AI/rocm-systems
2
0
Fork 0
Κώδικας Ζητήματα Pull Requests Δράσεις Πακέτα Έργα Κυκλοφορίες Wiki Δραστηριότητα
1.419 Υποβολές 3 Κλάδοι 0 Ετικέτες
09c4d50e508949c5c486ea0fc88aada9ca96e064
Γράφημα Υποβολών

724 Υποβολές

Συγγραφέας SHA1 Μήνυμα Ημερομηνία
Paul Emberson f7fb3392fb fix initOnceFunc setting incorrect result code (#1205) 
Addresses DMA-BUF support check unexpectedly failing

[ROCm/rccl commit: 435756af02]
 2024-06-07 16:47:19 -07:00
Nusrat Islam 7c029672a8 Merge pull request #1200 from nusislam/multi-node-256-fix 
graph: fix multi-node channel count

[ROCm/rccl commit: 9660e2e2dc]
 2024-06-07 14:34:20 -05:00
Nilesh M Negi 9b88e59ea4 Fix min_nchannels bug for gfx94* nranks=4 (#1202) 
Signed-off-by: nileshnegi <Nilesh.Negi@amd.com>

[ROCm/rccl commit: d9661c17e6]
 2024-06-07 14:31:28 -05:00
gilbertlee-amd efe851cfff Disabling NUMA maching for model 79 for some VM configs (#1204) 
[ROCm/rccl commit: 9b94a1052f]
 2024-06-06 17:15:04 -06:00
Nusrat Islam 3cbf67715f graph: restrict maxChannels to 64 for multi-node and RCCL_ENABLE_INTRANET=1 
[ROCm/rccl commit: 526cce9bf4]
 2024-06-06 10:58:41 -05:00
Nusrat Islam bf35178250 graph: fix multi-node minChannel count 
[ROCm/rccl commit: 6ab20a7c6b]
 2024-06-06 10:56:39 -05:00
Nusrat Islam 62cbb3bcdd set MIN_NCHANNEL limit to 64 for multi-node 
[ROCm/rccl commit: 9746d8ca3f]
 2024-06-03 13:05:05 -05:00
Nusrat Islam b34fd115a1 doubling debug buffer size with increased channels 
[ROCm/rccl commit: 0634c5c8e1]
 2024-06-03 13:05:05 -05:00
Nusrat Islam 48821ad0d7 set MAXCHANNELS to 128 
[ROCm/rccl commit: ef442f8f92]
 2024-06-03 13:05:05 -05:00
Nusrat Islam 2447161fb4 graph: restrict MAXCHANNELS for certain platforms 
[ROCm/rccl commit: 9f654f6cf5]
 2024-06-03 13:05:01 -05:00
Nusrat Islam ed2f96bc6a device: update the logic for channelId assignment 
[ROCm/rccl commit: 48859a97b1]
 2024-06-03 13:03:18 -05:00
Nusrat Islam b0b2aa1166 add 256 channels support 
[ROCm/rccl commit: 506f16c506]
 2024-06-03 13:03:18 -05:00
gilbertlee-amd a009e43f3c Changing channel stride for MI300X multinode (#1196) 
* Shuffling MI300X multi-node channels
* Updating tree channel logic

[ROCm/rccl commit: 0948eecbba]
 2024-06-03 10:00:55 -06:00
ClementLinCF 4f56aa5f8c Optimize NCHANNELS and MSCCL config for gfx942 80CUs (#1195) 
* Optimize NCHANNELS and MSCCL config for gfx942 80CUs

Set appropriately for different NCCL_MIN_NCHANNELS and MSCCL config,
potentially improving communication perf on the MI300x 80CUs

* Delete tools/msccl-algorithms/allreduce_1step_mccl_8_2_16777216_LL.xml

* Change the factor of gfx94 and update msccl config

[ROCm/rccl commit: cab25f919e]
 2024-06-01 07:07:46 -07:00
Nilesh M Negi 8518ef4afb [MSCCL]: Move scratch buffer debug msgs to TRACE (#1189) 
Signed-off-by: nileshnegi <Nilesh.Negi@amd.com>

[ROCm/rccl commit: 1249a6c3fd]
 2024-05-31 17:54:23 -05:00
gilbertlee-amd b35e2b8c4b Addressing possible out-of-bounds mem access during channel duplication (#1193) 
[ROCm/rccl commit: 354e0b29a6]
 2024-05-30 14:02:14 -06:00
Wenkai Du a724f1ebb7 Add ring simple chunk size tuning (#1180) 
* Add ring simple chunk size tuning

* modifying the tuning table to improve the performance of broadcast for 8MB to 32MB for single-node MI300X after ring simple chunk size tuning

* modifying the tuning table to improve the performance of reduce for 1MB to 4MB for single-node MI300X after ring simple chunk size tuning

---------

Co-authored-by: PedramAlizadeh <pmohamma@amd.com>

[ROCm/rccl commit: 73221b4230]
 2024-05-29 07:59:47 -07:00
Wenkai Du ae6c372406 Make WSL detection thread safe (#1178) 
* Make WSL detection thread safe

* Change static to beginning

* Switch to use atomics

[ROCm/rccl commit: 8f099b1adb]
 2024-05-28 17:23:50 -07:00
Edgar Gabriel 7f4bec7682 Merge pull request #1175 from edgargabriel/topic/alt_rsmi 
add alternative to rocm_smi_lib

[ROCm/rccl commit: a78c4f5e88]
 2024-05-28 09:36:55 -05:00
Wenkai Du bc29c89d9d Report error when collective is not enabled in build (#1177) 
* Report error when collective is not enabled in build

* Fix typo

[ROCm/rccl commit: eeea3b693b]
 2024-05-16 10:11:12 -07:00
Edgar Gabriel c31ac0d17b add alternative to rocm_smi_lib 
[ROCm/rccl commit: 9ad913bfa8]
 2024-05-14 13:51:41 -07:00
Wenkai Du 984f87a5d5 Support WSL2 (#1173) 
[ROCm/rccl commit: ecafc1969c]
 2024-05-10 07:31:12 -07:00
mberenjk 42d9f2458a Changing the failure print to reflect a bad installation of MSCCL (#1160) 
Co-authored-by: Marzieh Berenjkoub <mberenjk@amd.com>

[ROCm/rccl commit: 87d01e6bf5]
 2024-05-08 16:56:26 -05:00
Wenkai Du dd5d66d01a Use rocm-smi thread only mutex when available (#1169) 
[ROCm/rccl commit: a64aab5f63]
 2024-05-08 14:32:24 -07:00
Pedram Alizadeh 43d875b422 modifying the tuning table to improve the performance of broadcast for 1MB to 64MB for single-node MI300X (#1172) 
[ROCm/rccl commit: 73acf3eeec]
 2024-05-08 15:49:33 -04:00
mberenjk 00c0f3d67f Adding ASAN changes to address memory leak issue" (#1170) 
Co-authored-by: akolliasAMD <akollias@amd.com>

[ROCm/rccl commit: 408278209d]
 2024-05-08 09:16:00 -05:00
Wenkai Du 63d8de3bd3 Add compiler warning for uninitialized variable and fix (#1163) 
* Add compiler warning for uninitialized variable and fix

* Add -Wsometimes-uninitialized

* Convert warning to error

[ROCm/rccl commit: b18784d8b8]
 2024-05-08 07:00:25 -07:00
Wenkai Du eab9978e48 Use normal permute path when one NIC per GPU (#1171) 
[ROCm/rccl commit: f679db6ff6]
 2024-05-08 06:59:57 -07:00
Wenkai Du 0ff5fc0bad npkit: add broadcast trace (#1166) 
[ROCm/rccl commit: a0cef69110]
 2024-05-07 14:00:16 -07:00
paklui dd8e937948 fix typo for DMABUF_ENABLE 
[ROCm/rccl commit: 140b7dd40f]
 2024-05-06 13:27:50 -07:00
Wenkai Du c782aba364 Bypass NVIDIA Ampere related tuning (#1165) 
[ROCm/rccl commit: b513c3970a]
 2024-05-03 17:57:16 -07:00
Wenkai Du 7c811a7582 Fix ignore NUMA not being observed for NICs during model matching (#1164) 
[ROCm/rccl commit: bb58b1c258]
 2024-05-03 16:42:07 -07:00
Wenkai Du 9638535690 Fix build error when roctracer-dev package is not installed (#1161) 
[ROCm/rccl commit: 6f5a8ce1fb]
 2024-05-01 13:55:09 -07:00
Wenkai Du 3906e992f8 MSCCL: add support for out-of-place all reduce (#1156) 
[ROCm/rccl commit: 4e1b8c1cbb]
 2024-04-28 19:49:09 -07:00
Wenkai Du 703014e960 Add back tree simple chunk size tuning (#1157) 
[ROCm/rccl commit: cd6e840e0b]
 2024-04-28 19:48:53 -07:00
Nilesh M Negi b99b89e7a2 [GRAPH] Reduce NCCL_TOPO_MAX_NODES to 64 (#1153) 
Signed-off-by: nileshnegi <Nilesh.Negi@amd.com>

[ROCm/rccl commit: b90436d292]
 2024-04-27 23:41:11 -05:00
Tim afeaa17475 Merge pull request #1158 from AtlantaPepsi/NPKit_fix 
Prevent segfault from npkit-enabled rccl build

[ROCm/rccl commit: cc39e91c6f]
 2024-04-26 12:44:04 -04:00
AtlantaPepsi 8cf28704ce prevent segfault from npkit-enabled rccl build 
Signed-off-by: AtlantaPepsi <timhu102@amd.com>


[ROCm/rccl commit: 67246649ac]
 2024-04-26 10:54:27 -05:00
Wenkai Du 3c94f98688 Revert "Use relaxed atomics for LL on GFX11 (#859)" (#1148) 
This reverts commit 5983f0e371.

Use inline asm for 128b load on GFX11 for better peformance.

[ROCm/rccl commit: f330b82985]
 2024-04-26 07:49:55 -07:00
Shilei Tian 9a203f439c SWDEV-455705: Fix an UB that could lead to miscompilation (#1155) 
[ROCm/rccl commit: efe99057b0]
 2024-04-25 10:10:01 -07:00
Wenkai Du e494f29235 Replace __HIP_PLATFORM_HCC__ with __HIP_PLATFORM_AMD__ (#1154) 
[ROCm/rccl commit: 9e0c9b4ed8]
 2024-04-25 07:19:18 -07:00
BertanDogancay 36f9492cda Merge remote-tracking branch 'nccl/master' into develop 
[ROCm/rccl commit: e1a835910e]
 2024-04-23 13:34:00 -07:00
Wenkai Du 35f8d269f8 Use hipExtMallocWithFlags to allocate host memory on APU (#1149) 
Also use SM60 as CUDA compatibility level.

[ROCm/rccl commit: 220066197a]
 2024-04-17 16:56:38 -07:00
gilbertlee-amd 422a7ffcbb Rail optimization for rings (#1140) 
- Modifies the ring creation algorithm to be friendlier to rail-optimized topologies (should not affect classic fabric topologies)

[ROCm/rccl commit: 4cb62f999a]
 2024-04-15 12:03:57 -06:00
mberenjk da835cff9c replacing rccl_bfloat16 with hip_bfloat16 (#1126) 
Co-authored-by: mberenjk <mberenjk@amd.com>

[ROCm/rccl commit: 428837ffe4]
 2024-04-11 11:30:37 -05:00
Wenkai Du 99c7fc29ba NPKit: doubling size of event buffers following MAXCHANNELS change (#1135) 
[ROCm/rccl commit: 0ce68f21d4]
 2024-04-09 08:02:58 -07:00
Wenkai Du 0941d6bc6e Fix buffer overflow when parsing kernel cmdline (#1133) 
[ROCm/rccl commit: 137571fa01]
 2024-04-08 11:12:20 -07:00
Wenkai Du 43bbee4dcc Remove hipEventDisableSystemFence (#1122) 
There is no indication that disabling system fence has any latency improvement.
Removing it per recommendation from HIP.

[ROCm/rccl commit: 5976f757dd]
 2024-03-25 08:01:57 -07:00
FrankJ 894d3459a5 [bugfix]save undo plans in some case 
[ROCm/rccl commit: 9ef920a77b]
 2024-03-12 00:00:16 +08:00
Andy li e373bd44bf Enable fp8 support (#1101) 
* initial checkin

* resolve cr comments

* resolve the build issue

* fix the data correctless issue

* update fp8 header file and update the unit test for fp8 support

* remove fp16 from fp8 headers

* fix ut issue and catch up the latest code from develop

* udate according to cr comments

* update ut according to cr comments

* update num floats for each SumPostDiv from 4 to 6

* update fp8 header file name

* fix the typo

[ROCm/rccl commit: 6777e65c1d]
 2024-03-08 15:17:53 -08:00