We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
1 parent a593a4f commit 30d921aCopy full SHA for 30d921a
Makefile
@@ -169,6 +169,8 @@ ifdef LLAMA_HIPBLAS
169
LDFLAGS += -L/opt/rocm/lib -Wl,-rpath=$(ROCM_PATH)/lib -lhipblas -lamdhip64
170
OBJS += ggml-cuda.o
171
ggml-cuda.o: CXXFLAGS += $(addprefix --offload-arch=,$(GPU_TARGETS))
172
+ggml-cuda.o: CXXFLAGS += -DGGML_CUDA_DMMV_X=64
173
+ggml-cuda.o: CXXFLAGS += -DGGML_CUDA_DMMV_Y=1
174
ggml-cuda.o: ggml-cuda.cu ggml-cuda.h
175
$(CXX) $(CXXFLAGS) -x hip -c -o $@ $<
176
endif
0 commit comments