When generating the triton template for convolution, we check `V.graph.sizevars.statically_known_equals(in_chan * groups, x.get_size()[1]) `. Note that in this check, we should consider the groups.
This check verifies, at compile time, that the total number of input channels expected by the convolution weights (in_chan * groups) exactly matches the number of channels in the input tensor (x.get_size()[1]).
This fix is good in general as it allows for conv triton template to be generated when `groups> 1`. It's also required for unified runtime to use AOTI as a backend delegate, because unified runtime is libtorch-free, so we cannot use the ATEN fallback of conv2d.
```
python test/inductor/test_select_algorithm.py -k test_convolution2_group
```
Pull Request resolved: https://github.com/pytorch/pytorch/pull/163094
Approved by: https://github.com/SherlockNoMad
Summary: use `self.inductor_meta_common()` to call the static method, since the custom subclasses may overwrite the method to be an instance method
Test Plan:
```
caffe2/test/inductor:select_algorithm -- test_finalized_subclass_hooks
```
Rollback Plan:
Differential Revision: D80375351
Pull Request resolved: https://github.com/pytorch/pytorch/pull/160895
Approved by: https://github.com/eellison, https://github.com/blaine-rister
# Moativation
This PR is used to enable _int_mm on Intel GPU. And _int_mm is used by int8 quantization on torchao.
# Model Test Result:
We run meta-llama/Llama-3.1-8B-Instruct on Intel GPU and A100 using torchao int8-dynamic-quantization. The model configs as below:
Precision : torch.bfloat16
quantization configuration : Int8DynamicActivationInt8WeightConfig
dataset : wikitext
Result:
The perplexity values for Intel GPU and A100 are 9.582953453063965 and 9.57755184173584, respectively.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/157769
Approved by: https://github.com/EikanWang, https://github.com/desertfire
This refactor ensures all registered template hooks have been finalised before accessing the code object of the template. In `simd.SimdScheduling.codegen_template` the template hooks are finalised manually with `template.finalize_hook(hook_name)` calls, so it is the responsibility of the caller to finalise all the template hooks. This PR adds:
- `RenderPartial.finalize_remaining` a function that can be called at the end to finalise the remaining active hooks after a selection of hooks have been finalised manually.
- A test with a custom template implementation that registers custom hooks that the scheduler needs to finalise. This test should fail if the scheduler does not finalise the registered custom hook.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/157270
Approved by: https://github.com/eellison
Summary:
# Why
- keep code cleaner
- modular way to hook up preprocessing steps
- expand testability of flows that change which choices are provided e.g. to test performance models and lookup tables by running torch.compile
# What
- similar to feedback_saver_fns, now there are preprocessing_fns
- the existing regex logic is exported into those as a proof of concept
Test Plan:
```
buck2 run mode/opt scripts/coconutruben/torchmm:experiment 2>&1 | tee /tmp/epx038
```
This does not exercise the logic, it just shows that it's safe right now
Rollback Plan:
Differential Revision: D76946993
Pull Request resolved: https://github.com/pytorch/pytorch/pull/156464
Approved by: https://github.com/masnesral
Summary:
Currently only `num_warps` and `num_stages` are supported as one of the kernel options for inductor auto-tuning using `TritonTemplate`.
In order to allow warp-specialization kernel options should allow specifying `num_consumer_groups` and `num_buffers_warp_spec` as well.
NOTE: Currently gating changes to FBCODE using HAS_WARP_SPEC which is only available on triton/release-3.3.x
Test Plan:
## Unit test
Added tests for `test_triton_template_warp_specialization` to verify generated kenrnel contains configs for `num_consumer_groups` and `num_buffers_warp_spec`.
## Functional Testing
Specific to flexattention.
```
import torch
from torch.nn.attention.flex_attention import flex_attention
from triton.testing import do_bench
make_tensor = lambda: torch.rand(8, 16, 8192, 128, device="cuda", dtype=torch.bfloat16)
q, k, v = make_tensor(), make_tensor(), make_tensor()
flex_compiled = torch.compile(flex_attention, fullgraph=True)
print(do_bench(lambda: flex_compiled(q, k, v, kernel_options={"num_warps": 4})))
```
triton do_bench results:
- default compile: 15.176783561706543
- with warp-spec: 9.452800750732422
## Extra notes
- generated triton kernel using `TORCH_LOGS=output_code`: P1740612877
- TTGIR for fused kernel: P1740614685
Differential Revision: D71982587
Pull Request resolved: https://github.com/pytorch/pytorch/pull/150122
Approved by: https://github.com/eellison, https://github.com/zou3519, https://github.com/jansel
Summary:
Currently only `num_warps` and `num_stages` are supported as one of the kernel options for inductor auto-tuning using `TritonTemplate`. In order to allow warp-specialization kernel options should allow specifying `num_consumer_groups` and `num_buffers_warp_spec` as well.
Test Plan:
## Unit test
Added tests for `test_triton_template_warp_specialization` to verify generated kenrnel contains configs for `num_consumer_groups` and `num_buffers_warp_spec`.
## Functional Testing
Specific to flexattention.
```
import torch
from torch.nn.attention.flex_attention import flex_attention
from triton.testing import do_bench
make_tensor = lambda: torch.rand(8, 16, 8192, 128, device="cuda", dtype=torch.bfloat16)
q, k, v = make_tensor(), make_tensor(), make_tensor()
flex_compiled = torch.compile(flex_attention, fullgraph=True)
print(do_bench(lambda: flex_compiled(q, k, v, kernel_options={"num_warps": 4})))
```
triton do_bench results:
- default compile: 15.176783561706543
- with warp-spec: 9.452800750732422
## Extra notes
- generated triton kernel using `TORCH_LOGS=output_code`: P1740612877
- TTGIR for fused kernel: P1740614685
Differential Revision: D70212243
Pull Request resolved: https://github.com/pytorch/pytorch/pull/148503
Approved by: https://github.com/eellison
Summary:
Relands D69965761 / https://github.com/pytorch/pytorch/pull/147583
Before this PR, calling a triton kernel would look like:
```py
kernel.run(a, b, xnumel, grid=grid(xnumel), stream=stream0)
```
where the `grid=` was passed as a callable (function closure) arg. This PR removes the grid arg:
```py
kernel.run(a, b, xnumel, stream=stream0)
```
instead now the grid computation is included in the kernel launcher, with something like:
```py
def launcher(in_ptr0, out_ptr0, xnumel, stream):
grid_0 = ((xnumel + 1023) >> 10)
grid_1 = 1
grid_2 = 1
runner(grid_0, grid_1, grid_2, stream, function, metadata, None, launch_enter_hook, launch_exit_hook, in_ptr0, out_ptr0, xnumel)
```
This should be faster, since we remove multiple function/dict calls and are able to specialize the grid computation for each `triton.Config`.
It also allows us to unify the handling of grids between the Python and C++ wrapper code. Before this, C++ wrapper code didn't actually support dynamic grid sizes and instead burned in a static grid.
This unification allows this PR to be a net deletion of code.
Differential [disconnected] Revision: D70471332
Pull Request resolved: https://github.com/pytorch/pytorch/pull/148305
Approved by: https://github.com/shunting314, https://github.com/eellison
Summary:
Relands D69965761 / https://github.com/pytorch/pytorch/pull/147583
Before this PR, calling a triton kernel would look like:
```py
kernel.run(a, b, xnumel, grid=grid(xnumel), stream=stream0)
```
where the `grid=` was passed as a callable (function closure) arg. This PR removes the grid arg:
```py
kernel.run(a, b, xnumel, stream=stream0)
```
instead now the grid computation is included in the kernel launcher, with something like:
```py
def launcher(in_ptr0, out_ptr0, xnumel, stream):
grid_0 = ((xnumel + 1023) >> 10)
grid_1 = 1
grid_2 = 1
runner(grid_0, grid_1, grid_2, stream, function, metadata, None, launch_enter_hook, launch_exit_hook, in_ptr0, out_ptr0, xnumel)
```
This should be faster, since we remove multiple function/dict calls and are able to specialize the grid computation for each `triton.Config`.
It also allows us to unify the handling of grids between the Python and C++ wrapper code. Before this, C++ wrapper code didn't actually support dynamic grid sizes and instead burned in a static grid.
This unification allows this PR to be a net deletion of code.
Differential Revision: D70471332
Pull Request resolved: https://github.com/pytorch/pytorch/pull/148305
Approved by: https://github.com/shunting314, https://github.com/eellison
#147620 enabled `force_shape_pad` for triton kernel benchmark. Intel GPU supports this scenario. Hence, we need to enable the case in this PR. Otherwise, there would be a test case regression for Intel GPU as #147620 has been landed.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/148237
Approved by: https://github.com/jansel
Before this PR, calling a triton kernel would look like:
```py
kernel.run(a, b, xnumel, grid=grid(xnumel), stream=stream0)
```
where the `grid=` was passed as a callable (function closure) arg. This PR removes the grid arg:
```py
kernel.run(a, b, xnumel, stream=stream0)
```
instead now the grid computation is included in the kernel launcher, with something like:
```py
def launcher(in_ptr0, out_ptr0, xnumel, stream):
grid_0 = ((xnumel + 1023) >> 10)
grid_1 = 1
grid_2 = 1
runner(grid_0, grid_1, grid_2, stream, function, metadata, None, launch_enter_hook, launch_exit_hook, in_ptr0, out_ptr0, xnumel)
```
This should be faster, since we remove multiple function/dict calls and are able to specialize the grid computation for each `triton.Config`.
It also allows us to unify the handling of grids between the Python and C++ wrapper code. Before this, C++ wrapper code didn't actually support dynamic grid sizes and instead burned in a static grid.
This unification allows this PR to be a net deletion of code.
Note the attached diff contains some minor fbcode-only changes.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/147583
Approved by: https://github.com/eellison, https://github.com/shunting314
This PR aims to add the functionality support of max-autotune for XPU. The current triton templates and configurations are not well optimized for XPU, so the performance is not ready yet. Also the `mm_plus_mm` template have accuracy issues in some cases. We will address these issues in the next PRs.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/143266
Approved by: https://github.com/EikanWang, https://github.com/jansel
This PR aims to add the functionality support of max-autotune for XPU. The current triton templates and configurations are not well optimized for XPU, so the performance is not ready yet. Also the `mm_plus_mm` template have accuracy issues in some cases. We will address these issues in the next PRs.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/143266
Approved by: https://github.com/EikanWang, https://github.com/jansel
Summary: For the second-pass, we don't have to rerun the whole inductor flow again. This PR moves that second-pass to the codegen time. This change not only speeds up the compilation, but also removes kernel scheduling inconsistency between the two passes. Another future improvement is to make the second-pass reuse the scheduler and do the wrapper codegen only.
This is a copy of https://github.com/pytorch/pytorch/pull/113762 to land in github first.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/114067
Approved by: https://github.com/chenyang78
