hvaara
/ nn.linear.memory.debug.2025-05-27.ipynb
Created
May 27, 2025 17:01
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
hvaara
/ gist:340bc4bf740d97c15351db7b6759643d
Created
November 28, 2024 17:36
Memory format debug for https://github.com/pytorch/pytorch/issues/141471 via LTX-Video
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| 2024-11-28 05:45:11.621 python[37469:34111015] mps_3d_convolution:1:1:1:1:1:1:0:1:1:1:Contiguous:bf16[1,128,4,15,22]:bf16[512,128,3,3,3]:1:512 | |
| 2024-11-28 05:45:11.647 python[37469:34111015] mps_3d_convolution:1:1:1:1:1:1:0:1:1:1:Contiguous:bf16[1,512,4,15,22]:bf16[512,512,3,3,3]:1:512 | |
| 2024-11-28 05:45:11.649 python[37469:34111015] mps_3d_convolution:1:1:1:1:1:1:0:1:1:1:Contiguous:bf16[1,512,4,15,22]:bf16[512,512,3,3,3]:1:512 | |
| 2024-11-28 05:45:11.652 python[37469:34111015] mps_3d_convolution:1:1:1:1:1:1:0:1:1:1:Contiguous:bf16[1,512,4,15,22]:bf16[512,512,3,3,3]:1:512 | |
| 2024-11-28 05:45:11.652 python[37469:34111015] mps_3d_convolution:1:1:1:1:1:1:0:1:1:1:Contiguous:bf16[1,512,4,15,22]:bf16[512,512,3,3,3]:1:512 | |
| 2024-11-28 05:45:11.652 python[37469:34111015] mps_3d_convolution:1:1:1:1:1:1:0:1:1:1:Contiguous:bf16[1,512,4,15,22]:bf16[512,512,3,3,3]:1:512 | |
| 2024-11-28 05:45:11.653 python[37469:34111015] mps_3d_convolution:1:1:1:1:1:1:0:1:1:1:Contiguous:bf16[1,512,4,15,22]:bf16[512,512,3,3,3]:1:512 | |
| 2024-11-28 05:45:11.653 |
hvaara
/ gist:e3002c3258c689f9b6a2dd56fb15317c
Created
September 16, 2024 01:04
Tinygrad fused kernel example
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| METAL_XCODE=1 DISABLE_COMPILER_CACHE=1 DEBUG=4 python3 -c "from tinygrad import Tensor; | |
| N = 1024; a, b = Tensor.rand(N, N), Tensor.rand(N, N); | |
| c = (a.reshape(N, 1, N) * b.T.reshape(1, N, N)).sum(axis=2); | |
| print((c.numpy() - (a.numpy() @ b.numpy())).mean())" | |
| opened device METAL from pid:15248 | |
| opened device NPY from pid:15248 | |
| *** CUSTOM 1 custom_random mem 0.00 GB | |
| *** CUSTOM 2 custom_random mem 0.01 GB | |
| TENSOR CORES [(1, 1024, 1)] [(0, 1024, 1024)] WMMA_8_8_8_float_float | |
| r_32_8_2_4_2_2_4_128_8_2_4_4 |
hvaara
/ gist:9db8389c15cf9a665a090ba85926a18d
Created
September 12, 2024 23:47
Example collect_env.py output
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| ~/dev/pytorch/pytorch main *12 !1 git branch | grep -e '^*' ✔ 8s pytorchdev 01:45:33 | |
| * main | |
| ~/dev/pytorch/pytorch main *12 !1 python torch/utils/collect_env.py INT ✘ pytorchdev 01:45:22 | |
| Collecting environment information... | |
| PyTorch version: 2.5.0a0+gitdd47f6f | |
| Is debug build: False | |
| CUDA used to build PyTorch: None | |
| ROCM used to build PyTorch: N/A | |
| OS: macOS 14.6.1 (arm64) |
hvaara
/ repro-278bc98-134580
Created
August 27, 2024 16:50
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| python test/test_modules.py -v -k test_memory_format_nn_BatchNorm2d_ | |
| test_memory_format_nn_BatchNorm2d_eval_mode_cpu_float32 (__main__.TestModuleCPU.test_memory_format_nn_BatchNorm2d_eval_mode_cpu_float32) ... ok | |
| test_memory_format_nn_BatchNorm2d_eval_mode_cpu_float64 (__main__.TestModuleCPU.test_memory_format_nn_BatchNorm2d_eval_mode_cpu_float64) ... ok | |
| test_memory_format_nn_BatchNorm2d_train_mode_cpu_float32 (__main__.TestModuleCPU.test_memory_format_nn_BatchNorm2d_train_mode_cpu_float32) ... ok | |
| test_memory_format_nn_BatchNorm2d_train_mode_cpu_float64 (__main__.TestModuleCPU.test_memory_format_nn_BatchNorm2d_train_mode_cpu_float64) ... ok | |
| test_memory_format_nn_BatchNorm2d_eval_mode_mps_float16 (__main__.TestModuleMPS.test_memory_format_nn_BatchNorm2d_eval_mode_mps_float16) ... ok | |
| test_memory_format_nn_BatchNorm2d_eval_mode_mps_float32 (__main__.TestModuleMPS.test_memory_format_nn_BatchNorm2d_eval_mode_mps_float32) ... ok | |
| test_memory_format_nn_BatchNorm2d_train_mode_mps_float16 (__main__.TestModuleMPS.test_m |
hvaara
/ repro-43f78bf-134580
Created
August 27, 2024 16:46
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| python test/test_modules.py -v -k test_memory_format_nn_BatchNorm2d_ | |
| test_memory_format_nn_BatchNorm2d_eval_mode_cpu_float32 (__main__.TestModuleCPU.test_memory_format_nn_BatchNorm2d_eval_mode_cpu_float32) ... ok | |
| test_memory_format_nn_BatchNorm2d_eval_mode_cpu_float64 (__main__.TestModuleCPU.test_memory_format_nn_BatchNorm2d_eval_mode_cpu_float64) ... ok | |
| test_memory_format_nn_BatchNorm2d_train_mode_cpu_float32 (__main__.TestModuleCPU.test_memory_format_nn_BatchNorm2d_train_mode_cpu_float32) ... ok | |
| test_memory_format_nn_BatchNorm2d_train_mode_cpu_float64 (__main__.TestModuleCPU.test_memory_format_nn_BatchNorm2d_train_mode_cpu_float64) ... ok | |
| test_memory_format_nn_BatchNorm2d_eval_mode_mps_float16 (__main__.TestModuleMPS.test_memory_format_nn_BatchNorm2d_eval_mode_mps_float16) ... FAIL | |
| test_memory_format_nn_BatchNorm2d_eval_mode_mps_float32 (__main__.TestModuleMPS.test_memory_format_nn_BatchNorm2d_eval_mode_mps_float32) ... FAIL | |
| test_memory_format_nn_BatchNorm2d_train_mode_mps_float16 (__main__.TestModuleMPS.te |
hvaara
/ repro-278bc98-134580
Created
August 27, 2024 16:42
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| test_memory_format_nn_BatchNorm2d_eval_mode_cpu_float32 (__main__.TestModuleCPU.test_memory_format_nn_BatchNorm2d_eval_mode_cpu_float32) ... ok | |
| test_memory_format_nn_BatchNorm2d_eval_mode_cpu_float64 (__main__.TestModuleCPU.test_memory_format_nn_BatchNorm2d_eval_mode_cpu_float64) ... ok | |
| test_memory_format_nn_BatchNorm2d_train_mode_cpu_float32 (__main__.TestModuleCPU.test_memory_format_nn_BatchNorm2d_train_mode_cpu_float32) ... ok | |
| test_memory_format_nn_BatchNorm2d_train_mode_cpu_float64 (__main__.TestModuleCPU.test_memory_format_nn_BatchNorm2d_train_mode_cpu_float64) ... ok | |
| test_memory_format_nn_BatchNorm2d_eval_mode_mps_float16 (__main__.TestModuleMPS.test_memory_format_nn_BatchNorm2d_eval_mode_mps_float16) ... FAIL | |
| test_memory_format_nn_BatchNorm2d_eval_mode_mps_float32 (__main__.TestModuleMPS.test_memory_format_nn_BatchNorm2d_eval_mode_mps_float32) ... FAIL | |
| test_memory_format_nn_BatchNorm2d_train_mode_mps_float16 (__main__.TestModuleMPS.test_memory_format_nn_BatchNorm2d_train_mode_mps_float16) ... FAIL | |
| test |
hvaara
/ gist:26a4f5d4dc82a0dd550b0408d645e8fe
Created
August 27, 2024 16:39
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| python test/test_modules.py -v -k test_memory_format_nn_BatchNorm2d_ | |
| test_memory_format_nn_BatchNorm2d_eval_mode_cpu_float32 (__main__.TestModuleCPU.test_memory_format_nn_BatchNorm2d_eval_mode_cpu_float32) ... ok | |
| test_memory_format_nn_BatchNorm2d_eval_mode_cpu_float64 (__main__.TestModuleCPU.test_memory_format_nn_BatchNorm2d_eval_mode_cpu_float64) ... ok | |
| test_memory_format_nn_BatchNorm2d_train_mode_cpu_float32 (__main__.TestModuleCPU.test_memory_format_nn_BatchNorm2d_train_mode_cpu_float32) ... ok | |
| test_memory_format_nn_BatchNorm2d_train_mode_cpu_float64 (__main__.TestModuleCPU.test_memory_format_nn_BatchNorm2d_train_mode_cpu_float64) ... ok | |
| test_memory_format_nn_BatchNorm2d_eval_mode_mps_float16 (__main__.TestModuleMPS.test_memory_format_nn_BatchNorm2d_eval_mode_mps_float16) ... FAIL | |
| test_memory_format_nn_BatchNorm2d_eval_mode_mps_float32 (__main__.TestModuleMPS.test_memory_format_nn_BatchNorm2d_eval_mode_mps_float32) ... FAIL | |
| test_memory_format_nn_BatchNorm2d_train_mode_mps_float16 (__main__.TestModuleMPS.te |
hvaara
/ gist:140a09f0fd5b4c36932bebdfc9275c3c
Last active
August 27, 2024 15:53
Debug for https://github.com/pytorch/pytorch/issues/134580
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| git log v2.4.0..HEAD --oneline | |
| d74039f7010 (HEAD -> repro-24-134580, origin/repro-24-134580) Repro case for #134580 | |
| ccdbe084a9e Skip memory_format tests | |
| 49f0d3f1111 Update common_modules.py | |
| 630cc4ea8b2 Update test_nn.py | |
| b702a483965 Use newer `toAccumulateType` signature in Normalization.cpp | |
| $ git diff v2.4.0 | |
| diff --git a/aten/src/ATen/native/Normalization.cpp b/aten/src/ATen/native/Normalization.cpp | |
| index e9e7c001837..16ada4cead5 100644 |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| from collections.abc import Callable, Sequence | |
| import enum | |
| import types | |
| from typing import Annotated, overload | |
| import numpy | |
| from numpy.typing import ArrayLike | |
| from . import ( | |
| distributed as distributed, |
NewerOlder