Device memory allocation in rocBLAS

rocBLAS uses per-handle device memory allocation to manage temporary memory efficiently. Each handle maintains its own memory and executes kernels sequentially in a single stream, allowing memory reuse across kernels.

There are two memory allocation schemes:

• rocBLAS_managed(default): By default, rocBLAS internally manages memory, allocating more if needed. Allocated memory persists with the handle for reuse.

• user_owned: Users allocate memory and provide it to rocBLAS via rocblas_set_workspace.

rocBLAS_managed is the default scheme. This scheme uses hipMallocAsync and hipFreeAsync (stream-order allocation) to allocate and free memory in stream order, avoiding global synchronization. This enables seamless stream switching without needing hipStreamSynchronize().

The following computational functions use temporary device memory.

Function	Use of temporary device memory
L1 reduction functions rocblas_Xasum rocblas_Xasum_batched rocblas_Xasum_strided_batched rocblas_Xdot rocblas_Xdot_batched rocblas_Xdot_strided_batched rocblas_Xmax rocblas_Xmax_batched rocblas_Xmax_strided_batched rocblas_Xmin rocblas_Xmin_batched rocblas_Xmin_strided_batched rocblas_Xnrm2 rocblas_Xnrm2_batched rocblas_Xnrm2_strided_batched rocblas_dot_ex rocblas_dot_batched_ex rocblas_dot_strided_batched_ex rocblas_nrm2_ex rocblas_nrm2_batched_ex rocblas_nrm2_strided_batched_ex	Reduction array
L2 functions rocblas_Xgemv (optional) rocblas_Xgemv_batched rocblas_Xgemv_strided_batched rocblas_Xtbmv rocblas_Xtbmv_batched rocblas_Xtbmv_strided_batched rocblas_Xtpmv rocblas_Xtpmv_batched rocblas_Xtpmv_strided_batched rocblas_Xtrmv rocblas_Xtrmv_batched rocblas_Xtrmv_strided_batched rocblas_Xtrsv rocblas_Xtrsv_batched rocblas_Xtrsv_strided_batched rocblas_Xhemv rocblas_Xhemv_batched rocblas_Xhemv_strided_batched rocblas_Xsymv rocblas_Xsymv_batched rocblas_Xsymv_strided_batched rocblas_Xtrsv_ex rocblas_Xtrsv_batched_ex rocblas_Xtrsv_strided_batched_ex	Result array before overwriting input Column reductions of skinny transposed matrices applicable for gemv functions
L3 GEMM-based functions rocblas_Xtrsm rocblas_Xtrsm_batched rocblas_Xtrsm_strided_batched rocblas_Xsymm rocblas_Xsymm_batched rocblas_Xsymm_strided_batched rocblas_Xsyrk rocblas_Xsyrk_batched rocblas_Xsyrk_strided_batched rocblas_Xsyr2k rocblas_Xsyr2k_batched rocblas_Xsyr2k_strided_batched rocblas_Xsyrkx rocblas_Xsyrkx_batched rocblas_Xsyrkx_strided_batched rocblas_Xtrmm rocblas_Xtrmm_batched rocblas_Xtrmm_strided_batched rocblas_Xhemm rocblas_Xhemm_batched rocblas_Xhemm_strided_batched rocblas_Xherk rocblas_Xherk_batched rocblas_Xherk_strided_batched rocblas_Xher2k rocblas_Xher2k_batched rocblas_Xher2k_strided_batched rocblas_Xherkx rocblas_Xherkx_batched rocblas_Xherkx_strided_batched rocblas_Xgemm rocblas_Xgemm_batched rocblas_Xgemm_strided_batched rocblas_gemm_ex rocblas_gemm_ex_batched rocblas_gemm_ex_strided_batched rocblas_Xtrtri rocblas_Xtrtri_batched rocblas_Xtrtri_strided_batched	Block of matrix

Memory allocation functions

rocBLAS includes functions for manually setting the memory size and determining the memory requirements.

Function for setting a user-owned workspace

• rocblas_set_workspace

Functions for determining memory requirements

• rocblas_start_device_memory_size_query

• rocblas_stop_device_memory_size_query

• rocblas_is_managing_device_memory

See the API section for information about these functions.

rocBLAS function return values for insufficient device memory

If the user manually allocates (user-owned scheme) using rocblas_set_workspace(rocblas_handle handle, void* addr, size_t size), that size is used as the limit and no resizing or synchronizing ever occurs. The following two function return values indicate insufficient memory:

• rocblas_status == rocblas_status_memory_error : indicates there is insufficient device memory for a rocBLAS function.

• rocblas_status == rocblas_status_perf_degraded : indicates that a slower algorithm was used because of insufficient device memory for the optimal algorithm.

Switching streams without synchronization

Stream-order memory allocation lets the application switch streams without having to call hipStreamSynchronize().