ROCm Compile Time Reduction

[Spaarsh]

As per https://github.com/cp2k/dbcsr/discussions/933 discussion, this PR attempts to reduce compile time for rocm-enabled dbcsr builds with multiple target architectures.

[Spaarsh]

Currently, cmake successfully but make results in a failure with these errors:

In file included from /dbcsr/src/acc/cuda_hip/acc_mem.cpp:13:
/dbcsr/src/acc/cuda_hip/acc_mem.cpp: In function 'int c_dbcsr_acc_host_mem_allocate(void**, size_t, void*)':
/dbcsr/src/acc/cuda_hip/../hip/acc_hip.h:23:16: error: 'hipHostAllocDefault' was not declared in this scope; did you mean 'hipHostMallocDefault'?
   23 | #define ACC(x) hip##x
      |                ^~~
/dbcsr/src/acc/cuda_hip/acc_mem.cpp:49:23: note: in expansion of macro 'ACC'
   49 |   unsigned int flag = ACC(HostAllocDefault);
      |                       ^~~

[alazzaro]

Thanks for your contribution. However, given the current problems, I will keep this PR for a new release of DBCSR.