Episode 1: What is OpenCL™?

In this video, you learn what OpenCL™ is and why it was designed the way it
is. We go through design goals and the execution model of OpenCL™. Topics
covered include basic terminology, data-parallelism, execution domains and
synchronization.

[harshivpatel]

4:07 global work items must be independent--- is this analogous to a fragment shader operation being independent of other fragments?

[orbita]

what is the difference between a work item and a kernel in this context?

[B.D.M]

how to learn ml using amd gpu as i have dont have support of pytorch due to lack of support in cuda cores of my 6650xt gpu . is there any alternative way to learn ml better than nvedia gpu. olease guide me if possible or make a vedio upon that

[hilmar]

Is it still a thing in 2021? I read that AMD quit supporting OPENCL but I would like to use it on an AMD GPU, does anybody know this?

[RobotCentral]

Awesome stuff. Thank you for the primer. I'm going to have to watch this a few times and practice it to get it down.

[projectjt]

Are most of AMD's software technologies based off of OpenCL? (Mantle, TressFX, and TrueAudio, etc.)

[TylerBata]

is openCL supported in intel HD graphics?
If this is so then since the RAM is shared on intel HD graphics then can both the CPU and the intel HD graphics use the RAM?

PS. I am trying to wright a ray caster in C++ and would like to learn openCL as well any!

[vertigo]

Data Parallelism - 5 minutes of video time.
Task Parallelism - 44 seconds of video time.

It's obvious which one is more complex and will require more youtube videos :)

[submarmazaj]

No use for 3d matrixex, - have to do dynamic re+compression or will take too much of RAM.

[JiaxingQi]

This guy wears a shirt with a ATi logo.

[soldirix]

lol, talking about a pencil
subtitles ftw :p.

[dashrek]

Looks like David Faustino from Bundy family.