[JIT][fuser] Emit math functions specific to output type #18815
Closed
Conversation
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
facebook-github-bot
added
the
oncall: jit
jamesr66a
force-pushed
the
double_fuser
branch
7 times, most recently
from
49d2db4 to
5a12bd3
Compare
zdevito
reviewed
Apr 4, 2019
torch/csrc/jit/fuser/codegen.cpp
Outdated
There was a problem hiding this comment.
Refactor this into a function please.
torch/csrc/jit/fuser/codegen.cpp
Outdated
There was a problem hiding this comment.
struct Template {
Template(const char* flor_float, const char* for_double);
const char * for_float;
const char * for_double;
};
Then the lines can be much shorter:
{aten::sigmoid, {"1.f / (1.f + expf(-${0}))", "1. / (1. + exp(-${0}))"}},
5be3a9a to
35a54fa
Compare
There was a problem hiding this comment.
@jamesr66a has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.
suo
approved these changes
Apr 5, 2019
torch/csrc/jit/fuser/codegen.cpp
Outdated
torch/csrc/jit/fuser/codegen.cpp
Outdated
There was a problem hiding this comment.
Can you add a comment explaining what this struct is for
test/test_jit.py
Outdated
There was a problem hiding this comment.
Can you add a brief comment explaining what tests are supposed to check
jamesr66a
force-pushed
the
double_fuser
branch
2 times, most recently
from
43fbbac to
696a4a0
Compare
There was a problem hiding this comment.
@jamesr66a is landing this pull request. If you are a Facebook employee, you can view this diff on Phabricator.
There was a problem hiding this comment.
@jamesr66a is landing this pull request. If you are a Facebook employee, you can view this diff on Phabricator.
|
@jamesr66a merged this pull request in 34382e4. |
facebook-github-bot
pushed a commit
that referenced
this pull request
Apr 7, 2019
Summary: Stacked on #18815 and #18811. This makes it so that we emit a higher-precision literal for float values in the fusion kernel, as well as assign that to a `double` variable. This prevents us from losing precision for values such as `pi`, but with the previous fixes this will also get downcasted to `float` if downstream operations require it. Therefore, we should not lose performance because of implicit promotions Pull Request resolved: #18817 Differential Revision: D14820842 Pulled By: jamesr66a fbshipit-source-id: 519671c6ca5e7adac746a4c4c72760a6d91e332f
zhangguanheng66
pushed a commit
to zhangguanheng66/pytorch
that referenced
this pull request
May 6, 2019
Summary: Stacked on pytorch#18811 This makes it so that we only emit the *f variants of math functions if the output value's type is FloatTensor, otherwise we call the double variants to prevent loss of precision. This fixes more numerical issues Pull Request resolved: pytorch#18815 Differential Revision: D14816965 Pulled By: jamesr66a fbshipit-source-id: 464be644168875ede987142281fb2168f4041e81
zhangguanheng66
pushed a commit
to zhangguanheng66/pytorch
that referenced
this pull request
May 6, 2019
Summary: Stacked on pytorch#18815 and pytorch#18811. This makes it so that we emit a higher-precision literal for float values in the fusion kernel, as well as assign that to a `double` variable. This prevents us from losing precision for values such as `pi`, but with the previous fixes this will also get downcasted to `float` if downstream operations require it. Therefore, we should not lose performance because of implicit promotions Pull Request resolved: pytorch#18817 Differential Revision: D14820842 Pulled By: jamesr66a fbshipit-source-id: 519671c6ca5e7adac746a4c4c72760a6d91e332f
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Stacked on #18811
This makes it so that we only emit the *f variants of math functions if the output value's type is FloatTensor, otherwise we call the double variants to prevent loss of precision. This fixes more numerical issues