07:23 pmoreau: imirkin Indeed: it‘s the final stretch at the moment regarding thesis writing and then defense preparation, so not much time on my hands until mid January.
07:28 pmoreau: Regarding the CTS, here is what is remaining to pass the basic test (https://gitlab.freedesktop.org/pmoreau/mesa/-/snippets/1865; raw version but without the comments: https://gitlab.freedesktop.org/pmoreau/mesa/-/snippets/1865/raw/main/Failures) One main issue is the compMask not working well for 64-bit values split into 16-bit chunks: the 16-bit chunks get the correct offset within the 32-bit value, but not the offset of the 32-bit value
07:28 pmoreau: within the 64-bit one.
07:30 pmoreau: For the CTS, I would recommend using https://github.com/pierremoreau/OpenCL-CTS/tree/support_opencl_1.0_and_1.1 which adds missing versions to some OpenCL 1.1 and OpenCL 1.2 tests, to avoid running those when testing 1.0 for example; I will need to rebase it on the latest master.
07:30 pmoreau: As for regular patches, there is the MR you discussed earlier, and then I have some additional ones in https://gitlab.freedesktop.org/pmoreau/mesa/-/tree/nv50_compute_support.
07:31 imirkin: pmoreau: thanks for the info :)
07:31 pmoreau: Thank you for looking into it. :-)
07:32 pmoreau: Feel free to ping me if there's anything, and I would be happy to answer when I can.
07:36 imirkin: will do
19:03 Ermine: You most probably know that: https://www.phoronix.com/scan.php?page=news_item&px=NVIDIA-Confirms-Kepler-Ending . Is there any chance they will release some documentation on kepler?