Summary
Is OpenCL support for the Mali-T628 (for example as found in the Exynos 5420 SoC on the Arndale Octa board) available? If so, how to set it up?
More details
According to the vendor, OpenCL should be supported, but the Arndale Octa Wiki does not state how this can be achieved.
I am using the latest Linaro developer build and installed Mali drivers that contain OpenCL libraries for Mali T604. According to this guide, the driver actually contains references to the Mali T628. So I tried to create the udev rule as specified, which is supposed to solve a permission problem with /dev/mali0, but I found that there is no /dev/mali0 on my installation at all. So my conclusion is that the driver indeed does not support T628.
When I execute a clinfo utility, clGetDeviceInfo returns CL_OUT_OF_HOST_MEMORY for some device properties. Why can I query the GPU for some characteristics, but does this fail for some others? When running a normal application, the same error appears when trying to create an OpenCL Context.
I was surprised to find this topic, where yoshi seems to have OpenCL working and can run benchmarks on his Arndale Octa board. How is this possible if there is no driver available? Or am I just missing something? I hope that you can help me to also establish a working OpenCL development environment.
Hi Chris,
Where can we find benchmark numbers for examples kernels provided with SDK?. We ran "sobel" filter example for 1024x1024 image, profile number measured across kernel trigger and wait measures as 5msec. And Laplace transform for 4096x2048 image, number measures as 10msec. Is these numbers expected for for Mali T628 @ 533Mhz GPU? We measured the numbered across clEnqueueNDRange() and immediate clFinish().
Thanks,
Veeranna
Hi Veerannah,
We don't have stock performance numbers for the SDK samples to hand. However what I will do later this week is run Sobel and Laplace on a platform here and get some numbers back to you.
HTH, Tim
Hi Tim,
We really appreciate your quick reply. Looking forward to see the performance numbers.
Hi Veeranna,
I ran the Sobel filter as provided by the Mali OpenCL SDK, modifying it to run a 1024x1024 image and I see the following...
Queued time: 0.11ms
Wait time: 0.222292ms
Run time: 0.680683ms
The Mali OpenCL SDK doesn't include a Laplace transform... which one are you using?
Best regards, Tim
I realise I didn't include the platform details. I'm running on a Mali-T604 platform with the GPU clocked @ 533MHz. I'm using a later version of the drivers than you, I believe, and that would likely mean the performance you would expect on your platform is slightly worse. On balance though I would expect the performance on your Mali-T628 system to be roughly equivalent to mine.
Thank you for providing information on performance number. With "cl_event" profiling we are getting numbers as
Wait time: 0.2725ms
Run time: 0.7586ms
But if we measure numbers across clEnqueueNDRangeKernel() and clfinish() using gettimeofday we are seeing number is 1.5msec. Is this API/driver overhead?
Is this API/driver overhead?
Probably, yes. When I try this I'm seeing the same sort of additional time using gettimeofday vs the profiling values output. With the Sobel sample I looked at the difference in this additional time when I increased the size of the image being processed. It appeared to stay roughly the same, which would back up the theory that this is due to additional overhead in the API.
Hope that helps,
Tim
Thank you for the information.