The CI test machines only have 2 MPI slots. MPI tests were failing
when requesting 4 CPUs. Update these tests to request 2 CPUs.
[ROCm/rocprofiler-systems commit: aeb346b6d6]
-m64 only applies to x86-64, it's more reasonable to enable it only on x86_64 host.
It fixes build on some other platforms as well.
Provided by user r-value on github
https: //github.com/RadeonOpenCompute/rocminfo/pull/63
Change-Id: I9c1c40d3fa39b0a61d28041fe4998b5e1ad0cdcd
[ROCm/rocminfo commit: 17de0f9097]
-m64 only applies to x86-64, it's more reasonable to enable it only on x86_64 host.
It fixes build on some other platforms as well.
Provided by user r-value on github
https: //github.com/RadeonOpenCompute/rocminfo/pull/63
Change-Id: I9c1c40d3fa39b0a61d28041fe4998b5e1ad0cdcd
Single OCL context with multiple devices should use direct map
even with persistent memory, because cache coherency layer.
Change-Id: I02a48830440203a1dea176acd4dff4452b70f133
[ROCm/clr commit: 49037b9762]
Single OCL context with multiple devices should use direct map
even with persistent memory, because cache coherency layer.
Change-Id: I02a48830440203a1dea176acd4dff4452b70f133
- Make sure persistent memory from resource cache is properly adjusted
in free memory calculation.
Change-Id: I74ef68975ccde4694fb1cb904617c418e85dfc9f
[ROCm/clr commit: 85c15d720d]
- Make sure persistent memory from resource cache is properly adjusted
in free memory calculation.
Change-Id: I74ef68975ccde4694fb1cb904617c418e85dfc9f
Ignore the iGPU when iterating rocminfo and correct the bounds check
when matching results with HIP API. Also do not compare bytes directly,
use string to do a string compare as it can go out of bounds.
Disabling this test on windows till I get hold of a working windows
machine.
Change-Id: Icf811e0310776567afecea92e1d54e8438b39235
[ROCm/hip-tests commit: af3ac29c40]
Ignore the iGPU when iterating rocminfo and correct the bounds check
when matching results with HIP API. Also do not compare bytes directly,
use string to do a string compare as it can go out of bounds.
Disabling this test on windows till I get hold of a working windows
machine.
Change-Id: Icf811e0310776567afecea92e1d54e8438b39235
Occationally "Memory access fault by GPU node-8 (Agent handle: 0x23a5640) on address 0x7f461ec00000. Reason: Page not present or supervisor privilege" can be seen from gfx906 CI
[ROCm/rccl commit: e5bf56c6d8]
Occationally "Memory access fault by GPU node-8 (Agent handle: 0x23a5640) on address 0x7f461ec00000. Reason: Page not present or supervisor privilege" can be seen from gfx906 CI
1) Move needsProxy initialization before mscclSetupConnections since the latter
will revise it later.
2) Remove mscclAvailable check in net.cc since it's no more required and caused
non-shared buffer allocated for MSCCL which is not expected.
[ROCm/rccl commit: 70bbeb4773]
1) Move needsProxy initialization before mscclSetupConnections since the latter
will revise it later.
2) Remove mscclAvailable check in net.cc since it's no more required and caused
non-shared buffer allocated for MSCCL which is not expected.
Adds support and implement APIs for 'gpu_metrics_v1_5'
Code changes related to the following:
* gpu metrics 1.5 support
* Unit tests
* Examples
Build changes related to the following: None
Change-Id: Ie8917dd63c1dd1a94467b100fa44b634cebe62b6
Signed-off-by: Oliveira, Daniel <daniel.oliveira@amd.com>
[ROCm/rocm_smi_lib commit: 373621aed3]
Adds support and implement APIs for 'gpu_metrics_v1_5'
Code changes related to the following:
* gpu metrics 1.5 support
* Unit tests
* Examples
Build changes related to the following: None
Change-Id: Ie8917dd63c1dd1a94467b100fa44b634cebe62b6
Signed-off-by: Oliveira, Daniel <daniel.oliveira@amd.com>
Ben Richard
David Yat Sin
German
Wenkai Du
Aryan Salmanpour
sdashmiz
Jatin Chaudhary
Giovanni LB
Rahul Garg
Ranjith Ramakrishnan
ksankisa
Ziyue Yang
Benjamin Welton
Oliveira, Daniel