moved volta onto tile kernels, so building for cc7.0 can be avoided

this shouldn't do anything (+2 squashed commit)

Squashed commit:

[1cdcb302a] another attempt to tip the scales, part 2

[8f647b709] another attempt to tip the scales (volta)
This commit is contained in:
Concedo
2025-12-07 18:20:47 +08:00
parent 40d3d830a1
commit cd73613136
4 changed files with 8 additions and 10 deletions
-2
View File
@@ -226,7 +226,6 @@ NVCCFLAGS += -Wno-deprecated-gpu-targets \
-gencode arch=compute_35,code=compute_35 \
-gencode arch=compute_50,code=compute_50 \
-gencode arch=compute_61,code=compute_61 \
-gencode arch=compute_70,code=compute_70 \
-gencode arch=compute_75,code=compute_75 \
-DKCPP_LIMIT_CUDA_MAX_ARCH=750
@@ -234,7 +233,6 @@ else ifdef LLAMA_ARCHES_CU12
NVCCFLAGS += -Wno-deprecated-gpu-targets \
-gencode arch=compute_50,code=compute_50 \
-gencode arch=compute_61,code=compute_61 \
-gencode arch=compute_70,code=compute_70 \
-gencode arch=compute_75,code=compute_75 \
-gencode arch=compute_80,code=compute_80 \
-DKCPP_LIMIT_CUDA_MAX_ARCH=800