The AMP inserts `_autocast_to_reduced_precision` and `_autocast_to_full_precision` automatically. The aten implementation provides a fast path to bypass the conversion if the tensor data type has been the reduced/full precision. But NNC always does the conversion which could bring >5% E2E performance regression.
This PR is to address the performance issue like aten. We will not pull `_autocast_to_reduced_precision` and `_autocast_to_full_precision` into NNC fusion group and fallback to aten to trigger its fast path if the tensor data type has been the reduced/full precision.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/85140
Approved by: https://github.com/frank-wei
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/72390
This class didn't add much value and only caused more boilerplate code.
This change removes the class and updates all the use cases with
uses of `ExprHandle`.
A side effect of this change is different names in loop variables, which
caused massive mechanical changes in our tests.
Test Plan: Imported from OSS
Reviewed By: navahgar
Differential Revision: D34030296
Pulled By: ZolotukhinM
fbshipit-source-id: 2ba4e313506a43ab129a10d99e72b638b7d40108
(cherry picked from commit c2ec46a058)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/71186
So far we've only supported scalar inputs, but couldn't handle scalar outputs
or intermediates. This PR adds it.
Scalar outputs are returned as 0-dim tensors. If the kernel is invoked on a
stack of IValues, we correctly convert the results to scalar IValues when
needed. If the kernel is invoked with a vector of void* pointers, everything
works out of the box without any conversions.
Lowerings for scalar operators are a bit tricky. Usual lowerings return a pair
<Buf, Stmt> (aka Tensor), but for scalar operators we also want to have the
corresponding Var that the lowering function supposedly creates (in theory we
could just use Loads and Stores, but I'm worried it can affect performance as
there is no guarantee this will be optimized by LLVM). So, what we do here to
work around this is we return a fake buf + stmt that sets the corresponding
var. Then outside of the lowering we create a real buffer and generate a Store
to it with the value from the variable we passed as the base handle of the fake
buf. This real buffer is then treated as usual by the rest of the system and we
can use it if we need to return this scalar value as a kernel output. If we do
not need to return it, then the Store will be deleted by the DCE pass.
Differential Revision:
D33539324
D33539324
Test Plan: Imported from OSS
Reviewed By: navahgar
Pulled By: ZolotukhinM
fbshipit-source-id: ab4524b9820ce204f106effcf6232ed33d4ee223
(cherry picked from commit 7faa0939f0)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/70871
We had previously handled reusing memory in the optimized kernel execution path, but not yet handled it if we hit the unoptimized fallback.
Test Plan: Imported from OSS
Reviewed By: ngimel
Differential Revision: D33458652
Pulled By: eellison
fbshipit-source-id: 4eb62181ed02c95813a99638f5e2d0f9347b5c08
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/70535
This also fixes handling of inputs that happen to be outputs (they
require copy).
Test Plan: Imported from OSS
Reviewed By: pbelevich
Differential Revision: D33399116
Pulled By: ZolotukhinM
fbshipit-source-id: 9845838eb653b82ae47b527631b51893990d5319
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/69477
This diff adds a new run method to `TensorExprKernel` which takes in
output tensors as inputs and stores the output in those given tensors.
ghstack-source-id: 146107009
Test Plan: buck test mode/dev-nosan //caffe2/test/cpp/tensorexpr:tensorexpr -- --exact 'caffe2/test/cpp/tensorexpr:tensorexpr - Kernel.RunWithAllocatedOutputs'
Reviewed By: ZolotukhinM
Differential Revision: D32823890
fbshipit-source-id: edc1f4839785124048b034060feb71cb8c1be34f
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/68858
when executing with ir_eval, check for index out of bounds.
Test Plan: Imported from OSS
Reviewed By: ZolotukhinM
Differential Revision: D32657881
Pulled By: davidberard98
fbshipit-source-id: 62dd0f85bb182b34e9c9f795ff761081290f6922
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/67861
Previously submitted as https://github.com/pytorch/pytorch/pull/67197.
This got reverted because its failures were hidden by the failures of
another PR.
Test Plan: Imported from OSS
Reviewed By: ZolotukhinM
Differential Revision: D32178196
Pulled By: navahgar
fbshipit-source-id: cc8a5c68aed360d06289e69645461cfa773e1300
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/67734
The implementation of `aten::cat` op in NNC has to ignore tensors that have 0-size in any dimension.
Test Plan: `buck test mode/dev-nosan //caffe2/test/cpp/tensorexpr:tensorexpr -- --exact 'caffe2/test/cpp/tensorexpr:tensorexpr - Kernel.CatWithEmptyInputs'`
Reviewed By: ZolotukhinM
Differential Revision: D32122171
fbshipit-source-id: 90c697813bc504664673cdc262df6e7ce419c655
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/66744
Modified loops in files under fbsource/fbcode/caffe2/ from the format
`for(TYPE var=x0;var<x_max;x++)`
to the format
`for(const auto var: irange(xmax))`
This was achieved by running r-barnes's loop upgrader script (D28874212) with some modification to exclude all files under /torch/jit and a number of reversions or unused variable suppression warnings added by hand.
Test Plan: Sandcastle
Reviewed By: ngimel
Differential Revision: D31705358
fbshipit-source-id: d6ea350cbaa8f452fc78f238160e5374be637a48
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/66234
Modified loops in files under fbsource/fbcode/caffe2/ from the format
`for(TYPE var=x0;var<x_max;x++)`
to the format
`for(const auto var: irange(xmax))`
This was achieved by running r-barnes's loop upgrader script (D28874212) with some modification to exclude all files under /torch/jit and a number of reversions or unused variable suppression warnings added by hand.
bypass_size_limit
allow-large-files
Test Plan: Sandcastle
Reviewed By: ngimel
Differential Revision: D30652629
fbshipit-source-id: 0ae6c4bbbb554bad42e372792a6430e1acf15e3e
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/64717
This also exposed several bugs, which are fixed in this PR.
Differential Revision:
D30826408
D30826408
Test Plan: Imported from OSS
Reviewed By: navahgar
Pulled By: ZolotukhinM
fbshipit-source-id: a67ec5739aceed9ffdf0d24f77eb3787cefe4560
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/64627
This fixes the root cause of S242719
Test Plan: Imported from OSS
Reviewed By: ZolotukhinM
Differential Revision: D30801686
Pulled By: navahgar
fbshipit-source-id: b6d3ebdc7eb57116eaced53c2f35c7798bb17e80
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/64609
We've been using exceptions to indicate whether vectorization succeeded
or not, but that posed some problems with (e.g. we spent too much time
symbolicazing these exceptions). This change converts this mechanism to
a standard error return code.
Test Plan: Imported from OSS
Reviewed By: bertmaher
Differential Revision: D30795342
Pulled By: ZolotukhinM
fbshipit-source-id: 16e38b37bcdd78ceb438ac814cc377f35b058e17
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/64077
We were assuming kernel dimensions fit in 32 bits (the old fuser made
this assumption too), but we should be able to support 64.
ghstack-source-id: 136933272
Test Plan: unit tests; new IR level test with huge sizes
Reviewed By: ZolotukhinM
Differential Revision: D30596689
fbshipit-source-id: 23b7e393a2ebaecb0c391a6b1f0c4b05a98bcc94
Summary:
Fixes https://github.com/pytorch/pytorch/issues/63923
The input graph can contain constants whose names contain special characters. So, all names of constants in the input graph need to be sanitized.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/63990
Reviewed By: ZolotukhinM
Differential Revision: D30558432
Pulled By: navahgar
fbshipit-source-id: de5b0c23d50ee8997f40f2c0fc605dda3719186f
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/63776
I reverted this out of an abundance of caution because some test
failures occurred, but they were all due to precision issues fixed lower in
this stack. Let's try again.
I've rolled the elimination of the allow-parallelism-in-fusions toggle into
this diff since they're pretty tightly coupled.
ghstack-source-id: 136529847
Test Plan: CI
Reviewed By: huiguoo
Differential Revision: D30484555
fbshipit-source-id: 38fd33520f710585d1130c365a8c60c9ce794a59
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/63587
Now that there is no classes using KernelArena for memory management we
can remove it.
Differential Revision:
D30429115
D30429115
Test Plan: Imported from OSS
Reviewed By: navahgar
Pulled By: ZolotukhinM
fbshipit-source-id: 375f6f9294d27790645eeb7cb5a8e87047a57544
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/63586
This is another commit in transition from KernelArena memory management.
Tensor is essentially just a pair of <BufPtr, StmtPtr> and we don't need
to dynamically allocate it at all - it's cheap to pass it by value, and
that's what we're switching to in this commit.
After this change nothing uses KernelScope/KernelArena and they can be
safely removed.
Differential Revision:
D30429114
D30429114
Test Plan: Imported from OSS
Reviewed By: navahgar
Pulled By: ZolotukhinM
fbshipit-source-id: f90b859cfe863692b7beffbe9bd0e4143df1e819
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/63195
This helps us to later switch from using KernelArena with raw pointers
to shared pointers without having to change all our source files at
once.
The changes are mechanical and should not affect any functionality.
With this PR, we're changing the following:
* `Add*` --> `AddPtr`
* `new Add(...)` --> `alloc<Add>(...)`
* `dynamic_cast<Add*>` --> `to<Add>`
* `static_cast<Add*>` --> `static_to<Add>`
Due to some complications with args forwarding, some places became more
verbose, e.g.:
* `new Block({})` --> `new Block(std::vector<ExprPtr>())`
Test Plan: Imported from OSS
Reviewed By: navahgar
Differential Revision: D30292779
Pulled By: ZolotukhinM
fbshipit-source-id: 150301c7d2df56b608b035827b6a9a87f5e2d9e9
Summary:
As GoogleTest `TEST` macro is non-compliant with it as well as `DEFINE_DISPATCH`
All changes but the ones to `.clang-tidy` are generated using following script:
```
for i in `find . -type f -iname "*.c*" -or -iname "*.h"|xargs grep cppcoreguidelines-avoid-non-const-global-variables|cut -f1 -d:|sort|uniq`; do sed -i "/\/\/ NOLINTNEXTLINE(cppcoreguidelines-avoid-non-const-global-variables)/d" $i; done
```
Pull Request resolved: https://github.com/pytorch/pytorch/pull/62008
Reviewed By: driazati, r-barnes
Differential Revision: D29838584
Pulled By: malfet
fbshipit-source-id: 1b2f8602c945bd4ce50a9bfdd204755556e31d13
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/60804
The lowerings are stored as a map c10::Symbol -> std::function and the
signature of thoese functions match the signature of
`computeOperandValue`. Custom lowerings have higher priority over the
standard ones, i.e. we can redefine how a given op is lowered.
In general this feature is aimed at unblocking users whose models
contain ops that are not yet supported by NNC - it allows to quickly add
a custom lowering for a given op.
Test Plan: Imported from OSS
Reviewed By: bertmaher
Differential Revision: D29409580
Pulled By: ZolotukhinM
fbshipit-source-id: e8e8dc9d3cb9155cfbf5c08a4216ba1b5b791a60
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/59508
An assert that was triggering in a previous version is now relaxed to
take 0-dim tensors into account.
Test Plan: Imported from OSS
Reviewed By: bertmaher
Differential Revision: D28918342
Pulled By: ZolotukhinM
fbshipit-source-id: c09b62c9725d1603b0ec11fcc051e7c932af06ae
