From d052c2df34ab41ba285f70965663e5a0832f6ac9 Mon Sep 17 00:00:00 2001 From: Joshua Bakita Date: Thu, 19 Dec 2024 14:20:38 -0500 Subject: Bugfix stream-mask override, support old CUDA, and start Hopper support Use a different callback to intercept the TMD/QMD later in the launch pipeline. Major improvements: - Fix bug with next mask not overriding stream mask on CUDA 11.0+ - Add CUDA 6.5-10.2 support for next- and global-granularity partitioning masks on x86_64 and aarch64 Jetson - Remove libdl dependency - Partially support TMD/QMD Version 4 (Hopper) Minor improvements: - Check for sufficient CUDA version before before attempting to apply a next-granularity partitioning mask - Only check for sufficient CUDA version on the first call to `libsmctrl_set_next_mask()` or `libsmctrl_set_global_mask()`, rather than checking every time (lowers overheads) - Check that TMD version is sufficient before modifying it - Improve documentation Issues: - Partitioning mask bits have a different meaning in TMD/QMD Version 4 and require floorsweeping and remapping information to properly construct. This information will be forthcoming in future releases of libsmctrl and nvdebug. --- Makefile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'Makefile') diff --git a/Makefile b/Makefile index 0e9ee3a..0d9b9f6 100644 --- a/Makefile +++ b/Makefile @@ -3,7 +3,7 @@ CXX = g++ NVCC ?= nvcc # -fPIC is needed in all cases, as we may be linked into another shared library CFLAGS = -fPIC -LDFLAGS = -lcuda -I/usr/local/cuda/include -ldl +LDFLAGS = -lcuda -I/usr/local/cuda/include .PHONY: clean tests -- cgit v1.2.2