Will be interesting to see where VRAM usage settles out with the new OpenCL 3.0 specification that is being finalized. It is based on the old 1.2 specification that Nvidia settled on.
Will AMD cards in the future be limited to 27% of memory like Nvidia cards currently are?
They're making a new OpenCL? I thought OpenCL was being depricated in favour of Vulkan.
If this page takes an hour to load, reduce posts per page to 20 in your settings, then the tinpot 486 Einstein uses can handle it.
Yes, I know. I was replying in the context of GPGPU compute. Read the Anandtech link for a deep dive into the OpenCL 3.0 specification and learn more than you ever wanted about OpenCL.
it was more of a preemptive reply to Peter for when he inevitably googles "Vulkan" and reads the first sentence on the wikipedia page which says it does compute in very general terms. lol.
it was more of a preemptive reply to Peter for when he inevitably googles "Vulkan" and reads the first sentence on the wikipedia page which says it does compute in very general terms. lol.
[Spanks Ian or Steve (which one am I talking to?) very hard indeed]
If Wikipedia is wrong, correct it.
If this page takes an hour to load, reduce posts per page to 20 in your settings, then the tinpot 486 Einstein uses can handle it.
its not wrong, its just general and open to the wrong impression from people who don't bother reading more than the first sentence of wikipedia articles. wikipedia is not a tech journal, and mostly just gives basic information about a lot of things.
but the fact is that Vulkan, while it has compute capabilities, they are meant for the context of 3D rendering and not pure GPGPU. you probably can use Vulkan for GPGPU, but since it's not really designed for that, no one is doing it. everyone is using it for it's intended purpose, gaming.
its not wrong, its just general and open to the wrong impression from people who don't bother reading more than the first sentence of wikipedia articles.
Oy! Are you implying that I'm a lazy reader? I can guarantee that you'd be entirely correct.
Ian&Steve C. wrote:
wikipedia is not a tech journal, and mostly just gives basic information about a lot of things.
but the fact is that Vulkan, while it has compute capabilities, they are meant for the context of 3D rendering and not pure GPGPU. you probably can use Vulkan for GPGPU, but since it's not really designed for that, no one is doing it. everyone is using it for it's intended purpose, gaming.
So we're stuck with the OpenCL and Cuda wars.... if only people could get along and work together instead of reinventing the wheel?
If this page takes an hour to load, reduce posts per page to 20 in your settings, then the tinpot 486 Einstein uses can handle it.
So we're stuck with the OpenCL and Cuda wars.... if only people could get along and work together instead of reinventing the wheel?
But, but, but we would be completely without "wheelies" if people weren't re-inventing the wheel...
Tom M
A Proud member of the O.F.A. (Old Farts Association). Be well, do good work, and keep in touch.® (Garrison Keillor) I want some more patience. RIGHT NOW!
Keith Myers wrote: Will be
)
They're making a new OpenCL? I thought OpenCL was being depricated in favour of Vulkan.
If this page takes an hour to load, reduce posts per page to 20 in your settings, then the tinpot 486 Einstein uses can handle it.
https://www.extremetech.com/c
)
https://www.extremetech.com/computing/309842-opencl-3-0-kicks-off-with-a-huge-step-backwards
https://www.anandtech.com/show/15746/opencl-30-announced-hitting-reset-on-compute-frameworks
Quote: They're making a new
)
Vulkan IS NOT a compute API. It is a graphics API.
well Vulkan does have
)
well Vulkan does have compute, but it's not really for GPGPU, its more for doing computations that are feeding a 3D rendering process.
_________________________________________________________________________
Yes, I know. I was replying
)
Yes, I know. I was replying in the context of GPGPU compute. Read the Anandtech link for a deep dive into the OpenCL 3.0 specification and learn more than you ever wanted about OpenCL.
it was more of a preemptive
)
it was more of a preemptive reply to Peter for when he inevitably googles "Vulkan" and reads the first sentence on the wikipedia page which says it does compute in very general terms. lol.
_________________________________________________________________________
Ian&Steve C. wrote: it was
)
[Spanks Ian or Steve (which one am I talking to?) very hard indeed]
If Wikipedia is wrong, correct it.
If this page takes an hour to load, reduce posts per page to 20 in your settings, then the tinpot 486 Einstein uses can handle it.
its not wrong, its just
)
its not wrong, its just general and open to the wrong impression from people who don't bother reading more than the first sentence of wikipedia articles. wikipedia is not a tech journal, and mostly just gives basic information about a lot of things.
but the fact is that Vulkan, while it has compute capabilities, they are meant for the context of 3D rendering and not pure GPGPU. you probably can use Vulkan for GPGPU, but since it's not really designed for that, no one is doing it. everyone is using it for it's intended purpose, gaming.
_________________________________________________________________________
Ian&Steve C. wrote:its not
)
Oy! Are you implying that I'm a lazy reader? I can guarantee that you'd be entirely correct.
So we're stuck with the OpenCL and Cuda wars.... if only people could get along and work together instead of reinventing the wheel?
If this page takes an hour to load, reduce posts per page to 20 in your settings, then the tinpot 486 Einstein uses can handle it.
Peter Hucker wrote:So we're
)
But, but, but we would be completely without "wheelies" if people weren't re-inventing the wheel...
Tom M
A Proud member of the O.F.A. (Old Farts Association). Be well, do good work, and keep in touch.® (Garrison Keillor) I want some more patience. RIGHT NOW!