Skip to content

GH-90997: Improve inline cache performance for MSVC #96781

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
brandtbucher merged 3 commits into from
Sep 15, 2022
Merged

GH-90997: Improve inline cache performance for MSVC #96781

Changes from 1 commit
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
eaa91ed
Use memcpy for cache reads/writes
brandtbucher Sep 12, 2022
ae8dfea
blurb add
brandtbucher Sep 12, 2022
5985a4a
Update comment
brandtbucher Sep 14, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Prev Previous commit
Update comment
  • Loading branch information
@brandtbucher
brandtbucher committed Sep 14, 2022
commit 5985a4abaf326193d270570e74b72f3a321d806d
17 changes: 8 additions & 9 deletions Include/internal/pycore_code.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -285,15 +285,14 @@ PyAPI_FUNC(PyObject*) _Py_GetSpecializationStats(void);
#define EVAL_CALL_STAT_INC_IF_FUNCTION(name, callable) ((void)0)
#endif // !Py_STATS

// NOTE: These cache reading/writing utilities use memcpy to avoid voilating C's
// strict aliasing rules, while also avoiding the need to maintain big-endian
// versions of the same code. Compilers are smart enough to understand what
// we're really trying to do here (see https://blog.regehr.org/archives/959).

// When modifying these, great care must be taken to ensure that we don't break
// or slow down our inline caching! All of these functions should compile to
// simple "move" instructions on all supported compilers and platforms. You can
// use the Compiler Explorer at https://godbolt.org to help verify this.
// Utility functions for reading/writing 32/64-bit values in the inline caches.
// Great care should be taken to ensure that these functions remain correct and
// performant! They should compile to just "move" instructions on all supported
// compilers and platforms.

// We use memcpy to let the C compiler handle unaligned accesses and endianness
// issues for us. It also seems to produce better code than manual copying for
// most compilers (see https://blog.regehr.org/archives/959 for more info).

static inline void
write_u32(uint16_t *p, uint32_t val)
Expand Down