NVIDIA 470 Series Driver Looks Like It Will Bring OpenCL 3.0 Support

[piotrj3]

what does OpenCL 3.0 mean ???? they do openCL1.2 and claim it is 3.0 because all features are optional...
this means openCL3 is a complet failure. Nvidia only claims to support 3.0 because they just support 1,2 and claim it is 3.0
because they can all features are optional means they don'T support it.
the people try openCL and end up using CUDA instead because OpenCL is complete insanity(all features are optional)...

the only fix to this is: USE VULKAN COMPUTE....
Vulkan compute is the only sane way to stop this insanity they (AMD/Nvidia) can't make a distiction between gaming cards and compute cards...

to pump up the openCL number makes only sense if the features ARE NOT OPTIONAL...

any sane person should switch from OpenCL to Vulkan right now.

Funny how you talk about Vulkan when pretty much half of Vulkan ecosystem is based on optional extensions, and simply OpenCl 3.0 tries same thing.

Still to make either Vulkan compute or OpenCl usable you need by far most, dedicated great ecosystem and libraries. This is primary reason why CUDA is light years ahead - because Nvidia spent there big moneys to make them usable.

[Alexmitter]

what does OpenCL 3.0 mean ???? they do openCL1.2 and claim it is 3.0 because all features are optional...
this means openCL3 is a complet failure. Nvidia only claims to support 3.0 because they just support 1,2 and claim it is 3.0
because they can all features are optional means they don'T support it.
the people try openCL and end up using CUDA instead because OpenCL is complete insanity(all features are optional)...

the only fix to this is: USE VULKAN COMPUTE....
Vulkan compute is the only sane way to stop this insanity they (AMD/Nvidia) can't make a distiction between gaming cards and compute cards...

to pump up the openCL number makes only sense if the features ARE NOT OPTIONAL...

any sane person should switch from OpenCL to Vulkan right now.

I had a stronk reading this

[Slartifartblast]

You would think that an earlier generation of card would have better support than a new card, not worse;

Erm, no. That's why they make new stuff.

[d3coder]

Bingo... you get 100 of maximum 100 points... yes thats right Nvidia is sapotaging the OpenCL standard.

the motive is to push their own CUDA solution.

Let me guess, bad nvidia bribed amd to make sure amd's opencl driver is almost impossible to install and use on linux?

[Mthw]

If you've got an AMD card, yeah

I do have one, can you say how?

[andre30correia]

Are using Xorg or Wayland? I have a 1050 in my laptop i think it's powered down currently. (I think).

I'm not really paying that much attention to it though. How do you know it's ON?

I assume from the tone of the post that you understand to tell Xorg to use the Intel onboard or whatever etc.

I had problem up till now unaware they were caused by my stupid dell dock with displaylink gpu that makes the nvidia one not work

with old card, old than 16xx series in hybrid mod the nvidia card is always on even if not used, ofc it is at low clock but its on, with intel mod is off with nvidia mod is allways on as primary card. in hybrid mod using a external monitor its on and working at lowest clockest who makes the desktop slow, with bumblebee we are able to turn off complety the card and wake up when we need with this older models, no ideia why nvidia don't did the same here,maybe because external monitors support

[kieffer]

Let me guess, bad nvidia bribed amd to make sure amd's opencl driver is almost impossible to install and use on linux?

I installed recently (debian11, last December) a computer with an AMD GPU (and two nvidia) and the AMD works out of the box for OpenCL compute !

[schmidtbag]

Now that CUDA is dominant due to dubious business practices by Nvidia they now finally offer OpenCL 3 when it's no threat anymore to their strategy.

(No, I don't believe there were real showstoppers due to technical issues)

I'm curious what these "dubious" business practices are that you're referring to. As bug77 pointed out, don't blame Nvidia because Khronos and their partners did nothing to make OpenCL more competitive (until it was too late). Only until recently do Intel and AMD give a shit, now that they realize how much money they're losing. The fact of the matter is, CUDA is dominating the market and it's tailor-made for Nvidia's products.

You're not wrong OpenCL 3 isn't a threat, but it never was. Nvidia has nothing to lose by supporting it, because as long as they offer competitive performance-per-watt, they're still going to rake in customers. Corporations don't give a shit about open source, so Nvidia's proprietary nature is irrelevant to the success of compute APIs.

[ColdDistance]

I'm curious what these "dubious" business practices are that you're referring to. As bug77 pointed out, don't blame Nvidia because Khronos and their partners did nothing to make OpenCL more competitive (until it was too late). Only until recently do Intel and AMD give a shit, now that they realize how much money they're losing. The fact of the matter is, CUDA is dominating the market and it's tailor-made for Nvidia's products.

You're not wrong OpenCL 3 isn't a threat, but it never was. Nvidia has nothing to lose by supporting it, because as long as they offer competitive performance-per-watt, they're still going to rake in customers. Corporations don't give a shit about open source, so Nvidia's proprietary nature is irrelevant to the success of compute APIs.

A monopoly built on one sentence I hate in computing: "It just works". OK, CUDA is better, I'm not negating that, but CUDA is a monopoly built on a proprietary technology.

I see in the future thousands of developers crying and complaining because CUDA is a monopoly, a monopoly they fed for too many years.

[kieffer]

I'm curious what these "dubious" business practices are that you're referring to.

I am sorry, but if you ask ...

* Nvidia's license which forbids gamer cards to be used within datacenter (except for blockchain). Illegal in most countries anyway. Please don't come with thermal/power issues, this is another business, we know how to deal with.
* Claim we want ECC memory (we don't) and after we set-up an experiment to count how often those events occur (we did not find one in a year on many GPUs), they claimed their have _evidences_ but never shared the protocol with us. We are a radiation facility where creating bit-flip is even part of our business !
* twist the arm of top manager and sell DGX system when technical staff (I am part of) tells: Not more than 2 GPU per computer.
(...)

Why can't we use OpenCL on ARM or PowerPC systems ? why is the driver version so restricted on those platforms ?

Nvidia is today what Intel was 10 years ago or IBM 20 years ago: the bad company.