Harish Kasiviswanathan 6409b00f8f Add indirect io_links ...

Connect (Peer-to-Peer) GPUs that belong to same NUMA node.
Connect all [GPU] <--> [Non Parent NUMA] node

Change-Id: Ib4b08a6545d28b7dce4c9b1a90378bfc51bed07e


[ROCm/ROCR-Runtime commit: 7042292c60]

2016-03-10 15:11:17 -05:00

..

debug.c

Disable scratch Host allocation - via debug registration flags.

2016-02-10 07:52:32 -05:00

events.c

Fix race in dGPU event page setup

2016-02-28 07:14:23 -05:00

fmm.c

Add support for hsaKmtRegisterGraphicsHandleToNodes

2016-03-10 11:16:02 -05:00

fmm.h

Add support for hsaKmtRegisterGraphicsHandleToNodes

2016-03-10 11:16:02 -05:00

globals.c

Support gfx802 dGPU

2015-08-30 14:13:53 +03:00

libhsakmt.c

Move from direct call to kfd_ioctl to wrapping kmtIoctl

2014-12-29 15:37:42 +02:00

libhsakmt.h

Make hsaKmtAllocMemory more compliant with the Thunk spec

2016-02-10 10:29:54 -05:00

libhsakmt.ver

Add support for hsaKmtRegisterGraphicsHandleToNodes

2016-03-10 11:16:02 -05:00

Makefile

Link libhsakmt with -z nodelete

2016-01-22 18:08:19 -05:00

memory.c

Add support for hsaKmtRegisterGraphicsHandleToNodes

2016-03-10 11:16:02 -05:00

openclose.c

Don't limit number of supported HSA Nodes

2016-01-15 17:27:43 -05:00

perfctr.c

Don't limit number of supported HSA Nodes

2016-01-15 17:27:43 -05:00

pmc_table.c

gfx803: Add performance counter information

2016-02-12 16:39:39 -05:00

pmc_table.h

Add pmc table module

2014-07-29 13:58:01 +03:00

queues.c

Make hsaKmtAllocMemory more compliant with the Thunk spec

2016-02-10 10:29:54 -05:00

time.c

Move from direct call to kfd_ioctl to wrapping kmtIoctl

2014-12-29 15:37:42 +02:00

topology.c

Add indirect io_links

2016-03-10 15:11:17 -05:00

version.c

Move from direct call to kfd_ioctl to wrapping kmtIoctl

2014-12-29 15:37:42 +02:00