A. Unique TensorFlower c7e57df256 Use device handle instead of gpu ordinal in GpuVirtualMemAllocator for
configuring peer access.

Check that peers support virtual address management and have peer access
at GpuVirtualMemAllocator creation time.

PiperOrigin-RevId: 353124035
Change-Id: I8549253b84af35a664e9a6810c6b40deba20a532
2021-01-21 16:11:33 -08:00
..
2020-12-23 16:17:52 -08:00
2020-03-20 13:19:21 -07:00
2020-03-20 13:19:21 -07:00
2020-01-16 14:36:52 +09:00