Skip to main content
Log in

Mojo function

tcgen05_alloc

tcgen05_alloc[cta_group: SIMD[int32, 1]](ptr_tmem_addr: UnsafePointer[SIMD[uint32, 1], address_space=AddressSpace(3), alignment=16], num_cols: SIMD[uint32, 1])

Allocates tensor memory for use with tcgen05 instructions.

Note: This function is only available on NVIDIA Blackwell GPUs (SM 100+).

Parameters:

  • cta_group (SIMD[int32, 1]): The cooperative thread array (CTA) group ID.

Args:

  • ptr_tmem_addr (UnsafePointer[SIMD[uint32, 1], address_space=AddressSpace(3), alignment=16]): Shared memory pointer to hold tensor memory address.
  • num_cols (SIMD[uint32, 1]): The number of columns to allocate.

Was this page helpful?