8000 PyTorch 2.2.0 release Triton distribution not found in nightly · Issue #120044 · pytorch/pytorch · GitHub
[go: up one dir, main page]
Skip to content

PyTorch 2.2.0 release Triton distribution not found in nightly #120044

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Closed
woongjin-c opened this issue Feb 16, 2024 · 9 comments
Closed

PyTorch 2.2.0 release Triton distribution not found in nightly #120044

woongjin-c opened this issue Feb 16, 2024 · 9 comments
Assignees
@atalman
Labels
module: build Build system issues needs reproduction Someone else needs to try reproducing the issue given the instructions. No action needed from user triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module

Comments

@woongjin-c
Copy link
woongjin-c commented Feb 16, 2024 •
edited by pytorch-bot bot
Loading

🐛 Describe the bug

When building PyTorch v2.2.0 from source, "make triton" cannot find matching distribution for Triton:

pip3 uninstall -y triton
WARNING: Skipping triton as it is not installed.
Looking in indexes: https://download.pytorch.org/whl/nightly/
ERROR: Could not find a version that satisfies the requirement pytorch-triton==2.2.0+bcad9dabe1 (from versions: 2.0.0+0d7e753227, 2.0.0+3aa3d7024e, 2.0.0+af76c989eb, 2.0.0+b8b470bc59, 2.0.0+c8bfe3f548, 2.0.0+d54c04abe2, 2.1.0, 2.1.0+2c32f43999, 2.1.0+3c400e7818, 2.1.0+440fd1bf20, 2.1.0+46672772b4, 2.1.0+6e4932cda8, 2.1.0+7d1a95b046, 2.1.0+9820899b38, 2.1.0+9e3e10c5ed, 2.1.0+bcad9dabe1, 2.1.0+e6216047b8, 2.1.0+e650d3708b, 2.2.0+e28a256d71, 3.0.0+901819d2b6)
ERROR: No matching distribution found for pytorch-triton==2.2.0+bcad9dabe1
make: *** [Makefile:37: triton] Error 1

Versions

PyTorch version: N/A
Is debug build: N/A
CUDA used to build PyTorch: N/A
ROCM used to build PyTorch: N/A

OS: Ubuntu 22.04.3 LTS (x86_64)
GCC version: (conda-forge gcc 12.1.0-17) 12.1.0
Clang version: Could not collect
CMake version: version 3.22.1
Libc version: glibc-2.35

Python version: 3.11.0 (main, Mar 1 2023, 18:26:19) [GCC 11.2.0] (64-bit runtime)
Python platform: Linux-5.15.0-92-generic-x86_64-with-glibc2.35
Is CUDA available: N/A
CUDA runtime version: 12.3.107
CUDA_MODULE_LOADING set to: N/A
GPU models and configuration: GPU 0: NVIDIA RTX A4000
Nvidia driver version: 535.154.05
cuDNN version: Could not collect
HIP runtime version: N/A
MIOpen runtime version: N/A
Is XNNPACK available: N/A

CPU:
Architecture: x86_64
CPU op-mode(s): 32-bit, 64-bit
Address sizes: 46 bits physical, 48 bits virtual
Byte Order: Little Endian
CPU(s): 96
On-line CPU(s) list: 0-95
Vendor ID: GenuineIntel
Model name: Intel(R) Xeon(R) Gold 6252 CPU @ 2.10GHz
CPU family: 6
Model: 85
Thread(s) per core: 2
Core(s) per socket: 24
Socket(s): 2
Stepping: 7
CPU max MHz: 3700.0000
CPU min MHz: 1000.0000
BogoMIPS: 4200.00
Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc art arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid dca sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch cpuid_fault epb cat_l3 cdp_l3 invpcid_single intel_ppin ssbd mba ibrs ibpb stibp ibrs_enhanced tpr_shadow vnmi flexpriority ept vpid ept_ad fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid cqm mpx rdt_a avx512f avx512dq rdseed adx smap clflushopt clwb intel_pt avx512cd avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local dtherm ida arat pln pts pku ospke avx512_vnni md_clear flush_l1d arch_capabilities
Virtualization: VT-x
L1d cache: 1.5 MiB (48 instances)
L1i cache: 1.5 MiB (48 instances)
L2 cache: 48 MiB (48 instances)
L3 cache: 71.5 MiB (2 instances)
NUMA node(s): 2
NUMA node0 CPU(s): 0,2,4,6,8,10,12,14,16,18,20,22,24,26,28,30,32,34,36,38,40,42,44,46,48,50,52,54,56,58,60,62,64,66,68,70,72,74,76,78,80,82,84,86,88,90,92,94
NUMA node1 CPU(s): 1,3,5,7,9,11,13,15,17,19,21,23,25,27,29,31,33,35,37,39,41,43,45,47,49,51,53,55,57,59,61,63,65,67,69,71,73,75,77,79,81,83,85,87,89,91,93,95
Vulnerability Gather data sampling: Mitigation; Microcode
Vulnerability Itlb multihit: KVM: Mitigation: VMX disabled
Vulnerability L1tf: Not affected
Vulnerability Mds: Not affected
Vulnerability Meltdown: Not affected
Vulnerability Mmio stale data: Mitigation; Clear CPU buffers; SMT vulnerable
Vulnerability Retbleed: Mitigation; Enhanced IBRS
Vulnerability Spec rstack overflow: Not affected
Vulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl and seccomp
Vulnerability Spectre v1: Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2: Mitigation; Enhanced IBRS, IBPB conditional, RSB filling, PBRSB-eIBRS SW sequence
Vulnerability Srbds: Not affected
Vulnerability Tsx async abort: Mitigation; TSX disabled

Versions of relevant libraries:
[pip3] numpy==1.26.4
[pip3] pytorch-triton==2.1.0+e6216047b8
[pip3] torch==2.2.0
[pip3] torchaudio==2.2.0
[pip3] torchvision==0.17.0
[conda] mkl 2023.1.0 h213fc3f_46344
[conda] mkl-include 2023.2.0 intel_49495 intel
[conda] mkl-static 2023.2.0 intel_49495 intel
[conda] numpy 1.26.4 pypi_0 pypi
[conda] pytorch-triton 2.1.0+e6216047b8 pypi_0 pypi
[conda] torch 2.1.0a0+gita8e7c98 dev_0
[conda] torchaudio 2.2.0 pypi_0 pypi
[conda] torchvision 0.17.0 pypi_0 pypi

cc @malfet @seemethere
@malfet malfet added module: build Build system issues triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module enhancement Not as big of a feature, but technically not a bug. Should be easy to fix labels Feb 16, 2024
@malfet malfet self-assigned this Feb 16, 2024
@malfet
Copy link
Contributor
malfet commented Feb 16, 2024

Hmm, I wonder why it's search for it in plain nightly folder. But as Triton is CUDA independent we probably should just default it to cu121

@malfet malfet removed their assignment Feb 16, 2024
@malfet malfet added needs reproduction Someone else needs to try reproducing the issue given the instructions. No action needed from user and removed enhancement Not as big of a feature, but technically not a bug. Should be easy to fix labels Feb 16, 2024
@malfet
Copy link
Contributor
malfet commented Feb 16, 2024 •
edited
Loading

Actually, it's there in the index, but it should be 2.1.0 rather than 2.2.0: pytorch_triton-2.1.0+bcad9dabe1-cp310-cp310-linux_x86_64.whl

@woongjin-c do you have any local changes to your repo? If not, what commit hash are you on?

@woongjin-c
Copy link
Author

I didn't make any local changes, this is a clean checkout of the v2.2.0 tag with commit hash: 8ac9b20.

I also see in this 2.2.0 release tracker - https://github.com/pytorch/pytorch/issues/115300 others have had the same issue and that PyTorch 2.2.0 should use Triton release 2.2.x, a comment by @huydhn.

@woongjin-c woongjin-c 8000 closed this as completed Feb 16, 2024
@woongjin-c woongjin-c reopened this Feb 16, 2024
@Borda
Copy link
Contributor
Borda commented Feb 22, 2024

having the same when I checkout the latest 2.2.1 and run make triton

pytorch/scripts/install_triton_wheel.sh

Line 3 in 6c8c5ad

pip install --index-url https://download.pytorch.org/whl/nightly/ "pytorch-triton==$(cat .ci/docker/triton_version.txt)+$(head -c 10 .ci/docker/ci_commit_pins/triton.txt)"

the problem is that the referred version is

pytorch/.ci/docker/triton_version.txt

Line 1 in 6c8c5ad

2.2.0

but the append commit SHA is associated with pytorch_triton-2.1.0+bcad9dabe1

@malfet
Copy link
Contributor
malfet commented Feb 23, 2024

I'm technically AFK this week, will try to get a repro tomorrow morning.

@atalman
Copy link
Contributor
atalman commented Feb 23, 2024 •
edited
Loading

Nightly is using pytorch-triton=3.0.0:

pytorch-triton==3.0.0+901819d2b6 (from torch)

Release 2.2.0 is using triton=2.2.0:

triton-2.2.0-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (167.9 MB)

For release 2.2.x triton was built from this branch rather then pinned commit:
https://github.com/openai/triton/tree/release/2.2.x

@woongjin-c and @Borda
If you are building from source for pytorch release 2.2 this is matching version of triton: https://github.com/openai/triton/tree/release/2.2.x

You can refer to this workflow:
https://github.com/pytorch/pytorch/actions/runs/7921244390/job/21626107170#step:6:631

Will run our build from source script to confirm is changes are needed in release branch to make sure it works with release/2.2

@Borda
Copy link
Contributor
Borda commented Feb 23, 2024 •
edited
Loading

Release 2.2.0 is using triton=2.2.0

That would be my expectation, but it is not what make triton does if you checkout v2.2.0

Will run our build from source script to confirm is changes are needed in release branch to make sure it works with release/2.2

How about if I need to build particular release/tag v2.2.1?

@woongjin-c
Copy link
Author

Release v2.2.0 is using Triton 2.2.0, but the version doesn't exist in the distribution.

For tag v2.2.0, it looks for "pytorch-triton=2.2.0+bcad9dabe1", but only "pytorch-triton=2.2.0+e28a256d71" exists in the distribution as @Borda have mentioned as well.

@atalman atalman self-assigned this Feb 23, 2024
@atalman
Copy link
Contributor
atalman commented Mar 4, 2024

@woongjin-c and @Borda to fix this you need to change install_triton_wheel.sh this way in the release:

pip install --index-url https://download.pytorch.org/whl/test/ "pytorch-triton==$(cat .ci/docker/triton_version.txt)"

Will post a Release only PR to release 2.2.0 branch to resolve this

@atalman atalman mentioned this issue Mar 4, 2024
Closed
atalman added a commit to atalman/pytorch that referenced this issue Mar 5, 2024
@atalman
Fix make triton command on release branch (pytorch#121169)
7851745 
Fixes pytorch#120044

Should fix build from source instructions on release branch here: https://github.com/pytorch/pytorch#from-source

Please note we are using /test/ channel for release here to make sure it works, before actual release is completed.

Test main:
```
make triton
pip3 uninstall -y triton
WARNING: Skipping triton as it is not installed.
Looking in indexes: https://download.pytorch.org/whl/nightly/
Collecting pytorch-triton==3.0.0+a9bc1a3647
  Downloading https://download.pytorch.org/whl/nightly/pytorch_triton-3.0.0%2Ba9bc1a3647-cp310-cp310-linux_x86_64.whl (239.0 MB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 239.0/239.0 MB 8.7 MB/s eta 0:00:00
Requirement already satisfied: filelock in /home/atalman/miniconda3/envs/py310/lib/python3.10/site-packages (from pytorch-triton==3.0.0+a9bc1a3647) (3.13.1)
Installing collected packages: pytorch-triton
  Attempting uninstall: pytorch-triton
    Found existing installation: pytorch-triton 2.2.0
    Uninstalling pytorch-triton-2.2.0:
      Successfully uninstalled pytorch-triton-2.2.0
Successfully installed pytorch-triton-3.0.0+a9bc1a3647
```

Test release/2.2:
```
make triton
pip3 uninstall -y triton
WARNING: Skipping triton as it is not installed.
Looking in indexes: https://download.pytorch.org/whl/test/
Collecting pytorch-triton==2.2.0
  Using cached https://download.pytorch.org/whl/test/pytorch_triton-2.2.0-cp310-cp310-linux_x86_64.whl (183.1 MB)
Requirement already satisfied: filelock in /home/atalman/miniconda3/envs/py310/lib/python3.10/site-packages (from pytorch-triton==2.2.0) (3.13.1)
Installing collected packages: pytorch-triton
  Attempting uninstall: pytorch-triton
    Found existing installation: pytorch-triton 3.0.0+a9bc1a3647
    Uninstalling pytorch-triton-3.0.0+a9bc1a3647:
      Successfully uninstalled pytorch-triton-3.0.0+a9bc1a3647
Successfully installed pytorch-triton-2.2.0
```
Pull Request resolved: pytorch#121169
Approved by: https://github.com/seemethere
@atalman atalman mentioned this issue Mar 5, 2024
Merged
atalman added a commit that referenced this issue Mar 7, 2024
@atalman
Fix make triton command on release branch (#121169) (#121229)
56a2068 
Fixes #120044

Should fix build from source instructions on release branch here: https://github.com/pytorch/pytorch#from-source

Please note we are using /test/ channel for release here to make sure it works, before actual release is completed.

Test main:
```
make triton
pip3 uninstall -y triton
WARNING: Skipping triton as it is not installed.
Looking in indexes: https://download.pytorch.org/whl/nightly/
Collecting pytorch-triton==3.0.0+a9bc1a3647
  Downloading https://download.pytorch.org/whl/nightly/pytorch_triton-3.0.0%2Ba9bc1a3647-cp310-cp310-linux_x86_64.whl (239.0 MB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 239.0/239.0 MB 8.7 MB/s eta 0:00:00
Requirement already satisfied: filelock in /home/atalman/miniconda3/envs/py310/lib/python3.10/site-packages (from pytorch-triton==3.0.0+a9bc1a3647) (3.13.1)
Installing collected packages: pytorch-triton
  Attempting uninstall: pytorch-triton
    Found existing installation: pytorch-triton 2.2.0
    Uninstalling pytorch-triton-2.2.0:
      Successfully uninstalled pytorch-triton-2.2.0
Successfully installed pytorch-triton-3.0.0+a9bc1a3647
```
6517


Test release/2.2:
```
make triton
pip3 uninstall -y triton
WARNING: Skipping triton as it is not installed.
Looking in indexes: https://download.pytorch.org/whl/test/
Collecting pytorch-triton==2.2.0
  Using cached https://download.pytorch.org/whl/test/pytorch_triton-2.2.0-cp310-cp310-linux_x86_64.whl (183.1 MB)
Requirement already satisfied: filelock in /home/atalman/miniconda3/envs/py310/lib/python3.10/site-packages (from pytorch-triton==2.2.0) (3.13.1)
Installing collected packages: pytorch-triton
  Attempting uninstall: pytorch-triton
    Found existing installation: pytorch-triton 3.0.0+a9bc1a3647
    Uninstalling pytorch-triton-3.0.0+a9bc1a3647:
      Successfully uninstalled pytorch-triton-3.0.0+a9bc1a3647
Successfully installed pytorch-triton-2.2.0
```
Pull Request resolved: #121169
Approved by: https://github.com/seemethere
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@atalman atalman

Labels
module: build Build system issues needs reproduction Someone else needs to try reproducing the issue given the instructions. No action needed from user triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module
Projects
None yet
Milestone
No milestone
Development

Successfully merging a pull request may close this issue.

Fix make triton command on release branch atalman/pytorch
4 participants
@malfet @Borda @atalman @woongjin-c
0