1 |
On 2020-04-02 15:31, Marek Szuba wrote: |
2 |
|
3 |
> 3. Test if downstream applications are happy with the new unified OpenCL |
4 |
> headers required by the Khronos Group ICD loader and if they are, add |
5 |
> dev-libs/opencl-icd-loader to virtual/opencl as an alternative to ocl-icd; |
6 |
|
7 |
Quick update on this: today I have attempted to rebuild and test various |
8 |
OpenCL-dependent packages using the unified headers and |
9 |
dev-libs/opencl-icd-loader. The good news is, they have all been |
10 |
perfectly happy with them. The bad news is, I actually had to manually |
11 |
symlink all contents |
12 |
/usr/lib/OpenCL/vendors/opencl-icd-loader/include/CL/ into |
13 |
/usr/include/CL/ for this to work because eselect-opencl contains a |
14 |
hard-coded list of expected vendor header files. |
15 |
|
16 |
In other words, it will be necessary to either teach eselect-opencl how |
17 |
to handle the unified headers or make it possible for it to let the |
18 |
contents of /usr/include/CL be. Personally I am leaning towards the |
19 |
latter - it doesn't even handle legacy headers that well (it installs |
20 |
several versions into /usr/lib/OpenCL/global but in the end offers no |
21 |
way of switching between them), and in any case even its description |
22 |
suggests its job is to switch between implementations rather than handle |
23 |
global headers. |
24 |
|
25 |
-- |
26 |
MS |