Updates for 3.1 (#932)

This commit is contained in:
ANIKET SHIVAM
2023-04-29 06:34:27 -07:00
committed by GitHub
parent 6f8596ce3f
commit 7c04f95415
51 changed files with 1796 additions and 328 deletions

View File

@@ -227,4 +227,17 @@ elect_one_leader_sync()
#endif
}
// Store value to remote shared memory in the cluster
CUTE_DEVICE
void
store_shared_remote(uint32_t value, uint32_t smem_addr, uint32_t mbarrier_addr, uint32_t dst_cta_rank)
{
#if defined(CUTE_ARCH_CLUSTER_SM90_ENABLED)
uint32_t dsmem_addr = set_block_rank(smem_addr, dst_cta_rank);
uint32_t remote_barrier_addr = set_block_rank(mbarrier_addr, dst_cta_rank);
asm volatile("st.async.shared::cluster.mbarrier::complete_tx::bytes.u32 [%0], %1, [%2];"
: : "r"(dsmem_addr), "r"(value), "r"(remote_barrier_addr));
#endif
}
} // end namespace cute