AI/rocm-systems
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
13,501 Commits 3 Branches 0 Tags
e56ff3716d7b8d04ee60e42d2ebb6ebf5d0d58db
Commit Graph

13501 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Evgeny 3443b09c49 minor fix, hsa_ven_amd_aqlprofile_pfn_t alias 
Change-Id: I4f712cf39e577498da25e37323fcf847e39bfa19
 2018-09-27 17:22:46 -05:00
Evgeny 8e873e628b minor fix, hsa_ven_amd_aqlprofile_pfn_t alias 
Change-Id: I4f712cf39e577498da25e37323fcf847e39bfa19


[ROCm/rocprofiler commit: 3443b09c49]
 2018-09-27 17:22:46 -05:00
Evgeny 01aa4c4ffb moving to non-deprecated get extension table HSA API 
Change-Id: I356c6406eb5c2a1755202dc562298b30cb7acc06
 2018-09-26 13:20:36 -05:00
Evgeny da6e7f7da4 moving to non-deprecated get extension table HSA API 
Change-Id: I356c6406eb5c2a1755202dc562298b30cb7acc06


[ROCm/rocprofiler commit: 01aa4c4ffb]
 2018-09-26 13:20:36 -05:00
Evgeny fdbe277f2a hsa_ven_amd_aqlprofile_pfn_t alias 
Change-Id: Ia4a67ef0d2f8975f0e541e85c215afec76e9de5f
 2018-09-26 14:10:21 -04:00
Gang Ba 52ec7f805e drm/amdkfd: Added gfx904 and gfx803 for KFD. 
Change-Id: I4406dc70c776926feaecca3f2146d65259a80517
Signed-off-by: Gang Ba <gaba@amd.com>
 2018-09-25 08:17:44 -04:00
Mike Li 3144a84b9a all_gpu_id_array: Handle GPU resource management 
GPU Resource management can disable some of the GPU nodes.
The Kernel driver could be not aware of this.
Get from Kernel driver information of all the nodes and then filter it.

Change-Id: I4eeb126a5efce2192c35f5d2b72be1811e9ded32
Signed-off-by: Mike Li <Tianxinmike.Li@amd.com>
 2018-09-24 11:38:11 -04:00
Mike Li c3b47c0959 kfdtest: Handle GPU resource management 
Currently the FindDRMRenderNode function will access the sysfs
directly to find the render node. It doesn't work with the
GPU management changes. Have changed code to call hsaKmtGetNodeProperties
instead.

Change-Id: I3bb537a323bc1e8c49f38d8aabc60c13e268aecd
Signed-off-by: Mike Li <Tianxinmike.Li@amd.com>
 2018-09-24 11:38:11 -04:00
Mike Li f9bd960344 Output a error message only when open_drm_render_device failed unexpectedly. 
Change-Id: I5b9587a8d5c7a900e9ab8611a25d0c49d34b4cef
Signed-off-by: Mike Li <Tianxinmike.Li@amd.com>
 2018-09-24 11:36:11 -04:00
xinhui pan 918a45a430 kfdtest: add P2POverheadTest 
This is to measure the laterncy + overhead of sdma packet
consumption on p2p.
It is Similar with QueueLatency test. What's more, the queue's overhead
with different workload show more details.

test result on two gfx900.
[ RUN      ] KFDPerformanceTest.P2POverheadTest
[          ] Test (avg. ns) | Size	4	8	16	64	256	1024
[          ] -----------------------------------------------------------------------
[          ] [push]     [1 -> 0]	333	148	185	111	148	148
[          ] [push]     [1 -> 1]	370	222	333	74	148	111
[          ] [push]     [1 -> 2]	333	148	148	148	148	148
[          ] [push]     [2 -> 0]	111	333	259	148	148	148
[          ] [push]     [2 -> 1]	222	148	185	148	148	148
[          ] [push]     [2 -> 2]	222	111	370	111	74	148
[          ] [pull]     [1 -> 0]	370	296	296	148	185	148
[          ] [pull]     [1 -> 1]	185	333	222	148	222	148
[          ] [pull]     [1 -> 2]	222	444	259	148	185	111
[          ] [pull]     [2 -> 0]	148	148	148	148	148	148
[          ] [pull]     [2 -> 1]	148	148	148	148	148	148
[          ] [pull]     [2 -> 2]	185	148	148	74	222	296
[          ] [push|pull][1 -> 0]	1259	1222	1259	1074	1037	962
[          ] [push|pull][1 -> 1]	1037	1037	1037	740	740	1000
[          ] [push|pull][1 -> 2]	1259	1259	1296	1037	1000	1074
[          ] [push|pull][2 -> 0]	1037	1037	1037	1074	1037	1148
[          ] [push|pull][2 -> 1]	1037	1037	1037	1037	925	1074
[          ] [push|pull][2 -> 2]	666	666	740	740	703	925
[       OK ] KFDPerformanceTest.P2POverheadTest (459 ms)

Change-Id: I422263cb52f7ce184f6f1ff4466d04c239fbe9c9
Signed-off-by: xinhui pan <xinhui.pan@amd.com>
 2018-09-24 09:28:00 -04:00
Scott Linder 47f0e6f7d3 Apply dynamic relocations for STT_FUNC symbols 
Required to support function calls through GOT table.

Change-Id: I174a0269fdd67369d38fe41855b7bd01f350b839
 2018-09-23 21:42:32 -04:00
Gregory Stoner d72d7eb5a5 Update README.md 2018-09-23 20:35:10 -05:00
Gregory Stoner 1b12c8c981 Update README.md 2018-09-23 20:34:22 -05:00
Harish Kasiviswanathan fb79a0efe2 Topology: Use processors available to the process 
The existing call sysconf (_SC_NPROCESSORS_ONLN) provides the number of
processors available to the scheduler. When a KFD process is run under a
container environment, only a subset (cpuset) of processors are
available to the current process.

For getting CPU cache information use sched_getaffinity() to get the
number of processors available to the current process.

Change-Id: Ieac02f1f61c17e24ac34ba502968c69d3bc631cb
Signed-off-by: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>
 2018-09-21 10:31:59 -04:00
Evgeny Shcherbakov e56fc0adda Merge "standalone/sampling mode" into amd-master 2018-09-19 16:33:09 -04:00
Evgeny Shcherbakov e16847ef21 Merge "standalone/sampling mode" into amd-master 
[ROCm/rocprofiler commit: e56fc0adda]
 2018-09-19 16:33:09 -04:00
xinhui pan e5a541eaf2 kfdtest: Add P2P bandwidth test 
The test measures the bandwidth between GPUs. Currently we do not
care numa topology as some products really support across PCI-e root
complex p2p.

test result on two gfx900 system.
[ RUN      ] KFDPerformanceTest.P2PBandWidthTest
[          ] Copy from node to node by [push, NONE]
[          ] [1 -> 0] 6.13477 - 6.12695 GB/s
[          ] [1 -> 2] 3.77734 - 3.76855 GB/s
[          ] [2 -> 0] 6.67676 - 6.6543 GB/s
[          ] [2 -> 1] 6.14453 - 6.12793 GB/s
[          ] Copy from node to node by [pull, NONE]
[          ] [1 -> 0] 6.10547 - 6.08105 GB/s
[          ] [1 -> 2] 9.65527 - 9.65039 GB/s
[          ] [2 -> 0] 6.49805 - 6.4873 GB/s
[          ] [2 -> 1] 8.95508 - 8.85254 GB/s
[          ] Full duplex copy from node to node by [push|pull, NONE]
[          ] [1 -> 0] 11.0986 - 11.0986 GB/s
[          ] [1 -> 2] 7.54297 - 7.54297 GB/s
[          ] [2 -> 0] 12.0264 - 11.9639 GB/s
[          ] [2 -> 1] 12.0469 - 12.0371 GB/s
[          ] Full duplex copy from node to node by [push, push]
[          ] [1 <-> 2] 11.7324 - 11.4541 GB/s
[          ] Full duplex copy from node to node by [pull, pull]
[          ] [1 <-> 2] 11.4824 - 11.0508 GB/s
[          ] Copy from node to multiple nodes by [push, NONE]
[          ] [1 -> [0...2]] 5.625 - 5.73633 GB/s
[          ] [2 -> [0...2]] 6.45801 - 6.4707 GB/s
[          ] Copy from multiple nodes to node by [push, NONE]
[          ] [[1...2] -> 0] 12.8379 - 12.2578 GB/s

Now we can get more timestamp info like below.

Copy from node to node by [push, NONE]
[1 -> 0]
[1 : 0] #-#-#-#-#-#-#-#-##-#-#-#-#-#-#-#-#-##-#-#-#-#-#-#-#-##-#-#-#-#-#-#-#-#-##-#-#-#-#-#-#-#-##-#-#-#-#-###############################
[1 : 1] ####################################################################################################
[1 -> 2]
[1 : 0] #--#-#-#-#-#--#-#-#-#-#--#-#-#-#-#--#-#-#-#-#-#--#-#-#-#-#--#-#-#-#-#--#-#-#-#-#-#--#-#-#-#-#--#-#-######################################
[1 : 1] ##################################################################################################-#
[2 -> 0]
[2 : 0] ##-###-##-###-###-##-###-##-###-###-##-###-###-##-###-###-##-###-##-###-###-##-###-###-##-###-###-#################
[2 : 1] ###############################################################################-#############-###-##
[2 -> 1]
[2 : 0] ##-##-##-##-##-###-##-##-##-##-##-###-##-##-##-##-###-##-##-##-##-##-###-##-##-##-##-###-##-##-##-####################
[2 : 1] ################################################################################-###-############-##

[snip]

Full duplex copy from node to node by [push, push]
[1 <-> 2]
[1 : 0] #-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-##-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-##-#-#-#-#-#-#-#-#-#-#-#-####################################
[1 : 1] ################-###################################################-############-####-#############
[2 : 2] #-##-##-##-#-##-##-##-##-#-##-##-##-##-#-##-##-##-##-#-##-##-##-##-#-##-##-##-##-##-#-##-##-##-##-#-##-##-##-##-##-#-##################
[2 : 3] #####-######-#####-######-#####-######-#####-######-#####-######-#####-######-#####-######-#####-######-#####-#####-##
Full duplex copy from node to node by [pull, pull]
[1 <-> 2]
[1 : 0] ######################################################################-##-#-###############-####-###
[1 : 1] #-#-#-##-#-#-##-#-#-##-#-#-##-#-#-##-#-#-##-#-#-##-#-#-#-##-#-#-##-#-#-##-#-#-##-#-#-##-#-#-##-#-#-############################
[2 : 2] ##-##-##-##-###-##-##-##-##-###-##-##-##-###-##-##-##-##-###-##-##-##-###-##-##-##-##-###-##-##-##-##-###-##-##-##-###-##-##-############
[2 : 3] #-#-#-#-#-#-#-#-#-#-#-#-#-##-#-#-#-#-#-#-#-#-#-#-#-#-#-##-#-#-#-#-#-#-#-#-#-#-#-#-#-##-#-#-#-#-#-#-#-#-#-#-#-#-##-#-#-#-#-#-#-#########-#############
Copy from node to multiple nodes by [push, NONE]
[1 -> [0...2]]
[1 : 0] #-#--#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-#-###############################
[1 : 1] ########################################################################################-###-###-###
[2 -> [0...2]]
[2 : 0] ##-##-##-###-##-###-##-##-###-##-###-##-##-###-##-###-##-###-##-##-###-##-###-##-##-###-##-###-##-##################
[2 : 1] -################################################################################################-##
Copy from multiple nodes to node by [push, NONE]
[[1...2] -> 0]
[1 : 0] #-#-#-#-#-#-#-#-##-#-#-#-#-#-#-#-#-##-#-#-#-#-#-#-#-##-#-#-#-#-#-#-#-#-##-#-#-#-#-#-#-#-#-##-#-#-#-###############################
[1 : 1] ################################################################################################-#-#
[2 : 2] ##-##-##-###-##-##-###-##-##-##-###-##-##-###-##-##-###-##-##-###-##-##-##-###-##-##-###-##-##-###-##-##################
[2 : 3] #########################-#########################-#########################-#########################
[       OK ] KFDPerformanceTest.P2PBandWidthTest (15982 ms)

Change-Id: Ia90044191d51650ccb220476d31fb317aa3ad6ce
Signed-off-by: xinhui pan <xinhui.pan@amd.com>
 2018-09-19 12:03:05 +08:00
xinhui pan f618b3f075 kfdtest: add KFDTestUtilQueue 
Some infrastructures below,
Implement SdmaTimePacket which records the global GPU timestamp.

Introduce class AsyncMPSQ and AsyncMPMQ.
AsyncMPSQ is aka async multiple packet single queue. It takes a set of
packet when create and submits them to a GPU to run. While AsyncMPMQ is
aka async multiple packet multiple queue. It manages a set of AsyncMPSQ,
and use a forloop to do operations of AsyncMPSQ.

Implement sdma_multicopy helper functions.

Change-Id: I47e1d2ca9630113b2a1d85a0055f3f8ee629fb5f
Signed-off-by: xinhui pan <xinhui.pan@amd.com>
 2018-09-19 12:03:05 +08:00
Evgeny b481bfe5de standalone/sampling mode 
Change-Id: I24c358bfebacb5678fb3d31f2f94ae867a1c9ac9
 2018-09-18 20:33:41 -05:00
Evgeny 443b15bc97 standalone/sampling mode 
Change-Id: I24c358bfebacb5678fb3d31f2f94ae867a1c9ac9


[ROCm/rocprofiler commit: b481bfe5de]
 2018-09-18 20:33:41 -05:00
Ramesh Errabolu 01eea21d6c Capture number of Numa Nodes present on system 
Change-Id: Ic789a6b9da8e316cb483e50b0fe9faa03798f97c
 2018-09-18 16:27:30 -05:00
Xiaojie Yuan 247fa9f1e0 Use 'RecordProperty' to record performance scores 
For following test cases:
- KFDQMTest.QueueLatency
- KFDQMTest.BasicCuMaskingLinear
- KFDQMTest.BasicCuMaskingEven
- KFDMemoryTest.MMBandWidth
- KFDMemoryTest.MMapLarge
- KFDMemoryTest.MMBench

v2: xml element cannot start with a number, so change the key name of
    MMBandWidth and MMBench accordingly
    xml element cannot contain whitespaces, so trim whitespaces in "VRAM  "
v3: introduce KFDLog-like way to use KFDRecord

Change-Id: Ifc3ed5657621252a7b39dccf1ef4f50a92593f77
Signed-off-by: Xiaojie Yuan <xiaojie.yuan@amd.com>
 2018-09-18 17:41:14 +08:00
Ramesh Errabolu f007870792 ROCr changes to enable small BAR P2P over xGMI 
Change-Id: I6aaa3fe2565cdf7e15d58a7484d6bd5916ffff64
 2018-09-17 22:54:40 -04:00
Evgeny 81532bb6f5 VERSION_MINOR macro typo fix 
aqlprofile info ENABLE_CMD enum adding;

Change-Id: I7b19082144d2bd0bf7af7ddc282358168b225759
 2018-09-17 20:49:47 -04:00
Evgeny f1198b11bd adding rocprof link 
Change-Id: Idb37a39a19d0e75b7867fb390504596d67df6436
 2018-09-17 13:26:32 -05:00
Evgeny ad60f25d92 adding rocprof link 
Change-Id: Idb37a39a19d0e75b7867fb390504596d67df6436


[ROCm/rocprofiler commit: f1198b11bd]
 2018-09-17 13:26:32 -05:00
Gregory Stoner 88e3eadac7 Update README.md 2018-09-16 12:33:37 -05:00
Gregory Stoner 959468d0b6 Update README.md 2018-09-16 12:32:33 -05:00
cfreehill 65ee227637 Merge pull request #1 from RadeonOpenCompute/roc-1.9.x 
ROCm 1.9.0
 2018-09-16 00:17:05 -05:00
Chris Freehill 53b7ae1113 ROCm 1.9.0 update 2018-09-16 00:13:29 -05:00
xinhui pan a6287ba919 kfdtest: Do not set GTEST_FLAG throw_on_failure 
This change is from commit 62f7dc2a("kfdtest: Do not set GTEST_FLAG
throw_on_failure").
But it is unexpected to reverted by commit 414042ab("kfdtest: Clean up
comments"). So add this change back.

Fix: 414042ab

Change-Id: Ia9e99c9ca17b99aab62b4db55017018ddae43dfb
Signed-off-by: xinhui pan <xinhui.pan@amd.com>
 2018-09-11 10:25:56 +08:00
xinhui pan 07bd97a864 kfdtest: Fix queuelatency fail issue 
The timestamp written by releaseMemory packet might still not be visible
when we fetch it.
To fix this bug, use event-based wait.

Change-Id: If2324eb3b3a632c711ee4dff4d03a93d5306c289
Signed-off-by: xinhui pan <xinhui.pan@amd.com>
 2018-09-10 21:17:29 -04:00
Felix Kuehling be574169c1 libhsakmt: Fix segfault on gfx801 
Handle the case that svm.dgpu_aperture does not exist in vm_find_object.

Change-Id: Ic0983d4f321f1b6248514f2fa25162976e90bd75
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
 2018-09-10 14:39:05 -04:00
Harish Kasiviswanathan 1fda429726 kfdtest: GetNodeIoLinkProperties: Display NodeFrom 
Use the NodeFrom returned by hsaKmtGetNodeIoLinkProperties() to check
its correctness.

Change-Id: I6ce436dc7c5d5b192bee21156292bd3eff77f916
Signed-off-by: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>
 2018-09-10 09:44:24 -04:00
Harish Kasiviswanathan 7876bb70a9 Add cgroup support 
Some nodes are unavailable based on the task's cgroup hierarchy. Handle
this situation by ignoring those nodes

Change-Id: I72f9e822d2ec8cf15732df95e427d5549a75b55d
Signed-off-by: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>
 2018-09-06 16:56:32 -04:00
Harish Kasiviswanathan 866ef20054 iolinks: Handle GPU resource management 
With GPU resource management, some nodes are unavailable based on the
cgroup hierarchy of the task. Kernel via sysfs specifies all the
iolinks. Skip the links which are not accessible.

Also iolinks specified by the kernel refer to sysfs Node IDs. Map it to
relevant user Node IDs

v2: NodeFrom mapped from sysfs Node to User Node

Change-Id: I95312ee6ca51b89fe9e6ca2a9185c2ea1e94afc4
Signed-off-by: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>
 2018-09-06 16:56:07 -04:00
Harish Kasiviswanathan f84a99e953 Replace global variable _system with g_system 
Change-Id: I452090473a5b46b32204f7f916bdcfdd3e8a47bd
Signed-off-by: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>
 2018-09-06 16:56:07 -04:00
Evgeny 093caf20c5 fix reading of the input counters line without end-of-line symboll 
Change-Id: Idfa9c18d08af8308837124391885f35213c8a7e2
 2018-09-05 15:50:31 -05:00
Evgeny 8fa5f53f35 fix reading of the input counters line without end-of-line symboll 
Change-Id: Idfa9c18d08af8308837124391885f35213c8a7e2


[ROCm/rocprofiler commit: 093caf20c5]
 2018-09-05 15:50:31 -05:00
Sean Keely 3357cadeec Check fill addresses for alignment. 
Check was documented but missing.

Change-Id: I97951635d794fd22e20c25d20e9d0e35035254af
 2018-09-05 16:34:19 -04:00
xinhui pan 9c7cfc0df2 kfdtest: Add event-based synchronization mechanism to queues 
Wait4PacketConsumption now can accept an event to wait all packets subbmitted
to be processed.

Change-Id: I1497b7704e892b04d05811b8d3e4742237c1be57
Signed-off-by: xinhui pan <xinhui.pan@amd.com>
 2018-09-04 21:21:19 -04:00
Felix Kuehling a9bd6e6f8b Revert "libhsakmt: Try to use CPU addr as GPU addr for userptrs" 
This reverts commit ab181c46c0.

This fixes ambiguity when looking up GPU addresses with
hsaKmtQueryPointerInfo.

hsa_amd_agents_allow_access uses hsaKmtQueryPointerInfo, and
depends on finding the correct object from a GPU address. Finding
the wrong userptr object based on its CPU address leads to
incorrect GPU mappings and results in VM faults.


Change-Id: I7c5f571ee6e1f9d32687aa3eab6d96944ad032be
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
 2018-08-31 15:04:50 -04:00
Felix Kuehling 608dddbe9d kfdtest: Fix gfx902 blacklist 
Removed some tests from the blacklist that are now passing. Added two
new tests that hang the GPU.

Change-Id: I09e729590e5181311375058be492d387342ba2fe
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
 2018-08-31 15:04:50 -04:00
Felix Kuehling 855f1a32a9 libhsakmt: Fix and deduplicate object lookup code 
Added a helper vm_find_object that can be used everywhere we need to
lookup objects by their address and optionally size. This unifies
all subtly different, partially incomplete, or broken ways of doing
this in various functions:

* map
* unmap
* register
* deregister
* free
* get_mem_info
* set_mem_user_data

At the same time fix some subtle problems for userptr lookup that
got a bit more complex when the userptr address can match the GPU
address.


Change-Id: I98572d1734fc7688a1d68f6a784e02c8dee90af5
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
 2018-08-31 15:04:47 -04:00
Sean Keely 2843988dd7 Remove redundant initialization. 
LinkInfo is already initialized to zero in its default constructor.

Change-Id: Ifa4fb886cce9b474c6879c9c82744044ab394082
 2018-08-29 19:36:07 -04:00
Sean Keely 56ed5c8904 Refactor blocking sdma commands. 
Remove fence pool and use two signals.  Two signals allows overlapped
submission and copy while reducing thread busy polling.

Change-Id: Idb5f8e4c7f482a596ffce9e7799191fdd785a216
 2018-08-29 19:13:23 -04:00
Sean Keely e0839ab27e Implement SDMA copy rect for gfx9. 
Fix pitch overflow due to small element detection.
Add wide pitch 2D copy handling.
Cleanup code duplication.

Change-Id: I93b1584aba8e5964957eb7ab3544df806ca3e2f9
 2018-08-29 19:13:07 -04:00
shaoyunl 30a4ab39f3 thunk: Avoid create PCIe indirect link on none large bar target 
PCIe P2P (indirect) IOLinks should only be created if the remote GPU
is large-BAR

Change-Id: I55cbb5e37c5d41267583e07aca6bdcc708403029
Signed-off-by: shaoyunl <Shaoyun.Liu@amd.com>
 2018-08-29 16:31:55 -04:00
Shaoyun Liu 7796994f46 Thunk: Avoid add indirect link for the GPUS with xGMI link 
Change-Id: I06f511c55e28919512fda79b504566818dc2a5ab
Signed-off-by: Shaoyun Liu <Shaoyun.Liu@amd.com>
 2018-08-29 13:22:58 -04:00
xinhui pan a040a24243 kfdtest: Let BigBufferStressTest detect memory leak 
As it will alloc as much as small system memory to reach the allocation limit.
We can try to alloc memory several times to see if any allocation in
the previous step cause memory leak.

Also we test if GPU can access these memory correctly or not.

Change-Id: I309f9821b6bc99c212a6bfbc21fe3086ab589fd3
Signed-off-by: xinhui pan <xinhui.pan@amd.com>
 2018-08-28 22:50:42 -04:00