YuxinxinChen / ib_large_alloc_test

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

This test record a nvshmem error.

In this example, a class is created. In side the class, recv_queue and agg_queue are allocated by nvshmem_malloc. After the queues, a bunch of counters are allocated by nvshmem_malloc. However, when the size of queue is beyond certain size, nvshmem_uint32_p fails to update counters on remote GPU.

About


Languages

Language:Cuda 79.5%Language:Makefile 20.5%