add missing architectures supported by CUDA 12.6 #39

Job	Run time
m (make, gpu_nvidia, mkl, gcc)	16m 50s
m (make, gpu_nvidia, mkl, intel)	17m 1s
m (make, gpu_nvidia, mkl-ilp64, gcc)	16m 46s
m (make, gpu_nvidia, mkl-ilp64, intel)	16m 54s
m (make, gpu_nvidia, openblas, gcc)	34m 19s
m (make, gpu_nvidia, openblas, intel)	27m 45s
m (cmake, gpu_nvidia, mkl, gcc)	25m 54s
m (cmake, gpu_nvidia, mkl, intel)	21m 45s
m (cmake, gpu_nvidia, openblas, gcc)	21m 47s
m (cmake, gpu_nvidia, openblas, intel)	26m 9s
m (cmake, gpu_amd, mkl, gcc)	26m 44s
m (cmake, gpu_amd, mkl, intel)	17m 48s
m (cmake, gpu_amd, openblas, gcc)	18m 12s
m (cmake, gpu_amd, openblas, intel)	17m 45s
	5h 5m 39s

Provide feedback