llama.cpp/ggml
Oliver Simons 3732b85b09 Fix data-race in `soft_max_f32_parallelize_cols_single_row`
By using `tmp_vals` to store both max values and exponential
accumulator there was a potential data-race, where the exponential accumulator
for a given CTA may have written to `tmp_vals` before all others CTAs have
read the max value from it.

To avoid a third g.sync(), an additional temporary data-storage was
added. Given that there are syncs in place after writing to gmem, it is
guaranteed that the previous values for sums/max were read by all CTAs now.
2025-12-15 11:01:12 +01:00
..
cmake ggml: Skip backend library linking code when GGML_BACKEND_DL=ON (#15094) 2025-08-07 13:45:41 +02:00
include ggml-cpu : fix RISC-V Q4_0 repack select and RVV feature reporting (#17951) 2025-12-12 16:26:03 +02:00
src Fix data-race in `soft_max_f32_parallelize_cols_single_row` 2025-12-15 11:01:12 +01:00
.gitignore vulkan : cmake integration (#8119) 2024-07-13 18:12:39 +02:00
CMakeLists.txt cmake : set `CMAKE_RUNTIME_OUTPUT_DIRECTORY` for non standalone build (ggml/1394) 2025-12-14 08:33:51 +02:00