pytorch/torch/csrc/jit/jit_opt_limit.h
Peter Bell b2e79ed5ec Remove WindowsTorchApiMacro.h in favor of Export.h (#69585)
Summary:
Follow up to https://github.com/pytorch/pytorch/issues/68095

This also changes the files from the ATen folder to include c10's `Export.h` instead since they can't ever be exporting `TORCH_PYTHON_API`.

cc pietern mrshenli pritamdamania87 zhaojuanmao satgera rohan-varma gqchen aazzolini osalpekar jiayisuse SciPioneer H-Huang

Pull Request resolved: https://github.com/pytorch/pytorch/pull/69585

Reviewed By: mrshenli

Differential Revision: D32958594

Pulled By: albanD

fbshipit-source-id: 1ec7ef63764573fa2b486928955e3a1172150061
2021-12-09 17:30:09 -08:00

40 lines
1.4 KiB
C++

#pragma once
#include <torch/csrc/Export.h>
#include <string>
#include <unordered_map>
// `TorchScript` offers a simple optimization limit checker
// that can be configured through environment variable `PYTORCH_JIT_OPT_LIMIT`.
// The purpose is to limit how many optimization you can make per pass.
// This is useful for debugging any passes.
// Opt limit checker is enabled on a per file basis (hence per pass). For
// example, in `constant_propagation.cpp`, `PYTORCH_JIT_OPT_LIMIT` should be set
// to `constant_propagation=<opt_limt>` or, simply, to
// `constant_propagation=<opt_limit>` where <opt_limit> is the number of
// optimizations you want to make for the pass. (i.e.
// `PYTORCH_JIT_OPT_LIMIT="constant_propagation=<opt_limit>"`).
// Multiple files can be configured by separating each file name with a colon
// `:` as in the following example,
// `PYTORCH_JIT_OPT_LIMIT="constant_propagation=<opt_limit>:dead_code_elimination=<opt_limit>"`
// You can call opt limiter by calling JIT_OPT_ALLOWED. It will return true if
// we haven't reached the optimization limit yet. Otherwise, it will return
// false. Typical usage:
// if (!JIT_OPT_ALLOWED) {
// GRAPH_DUMP(...); //supplied from jit_log
// return;
// }
namespace torch {
namespace jit {
TORCH_API bool opt_limit(const char* pass_name);
#define JIT_OPT_ALLOWED opt_limit(__FILE__)
} // namespace jit
} // namespace torch