I think deus-ex has this but very very limited. System-shock also similar and also limited. I need something like path-of-exile style of development in a tense counter-strike-ish FPS.
Even a diablo-fps remake...
I guess it will have even lower boost frequencies but much better performance per core, especially with 1024 bit register (64 of them = 64kbit / 8kB fastest memory, maybe with 500-1000 GB/s bandwidth per core).
From OpenCL 1.2 to 2.0, C# GPGPU environment for multiple GPU loadbalancing(kernel partitioning) or task pooling, pipelining and whatever type of dynamic parallelism is needed:
Download: https://github.com/tugrul512bit/Cekirdekler
Wiki: https://github.com/tugrul512bit/Cekirdekler/wiki
Example:
When self-bootable gpus come, pci-e bandwidth bottleneck will be unimportant and more vram will be very important. Because even hdd/solidstate drives will be slapped near gpu for faster game loadings/savings. Directly to vram from hdd in a ramdisk manner.
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.