Module: Mesa Branch: master Commit: cf12faef614ab7cd9996410f1d161558a3853936 URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=cf12faef614ab7cd9996410f1d161558a3853936
Author: Jordan Justen <[email protected]> Date: Thu Oct 24 11:55:23 2019 -0700 intel/compiler: Restrict cs_threads to 64 Our current GPGPU_WALKER code only supports up to 64 threads. On HSW we could use up to 70 and TGL up to 112, but only if the walker is adjusted so the width does not exceed 64. Work to support this is in progress. Previous to this change, we might try to downgrade to SIMD8 if the SIMD16 shader spilled. Since HSW and TGL have the max number of threads above 64, we would then try to emit an invalid GPGPU walker command. Fixes: 932045061b5 ("i965/cs: Emit compute shader code and upload programs") Signed-off-by: Jordan Justen <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]> Reviewed-by: Paulo Zanoni <[email protected]> Tested-by: Paulo Zanoni <[email protected]> --- src/intel/compiler/brw_fs.cpp | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/src/intel/compiler/brw_fs.cpp b/src/intel/compiler/brw_fs.cpp index 07963ef6490..10a73251a44 100644 --- a/src/intel/compiler/brw_fs.cpp +++ b/src/intel/compiler/brw_fs.cpp @@ -8841,8 +8841,10 @@ brw_compile_cs(const struct brw_compiler *compiler, void *log_data, src_shader->info.cs.local_size[0] * src_shader->info.cs.local_size[1] * src_shader->info.cs.local_size[2]; + /* Limit max_threads to 64 for the GPGPU_WALKER command */ + const uint32_t max_threads = MIN2(64, compiler->devinfo->max_cs_threads); unsigned min_dispatch_width = - DIV_ROUND_UP(local_workgroup_size, compiler->devinfo->max_cs_threads); + DIV_ROUND_UP(local_workgroup_size, max_threads); min_dispatch_width = MAX2(8, min_dispatch_width); min_dispatch_width = util_next_power_of_two(min_dispatch_width); assert(min_dispatch_width <= 32); _______________________________________________ mesa-commit mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/mesa-commit
