topic Re:OpenCl in Intel® oneAPI DPC++/C++ Compiler

OpenCl

Ronen77 — Tue, 10 Nov 2020 12:47:21 GMT

Hi,

I am new to opencl, and would like to know if it is possible to integrate opencl with an existing application.

For instance, my application is developed with Visual Studio 2019, are there any GPU Api that can be called by the App to do some gpu calculations abd get the results back to the App?

Or, I would have to rewrite the entire App from scratch using opencl in order to get access to the

gpu integrated with the Intel chip/

Re:OpenCl

RahulV_intel — Wed, 11 Nov 2020 10:03:18 GMT

Hi,

Assuming that your application is developed in C/C++, you can install the oneAPI-basetoolkit, to perform the required computation on iGPU using the OpenCL backend. (env variable is SYCL_BE=PI_OPENCL)

Please refer to the links below to get started:

Download link: https://software.intel.com/content/www/us/en/develop/tools/oneapi/base-toolkit/get-the-toolkit.html

Reference guide for DPC++ (direct programming):

https://software.intel.com/content/www/us/en/develop/documentation/oneapi-dpcpp-cpp-compiler-dev-guide-and-reference/top.html

OneDPL for API based programming:

https://software.intel.com/content/www/us/en/develop/documentation/oneapi-dpcpp-library-guide/top.html

Also, could you let me know what exactly is the calculation that you want to perform on iGPU?

Thanks,

Rahul

Re:OpenCl

RahulV_intel — Tue, 17 Nov 2020 10:49:52 GMT

Hi,

Any updates on this?

Thanks,

Rahul

Re: Re:OpenCl

Ronen77 — Tue, 17 Nov 2020 11:44:45 GMT

Hi,

Download size of Download the Intel® oneAPI Base Toolkit is 22.5 GB, I don’t have space for this at the moment
Assuming that your application is developed in C/C++, you can install the oneAPI-basetoolkit, to perform the required computation on iGPU using the OpenCL backend. (env variable is SYCL_BE=PI_OPENCL)
I need to know whether the gpu that resides on the core-i7 chip is capabale to do the vector multiply. How big is it?
Is there any tutorial that explains how difficult to do some c = a + b parallel programming on top of the gpu which is on the local cpu?
The calculations is to do: B = A * A (the vector size is 1000-2000) and float
If it is too complicated, I might use the nvidia chip instead.
To remind you that the Application is done already, but i consider to enhance with the gpu

Regards,

Ronen Levy

Re: OpenCl

RahulV_intel — Thu, 19 Nov 2020 11:37:13 GMT

Hi,

Kindly refer to the below code samples and other resources to get started:

Vector add sample:

https://github.com/oneapi-src/oneAPI-samples/tree/master/DirectProgramming/DPC%2B%2B/DenseLinearAlgebra/vector-add

Other samples:

https://github.com/oneapi-src/oneAPI-samples/tree/master/DirectProgramming/DPC%2B%2B

Webinar links:

https://techdecoded.intel.io/essentials/dpc-part-1-an-introduction-to-the-new-programming-model/

https://techdecoded.intel.io/essentials/dpc-part-2-programming-best-practices/

DPC++ book:

https://link.springer.com/book/10.1007/978-1-4842-5574-2

Thanks,

Rahul

Re:OpenCl

RahulV_intel — Thu, 26 Nov 2020 05:21:12 GMT

Hi,

If the solution provided (above) helps, could you let me know if I can close this thread from my end?

Thanks,

Rahul

Re:OpenCl

RahulV_intel — Tue, 01 Dec 2020 08:39:12 GMT

Hi,

I have not heard back from you, so I will go ahead and close this thread from my end. Feel free to post a new question if you require any further assistance from Intel.

--Rahul