GitLab

* Adding USE_HSA_DMA to switch to using hsa_amd_memory_async_copy in lieu of hipMemcpyAsync
* Adding USE_GPU_DMA for A2A benchmark
* Adding largeBAR check and fix for 0-hop GPU-CPU links

* Fixing potential out-of-bounds write during topology detection
* Fixing CU_MASK for multi-XCD GPUs
* Adding sub-iterations via NUM_SUBITERATIONS
* Adding support for variable subexecutor Transfers
* Adding healthcheck preset

* Adding pcopy benchmark, fixing CPU kernel on null destination

* Adding targeted DMA engine support
* Fixing CUDA compilation for H100

* v1.45 New GFX kernel

* v1.41 Adding schmoo benchmark, fixing timing reports for variable-iteration modes

* Adding xccID output to SHOW_ITERATIONS

* Adding a2a preset benchmark, fixing some gfx941 values

* Validating source directly, extra nullptr deallocation checks

* Adding source prep kernel (USE_PREP_KERNEL)
* Adding nvcc-only compilation path
* Fix for NVIDIA - set shared mem usage to 0 by default
* Updating default fill pattern for source data
* Restoring missing example.cfg file

* Adding MIMO support, DMA executor, Null memory type