qmkl=parallel for vendors' GPU

[jinz2014]

Is is possible that users could simply add -qmkl=parallel running a SYCL program with BLAS functions on NVIDIA or AMD GPUs ?

For example,

icpx  -std=c++17 -Wall -fsycl --gcc-toolchain=  -fsycl-targets=nvptx64-nvidia-cuda -Xsycl-target-backend --cuda-gpu-arch=sm_90 -qmkl=parallel -O3 -DUSE_GPU -c main.cpp -o main.o
icpx  -std=c++17 -Wall -fsycl --gcc-toolchain=  -fsycl-targets=nvptx64-nvidia-cuda -Xsycl-target-backend --cuda-gpu-arch=sm_90 -qmkl=parallel -O3 -DUSE_GPU main.o -o main 
fatbinary warning : option 'image' has been deprecated
fatbinary warning : option 'image' has been deprecated
./main 268435456 1000

FP64 Dot
terminate called after throwing an instance of 'sycl::_V1::exception'
  what():  No kernel named _ZTSN6oneapi3mkl3gpu10l1_ker_usm23level1_reduction_kernelINS1_8usmMem_tIdLN4sycl3_V16access4modeE1024EEES9_NS4_IdLS8_1026EEESA_LNS2_10LEVEL1_APIE8ELl1ELl0ELNS2_11kernel_implE1ELl0EEE was found
make: *** [Makefile:93: run] Aborted (core dumped)