[CK Tile] Need GroupGEMM with make_argument on device

[zjing14]

Currently, CK Tile GroupGEMM prepares metadata on the host, which requires transferring meta data between the device and host back and forth.

composable_kernel/include/ck_tile/ops/gemm/kernel/grouped_gemm_kernel.hpp

Line 98 in 6b6fcd3

CK_TILE_HOST static auto MakeKargs(const std::vector<GroupedGemmHostArgs>& gemm_descs)

To avoid this overhead, we need groupGEMM kernel as old CK, which is a persistent kernel that reads GEMM shapes from device memory, and calculate offset and block_id on-the-fly.

[zjing14]

@aosewski Could you help? Move your groupGEMM design from old CK to CK Tile