Tree: 8c8cbc43a7
2 Commits (8c8cbc43a76fb6632e9fbaf56461bae50ab7223e)
| Author | SHA1 | Message | Date |
|---|---|---|---|
Protobuf Team Bot
|
6b00befa40 |
Internal change
PiperOrigin-RevId: 571462941 |
1 year ago |
Adam Cozzette
|
501ececd39 |
Reorganize upb file structure
This change moves almost everything in the `upb/` directory up one level, so that for example `upb/upb/generated_code_support.h` becomes just `upb/generated_code_support.h`. The only exceptions I made to this were that I left `upb/cmake` and `upb/BUILD` where they are, mostly because that avoids conflict with other files and the current locations seem reasonable for now. The `python/` directory is a little bit of a challenge because we had to merge the existing directory there with `upb/python/`. I made `upb/python/BUILD` into the BUILD file for the merged directory, and it effectively loads the contents of the other BUILD file via `python/build_targets.bzl`, but I plan to clean this up soon. PiperOrigin-RevId: 568651768 |
1 year ago |
|
Adam Cozzette
|
0d1ba13e83 |
Move everything into new upb/ directory
|
2 years ago |
Eric Salo
|
eb9e33ec42 |
upb: remove <stddef.h> and <stdlib.h> from port/def.inc
PiperOrigin-RevId: 559234283 |
2 years ago |
|
Eric Salo
|
d2f10df2dc |
upb: split apart upb/upb.hpp, delete :upb
PiperOrigin-RevId: 558945210 |
2 years ago |
|
Eric Salo
|
cc41a8bed3 |
upb: dedup the libraries
Remove array.h and map.h as hdrs from :collections_internal Remove alloc.h and arena.h as hdrs from :mem_internal (and add them to :mem) Remove common.h and decode.h and encode.h as hdrs from :wire_internal Lock down the visibility of :wire_internal to upb-only Merge :mini_descriptor_encode_internal into :mini_descriptor_internal PiperOrigin-RevId: 558235138 |
2 years ago |
|
Protobuf Team Bot
|
c552102d66 |
No public description
PiperOrigin-RevId: 556849883 |
2 years ago |
|
Adam Cozzette
|
5aca728f72 |
Reformat copyright headers
PiperOrigin-RevId: 554509301 |
2 years ago |
Jason Lunn
|
1b0011dfef |
Support ASAN detection on clang (#1424 An error occurred Clang and GCC differ on how they detect Address Sanitizer. Support both. Closes #1424 |
2 years ago |
|
Eric Salo
|
3817bbb572 |
upb: create upb/mem/BUILD
The next in a series of CLs to split upb/BUILD into subdirs. Create mem/internal/ Delete the deprecated upb/arena.h and upb/alloc.h stub headers PiperOrigin-RevId: 552864952 |
2 years ago |
Joshua Haberman
|
a6852ec919 |
Moved TypeScript direct bit read/writes to //third_party/upb/bits
This will move us towards keeping all these encapsulation breaks in a common place. This also removes the IFTTT blocks. Keeping IFTTT blocks for each language would become unmanageable, but going forward we will know to look in `//third_party/upb/bits` whenever the internal data structures in upb change. PiperOrigin-RevId: 547494495 |
2 years ago |
Matt Kulukundis
|
fa1c2a6d29 |
Switch upb_Arena_Fuse from a CAS based list insertion to an exchange based one
Second try with improved testing (Generated by http://go/benchy. Settings: --runs 20 --reference "srcfs" --perflab) ``` name old cpu/op new cpu/op delta BM_ArenaOneAlloc 18.2ns ± 2% 18.1ns ± 1% -0.72% (p=0.002 n=18+17) BM_ArenaInitialBlockOneAlloc 5.31ns ± 0% 5.30ns ± 1% ~ (p=0.345 n=16+19) BM_ArenaFuseUnbalanced/2 67.8ns ± 1% 68.0ns ± 0% +0.35% (p=0.011 n=16+17) BM_ArenaFuseUnbalanced/8 526ns ± 2% 524ns ± 1% ~ (p=0.708 n=18+17) BM_ArenaFuseUnbalanced/64 4.82µs ± 1% 4.84µs ± 1% +0.31% (p=0.049 n=16+17) BM_ArenaFuseUnbalanced/128 9.78µs ± 1% 9.82µs ± 1% +0.46% (p=0.001 n=17+17) BM_ArenaFuseBalanced/2 66.9ns ± 1% 67.2ns ± 1% +0.36% (p=0.025 n=17+16) BM_ArenaFuseBalanced/8 527ns ± 2% 529ns ± 1% ~ (p=0.081 n=17+19) BM_ArenaFuseBalanced/64 4.92µs ± 4% 4.88µs ± 2% ~ (p=0.184 n=18+17) BM_ArenaFuseBalanced/128 9.92µs ± 1% 9.91µs ± 1% ~ (p=0.883 n=16+19) BM_LoadAdsDescriptor_Upb<NoLayout> 5.89ms ± 2% 5.94ms ± 1% +0.88% (p=0.005 n=18+17) BM_LoadAdsDescriptor_Upb<WithLayout> 6.55ms ± 2% 6.55ms ± 1% ~ (p=0.961 n=17+18) BM_LoadAdsDescriptor_Proto2<NoLayout> 12.3ms ± 2% 12.4ms ± 1% ~ (p=0.226 n=18+18) BM_LoadAdsDescriptor_Proto2<WithLayout> 12.5ms ± 1% 12.6ms ± 1% +0.61% (p=0.005 n=17+19) BM_Parse_Upb_FileDesc<UseArena, Copy> 12.6µs ± 1% 12.7µs ± 2% ~ (p=0.219 n=17+18) BM_Parse_Upb_FileDesc<UseArena, Alias> 11.6µs ± 2% 11.6µs ± 3% ~ (p=0.721 n=16+18) BM_Parse_Upb_FileDesc<InitBlock, Copy> 12.4µs ± 1% 12.5µs ± 1% ~ (p=0.118 n=16+17) BM_Parse_Upb_FileDesc<InitBlock, Alias> 11.3µs ± 2% 11.4µs ± 1% ~ (p=0.327 n=18+19) BM_Parse_Proto2<FileDesc, NoArena, Copy> 25.2µs ± 2% 25.3µs ± 1% ~ (p=0.301 n=16+19) BM_Parse_Proto2<FileDesc, UseArena, Copy> 12.1µs ± 3% 12.1µs ± 2% ~ (p=0.869 n=18+19) BM_Parse_Proto2<FileDesc, InitBlock, Copy> 11.8µs ± 3% 11.8µs ± 3% ~ (p=0.462 n=18+18) BM_Parse_Proto2<FileDescSV, InitBlock, Alias> 13.2µs ± 1% 13.2µs ± 1% ~ (p=0.333 n=16+19) BM_SerializeDescriptor_Proto2 5.83µs ± 3% 5.86µs ± 4% ~ (p=0.496 n=18+20) BM_SerializeDescriptor_Upb 10.5µs ± 2% 10.4µs ± 1% -1.20% (p=0.000 n=18+16) name old time/op new time/op delta BM_ArenaOneAlloc 18.2ns ± 2% 18.1ns ± 0% -0.73% (p=0.010 n=18+17) BM_ArenaInitialBlockOneAlloc 5.32ns ± 0% 5.31ns ± 1% ~ (p=0.106 n=15+18) BM_ArenaFuseUnbalanced/2 67.9ns ± 1% 68.1ns ± 0% +0.31% (p=0.044 n=16+16) BM_ArenaFuseUnbalanced/8 527ns ± 2% 526ns ± 1% ~ (p=0.772 n=18+16) BM_ArenaFuseUnbalanced/64 4.83µs ± 1% 4.84µs ± 2% ~ (p=0.144 n=16+18) BM_ArenaFuseUnbalanced/128 9.79µs ± 1% 9.84µs ± 1% +0.52% (p=0.001 n=17+18) BM_ArenaFuseBalanced/2 67.0ns ± 1% 67.3ns ± 3% +0.41% (p=0.019 n=15+16) BM_ArenaFuseBalanced/8 528ns ± 2% 530ns ± 1% ~ (p=0.121 n=17+19) BM_ArenaFuseBalanced/64 4.93µs ± 4% 4.89µs ± 2% ~ (p=0.103 n=18+17) BM_ArenaFuseBalanced/128 9.93µs ± 1% 9.93µs ± 1% ~ (p=0.806 n=16+19) BM_LoadAdsDescriptor_Upb<NoLayout> 5.91ms ± 2% 5.96ms ± 1% +0.93% (p=0.002 n=18+16) BM_LoadAdsDescriptor_Upb<WithLayout> 6.57ms ± 2% 6.57ms ± 1% ~ (p=0.935 n=17+18) BM_LoadAdsDescriptor_Proto2<NoLayout> 12.4ms ± 2% 12.4ms ± 1% ~ (p=0.239 n=18+18) BM_LoadAdsDescriptor_Proto2<WithLayout> 12.5ms ± 2% 12.6ms ± 1% +0.43% (p=0.024 n=18+19) BM_Parse_Upb_FileDesc<UseArena, Copy> 12.7µs ± 2% 12.7µs ± 2% ~ (p=0.245 n=17+18) BM_Parse_Upb_FileDesc<UseArena, Alias> 11.6µs ± 2% 11.6µs ± 2% ~ (p=0.772 n=16+18) BM_Parse_Upb_FileDesc<InitBlock, Copy> 12.5µs ± 1% 12.5µs ± 1% ~ (p=0.136 n=16+17) BM_Parse_Upb_FileDesc<InitBlock, Alias> 11.4µs ± 2% 11.4µs ± 1% ~ (p=0.391 n=18+19) BM_Parse_Proto2<FileDesc, NoArena, Copy> 25.3µs ± 2% 25.4µs ± 1% ~ (p=0.403 n=16+19) BM_Parse_Proto2<FileDesc, UseArena, Copy> 12.1µs ± 2% 12.1µs ± 2% ~ (p=0.731 n=17+19) BM_Parse_Proto2<FileDesc, InitBlock, Copy> 11.9µs ± 3% 11.8µs ± 3% ~ (p=0.424 n=18+18) BM_Parse_Proto2<FileDescSV, InitBlock, Alias> 13.2µs ± 2% 13.3µs ± 1% ~ (p=0.683 n=16+19) BM_SerializeDescriptor_Proto2 5.84µs ± 3% 5.86µs ± 4% ~ (p=0.496 n=18+20) BM_SerializeDescriptor_Upb 10.5µs ± 2% 10.4µs ± 1% -1.27% (p=0.000 n=18+16) name old speed new speed delta BM_LoadAdsDescriptor_Upb<NoLayout> 133MB/s ± 2% 132MB/s ± 1% -0.97% (p=0.002 n=18+16) BM_LoadAdsDescriptor_Upb<WithLayout> 120MB/s ± 2% 120MB/s ± 1% ~ (p=0.961 n=17+18) BM_LoadAdsDescriptor_Proto2<NoLayout> 63.5MB/s ± 2% 63.3MB/s ± 1% ~ (p=0.226 n=18+18) BM_LoadAdsDescriptor_Proto2<WithLayout> 62.7MB/s ± 1% 62.4MB/s ± 1% -0.60% (p=0.005 n=17+19) BM_Parse_Upb_FileDesc<UseArena, Copy> 596MB/s ± 1% 594MB/s ± 2% ~ (p=0.219 n=17+18) BM_Parse_Upb_FileDesc<UseArena, Alias> 650MB/s ± 2% 649MB/s ± 3% ~ (p=0.721 n=16+18) BM_Parse_Upb_FileDesc<InitBlock, Copy> 605MB/s ± 1% 603MB/s ± 1% ~ (p=0.118 n=16+17) BM_Parse_Upb_FileDesc<InitBlock, Alias> 663MB/s ± 2% 661MB/s ± 1% ~ (p=0.327 n=18+19) BM_Parse_Proto2<FileDesc, NoArena, Copy> 298MB/s ± 2% 297MB/s ± 1% ~ (p=0.490 n=17+19) BM_Parse_Proto2<FileDesc, UseArena, Copy> 623MB/s ± 3% 624MB/s ± 2% ~ (p=0.869 n=18+19) BM_Parse_Proto2<FileDesc, InitBlock, Copy> 636MB/s ± 3% 637MB/s ± 3% ~ (p=0.462 n=18+18) BM_Parse_Proto2<FileDescSV, InitBlock, Alias> 570MB/s ± 1% 568MB/s ± 1% ~ (p=0.333 n=16+19) BM_SerializeDescriptor_Proto2 1.29GB/s ± 3% 1.29GB/s ± 4% ~ (p=0.496 n=18+20) BM_SerializeDescriptor_Upb 716MB/s ± 2% 725MB/s ± 1% +1.20% (p=0.000 n=18+16) ``` PiperOrigin-RevId: 525132431 |
2 years ago |
|
Protobuf Team Bot
|
750358b863 |
Switch upb_Arena_Fuse from a CAS based list insertion to an exchange based one
``` (Generated by http://go/benchy. Settings: --runs 20 --reference "srcfs" --perflab) name old cpu/op new cpu/op delta BM_ArenaOneAlloc 18.2ns ± 2% 18.1ns ± 1% -0.72% (p=0.002 n=18+17) BM_ArenaInitialBlockOneAlloc 5.31ns ± 0% 5.30ns ± 1% ~ (p=0.345 n=16+19) BM_ArenaFuseUnbalanced/2... *** PiperOrigin-RevId: 524406384 |
2 years ago |
|
Matt Kulukundis
|
52a8d214ca |
Switch upb_Arena_Fuse from a CAS based list insertion to an exchange based one
``` (Generated by http://go/benchy. Settings: --runs 20 --reference "srcfs" --perflab) name old cpu/op new cpu/op delta BM_ArenaOneAlloc 18.2ns ± 2% 18.1ns ± 1% -0.72% (p=0.002 n=18+17) BM_ArenaInitialBlockOneAlloc 5.31ns ± 0% 5.30ns ± 1% ~ (p=0.345 n=16+19) BM_ArenaFuseUnbalanced/2 67.8ns ± 1% 68.0ns ± 0% +0.35% (p=0.011 n=16+17) BM_ArenaFuseUnbalanced/8 526ns ± 2% 524ns ± 1% ~ (p=0.708 n=18+17) BM_ArenaFuseUnbalanced/64 4.82µs ± 1% 4.84µs ± 1% +0.31% (p=0.049 n=16+17) BM_ArenaFuseUnbalanced/128 9.78µs ± 1% 9.82µs ± 1% +0.46% (p=0.001 n=17+17) BM_ArenaFuseBalanced/2 66.9ns ± 1% 67.2ns ± 1% +0.36% (p=0.025 n=17+16) BM_ArenaFuseBalanced/8 527ns ± 2% 529ns ± 1% ~ (p=0.081 n=17+19) BM_ArenaFuseBalanced/64 4.92µs ± 4% 4.88µs ± 2% ~ (p=0.184 n=18+17) BM_ArenaFuseBalanced/128 9.92µs ± 1% 9.91µs ± 1% ~ (p=0.883 n=16+19) BM_LoadAdsDescriptor_Upb<NoLayout> 5.89ms ± 2% 5.94ms ± 1% +0.88% (p=0.005 n=18+17) BM_LoadAdsDescriptor_Upb<WithLayout> 6.55ms ± 2% 6.55ms ± 1% ~ (p=0.961 n=17+18) BM_LoadAdsDescriptor_Proto2<NoLayout> 12.3ms ± 2% 12.4ms ± 1% ~ (p=0.226 n=18+18) BM_LoadAdsDescriptor_Proto2<WithLayout> 12.5ms ± 1% 12.6ms ± 1% +0.61% (p=0.005 n=17+19) BM_Parse_Upb_FileDesc<UseArena, Copy> 12.6µs ± 1% 12.7µs ± 2% ~ (p=0.219 n=17+18) BM_Parse_Upb_FileDesc<UseArena, Alias> 11.6µs ± 2% 11.6µs ± 3% ~ (p=0.721 n=16+18) BM_Parse_Upb_FileDesc<InitBlock, Copy> 12.4µs ± 1% 12.5µs ± 1% ~ (p=0.118 n=16+17) BM_Parse_Upb_FileDesc<InitBlock, Alias> 11.3µs ± 2% 11.4µs ± 1% ~ (p=0.327 n=18+19) BM_Parse_Proto2<FileDesc, NoArena, Copy> 25.2µs ± 2% 25.3µs ± 1% ~ (p=0.301 n=16+19) BM_Parse_Proto2<FileDesc, UseArena, Copy> 12.1µs ± 3% 12.1µs ± 2% ~ (p=0.869 n=18+19) BM_Parse_Proto2<FileDesc, InitBlock, Copy> 11.8µs ± 3% 11.8µs ± 3% ~ (p=0.462 n=18+18) BM_Parse_Proto2<FileDescSV, InitBlock, Alias> 13.2µs ± 1% 13.2µs ± 1% ~ (p=0.333 n=16+19) BM_SerializeDescriptor_Proto2 5.83µs ± 3% 5.86µs ± 4% ~ (p=0.496 n=18+20) BM_SerializeDescriptor_Upb 10.5µs ± 2% 10.4µs ± 1% -1.20% (p=0.000 n=18+16) name old time/op new time/op delta BM_ArenaOneAlloc 18.2ns ± 2% 18.1ns ± 0% -0.73% (p=0.010 n=18+17) BM_ArenaInitialBlockOneAlloc 5.32ns ± 0% 5.31ns ± 1% ~ (p=0.106 n=15+18) BM_ArenaFuseUnbalanced/2 67.9ns ± 1% 68.1ns ± 0% +0.31% (p=0.044 n=16+16) BM_ArenaFuseUnbalanced/8 527ns ± 2% 526ns ± 1% ~ (p=0.772 n=18+16) BM_ArenaFuseUnbalanced/64 4.83µs ± 1% 4.84µs ± 2% ~ (p=0.144 n=16+18) BM_ArenaFuseUnbalanced/128 9.79µs ± 1% 9.84µs ± 1% +0.52% (p=0.001 n=17+18) BM_ArenaFuseBalanced/2 67.0ns ± 1% 67.3ns ± 3% +0.41% (p=0.019 n=15+16) BM_ArenaFuseBalanced/8 528ns ± 2% 530ns ± 1% ~ (p=0.121 n=17+19) BM_ArenaFuseBalanced/64 4.93µs ± 4% 4.89µs ± 2% ~ (p=0.103 n=18+17) BM_ArenaFuseBalanced/128 9.93µs ± 1% 9.93µs ± 1% ~ (p=0.806 n=16+19) BM_LoadAdsDescriptor_Upb<NoLayout> 5.91ms ± 2% 5.96ms ± 1% +0.93% (p=0.002 n=18+16) BM_LoadAdsDescriptor_Upb<WithLayout> 6.57ms ± 2% 6.57ms ± 1% ~ (p=0.935 n=17+18) BM_LoadAdsDescriptor_Proto2<NoLayout> 12.4ms ± 2% 12.4ms ± 1% ~ (p=0.239 n=18+18) BM_LoadAdsDescriptor_Proto2<WithLayout> 12.5ms ± 2% 12.6ms ± 1% +0.43% (p=0.024 n=18+19) BM_Parse_Upb_FileDesc<UseArena, Copy> 12.7µs ± 2% 12.7µs ± 2% ~ (p=0.245 n=17+18) BM_Parse_Upb_FileDesc<UseArena, Alias> 11.6µs ± 2% 11.6µs ± 2% ~ (p=0.772 n=16+18) BM_Parse_Upb_FileDesc<InitBlock, Copy> 12.5µs ± 1% 12.5µs ± 1% ~ (p=0.136 n=16+17) BM_Parse_Upb_FileDesc<InitBlock, Alias> 11.4µs ± 2% 11.4µs ± 1% ~ (p=0.391 n=18+19) BM_Parse_Proto2<FileDesc, NoArena, Copy> 25.3µs ± 2% 25.4µs ± 1% ~ (p=0.403 n=16+19) BM_Parse_Proto2<FileDesc, UseArena, Copy> 12.1µs ± 2% 12.1µs ± 2% ~ (p=0.731 n=17+19) BM_Parse_Proto2<FileDesc, InitBlock, Copy> 11.9µs ± 3% 11.8µs ± 3% ~ (p=0.424 n=18+18) BM_Parse_Proto2<FileDescSV, InitBlock, Alias> 13.2µs ± 2% 13.3µs ± 1% ~ (p=0.683 n=16+19) BM_SerializeDescriptor_Proto2 5.84µs ± 3% 5.86µs ± 4% ~ (p=0.496 n=18+20) BM_SerializeDescriptor_Upb 10.5µs ± 2% 10.4µs ± 1% -1.27% (p=0.000 n=18+16) name old speed new speed delta BM_LoadAdsDescriptor_Upb<NoLayout> 133MB/s ± 2% 132MB/s ± 1% -0.97% (p=0.002 n=18+16) BM_LoadAdsDescriptor_Upb<WithLayout> 120MB/s ± 2% 120MB/s ± 1% ~ (p=0.961 n=17+18) BM_LoadAdsDescriptor_Proto2<NoLayout> 63.5MB/s ± 2% 63.3MB/s ± 1% ~ (p=0.226 n=18+18) BM_LoadAdsDescriptor_Proto2<WithLayout> 62.7MB/s ± 1% 62.4MB/s ± 1% -0.60% (p=0.005 n=17+19) BM_Parse_Upb_FileDesc<UseArena, Copy> 596MB/s ± 1% 594MB/s ± 2% ~ (p=0.219 n=17+18) BM_Parse_Upb_FileDesc<UseArena, Alias> 650MB/s ± 2% 649MB/s ± 3% ~ (p=0.721 n=16+18) BM_Parse_Upb_FileDesc<InitBlock, Copy> 605MB/s ± 1% 603MB/s ± 1% ~ (p=0.118 n=16+17) BM_Parse_Upb_FileDesc<InitBlock, Alias> 663MB/s ± 2% 661MB/s ± 1% ~ (p=0.327 n=18+19) BM_Parse_Proto2<FileDesc, NoArena, Copy> 298MB/s ± 2% 297MB/s ± 1% ~ (p=0.490 n=17+19) BM_Parse_Proto2<FileDesc, UseArena, Copy> 623MB/s ± 3% 624MB/s ± 2% ~ (p=0.869 n=18+19) BM_Parse_Proto2<FileDesc, InitBlock, Copy> 636MB/s ± 3% 637MB/s ± 3% ~ (p=0.462 n=18+18) BM_Parse_Proto2<FileDescSV, InitBlock, Alias> 570MB/s ± 1% 568MB/s ± 1% ~ (p=0.333 n=16+19) BM_SerializeDescriptor_Proto2 1.29GB/s ± 3% 1.29GB/s ± 4% ~ (p=0.496 n=18+20) BM_SerializeDescriptor_Upb 716MB/s ± 2% 725MB/s ± 1% +1.20% (p=0.000 n=18+16) ``` PiperOrigin-RevId: 524398084 |
2 years ago |
|
Joshua Haberman
|
a918473c65 |
Implemented scalar fields for upb.js and Dart+upb.js.
PiperOrigin-RevId: 522658020 |
2 years ago |
|
Joshua Haberman
|
7ebeb6006b |
Fix the GitHub Actions for upb
An earlier change inadvertently disabled several GitHub Actions due to a syntax error. This broke the CI for upb because several expected tests were never running. I was not able to get the Windows CI test fully working, so I disabled it for now. PiperOrigin-RevId: 522652505 |
2 years ago |
|
Matt Kulukundis
|
71d1e8c85a |
Restructure fuse a tiny bit for aesthetics
* make internal functions start with `_` * make internal functions static * pull invariant tests to the very top name old cpu/op new cpu/op delta BM_ArenaOneAlloc 17.7ns ± 9% 17.9ns ±10% ~ (p=0.790 n=16+17) BM_ArenaInitialBlockOneAlloc 5.22ns ± 3% 5.63ns ±13% ~ (p=0.068 n=16+20) BM_ArenaFuseUnbalanced/2 68.4ns ± 3% 66.4ns ± 3% -3.05% (p=0.000 n=16+16) BM_ArenaFuseUnbalanced/8 515ns ± 3% 539ns ±16% ~ (p=0.496 n=18+20) BM_ArenaFuseUnbalanced/64 4.73µs ± 4% 4.71µs ± 4% ~ (p=0.444 n=16+17) BM_ArenaFuseUnbalanced/128 9.67µs ± 5% 9.59µs ± 6% ~ (p=0.217 n=17+16) BM_ArenaFuseBalanced/2 68.3ns ± 4% 66.9ns ± 5% -2.00% (p=0.020 n=18+18) BM_ArenaFuseBalanced/8 525ns ± 4% 519ns ± 3% ~ (p=0.128 n=16+16) BM_ArenaFuseBalanced/64 5.03µs ±17% 4.77µs ± 6% ~ (p=0.109 n=20+16) BM_ArenaFuseBalanced/128 10.2µs ±17% 9.6µs ± 3% ~ (p=0.058 n=20+16) name old time/op new time/op delta BM_ArenaOneAlloc 17.7ns ± 8% 17.9ns ±10% ~ (p=0.736 n=16+17) BM_ArenaInitialBlockOneAlloc 5.23ns ± 3% 5.66ns ±15% ~ (p=0.077 n=16+20) BM_ArenaFuseUnbalanced/2 68.6ns ± 3% 66.5ns ± 3% -3.07% (p=0.000 n=16+16) BM_ArenaFuseUnbalanced/8 516ns ± 3% 542ns ±18% ~ (p=0.496 n=18+20) BM_ArenaFuseUnbalanced/64 4.74µs ± 4% 4.72µs ± 4% ~ (p=0.423 n=16+17) BM_ArenaFuseUnbalanced/128 9.69µs ± 5% 9.61µs ± 7% ~ (p=0.191 n=17+16) BM_ArenaFuseBalanced/2 68.5ns ± 4% 67.1ns ± 5% -1.99% (p=0.022 n=18+18) BM_ArenaFuseBalanced/8 526ns ± 4% 520ns ± 3% ~ (p=0.157 n=16+16) BM_ArenaFuseBalanced/64 5.04µs ±18% 4.79µs ± 6% ~ (p=0.109 n=20+16) BM_ArenaFuseBalanced/128 10.2µs ±18% 9.7µs ± 3% ~ (p=0.062 n=20+16) PiperOrigin-RevId: 520541220 |
2 years ago |
|
Joshua Haberman
|
c642e43a5a |
Allow fuse/fuse races, so that upb_Arena is fully thread-compatible.
Previously upb_Arena was not thread-compatible when `upb_Arena_Fuse(a, b)` and `upb_Arena_Fuse(c, d)` executed in parallel if `b` and `c` were previously fused. This CL fixed that by allowing `upb_Arena_Fuse()` to run in parallel without limitations. Details on the design of the algorithm are captured in comments. The CL slightly improves the performance of `upb_Arena_Fuse()`. ``` name old cpu/op new cpu/op delta BM_ArenaOneAlloc 20.0ns ±19% 17.5ns ± 4% -12.30% (p=0.000 n=19+17) BM_ArenaInitialBlockOneAlloc 6.65ns ± 4% 5.17ns ± 3% -22.23% (p=0.000 n=18+17) BM_ArenaFuseUnbalanced/2 69.1ns ± 7% 68.5ns ± 4% ~ (p=0.327 n=18+19) BM_ArenaFuseUnbalanced/8 542ns ± 3% 513ns ± 4% -5.25% (p=0.000 n=18+18) BM_ArenaFuseUnbalanced/64 5.04µs ± 8% 4.74µs ± 4% -5.93% (p=0.000 n=17+17) BM_ArenaFuseUnbalanced/128 10.1µs ± 4% 9.6µs ± 4% -4.80% (p=0.000 n=18+17) BM_ArenaFuseBalanced/2 71.8ns ± 7% 68.4ns ± 6% -4.75% (p=0.000 n=17+17) BM_ArenaFuseBalanced/8 541ns ± 3% 519ns ± 3% -4.21% (p=0.000 n=18+17) BM_ArenaFuseBalanced/64 5.00µs ± 7% 4.86µs ± 4% -2.78% (p=0.003 n=17+18) BM_ArenaFuseBalanced/128 10.0µs ± 4% 9.7µs ± 4% -2.68% (p=0.001 n=16+18) BM_LoadAdsDescriptor_Upb<NoLayout> 5.52ms ± 2% 5.54ms ± 4% ~ (p=0.707 n=16+19) BM_LoadAdsDescriptor_Upb<WithLayout> 6.18ms ± 3% 6.15ms ± 3% ~ (p=0.501 n=18+18) BM_LoadAdsDescriptor_Proto2<NoLayout> 11.8ms ± 7% 11.7ms ± 5% ~ (p=0.330 n=16+18) BM_LoadAdsDescriptor_Proto2<WithLayout> 11.9ms ± 3% 11.8ms ± 3% ~ (p=0.303 n=18+17) BM_Parse_Upb_FileDesc<UseArena, Copy> 12.2µs ± 4% 12.3µs ± 4% ~ (p=0.935 n=17+18) BM_Parse_Upb_FileDesc<UseArena, Alias> 11.3µs ± 6% 11.3µs ± 3% ~ (p=0.873 n=16+17) BM_Parse_Upb_FileDesc<InitBlock, Copy> 12.1µs ± 4% 12.1µs ± 3% ~ (p=0.501 n=18+18) BM_Parse_Upb_FileDesc<InitBlock, Alias> 11.1µs ± 4% 11.1µs ± 2% ~ (p=0.297 n=18+16) BM_Parse_Proto2<FileDesc, NoArena, Copy> 24.2µs ± 3% 25.6µs ±16% ~ (p=0.177 n=17+20) BM_Parse_Proto2<FileDesc, UseArena, Copy> 11.6µs ± 3% 11.7µs ± 4% ~ (p=0.232 n=17+18) BM_Parse_Proto2<FileDesc, InitBlock, Copy> 11.5µs ± 7% 11.4µs ± 4% ~ (p=0.707 n=18+19) BM_Parse_Proto2<FileDescSV, InitBlock, Alias> 12.8µs ± 5% 13.0µs ±14% ~ (p=0.782 n=18+17) BM_SerializeDescriptor_Proto2 5.69µs ± 5% 5.76µs ± 6% ~ (p=0.143 n=18+18) BM_SerializeDescriptor_Upb 10.2µs ± 4% 10.2µs ± 3% ~ (p=0.613 n=18+17) name old time/op new time/op delta BM_ArenaOneAlloc 20.0ns ±19% 17.6ns ± 4% -12.37% (p=0.000 n=19+17) BM_ArenaInitialBlockOneAlloc 6.66ns ± 4% 5.18ns ± 3% -22.24% (p=0.000 n=18+17) BM_ArenaFuseUnbalanced/2 69.2ns ± 7% 68.6ns ± 4% ~ (p=0.343 n=18+19) BM_ArenaFuseUnbalanced/8 543ns ± 3% 515ns ± 4% -5.21% (p=0.000 n=18+18) BM_ArenaFuseUnbalanced/64 5.05µs ± 8% 4.75µs ± 4% -5.93% (p=0.000 n=17+17) BM_ArenaFuseUnbalanced/128 10.1µs ± 4% 9.6µs ± 4% -4.78% (p=0.000 n=18+17) BM_ArenaFuseBalanced/2 72.0ns ± 7% 68.6ns ± 6% -4.73% (p=0.000 n=17+17) BM_ArenaFuseBalanced/8 543ns ± 3% 520ns ± 3% -4.20% (p=0.000 n=18+17) BM_ArenaFuseBalanced/64 5.01µs ± 7% 4.87µs ± 4% -2.78% (p=0.004 n=17+18) BM_ArenaFuseBalanced/128 10.0µs ± 3% 9.8µs ± 4% -2.67% (p=0.001 n=16+18) BM_LoadAdsDescriptor_Upb<NoLayout> 5.53ms ± 2% 5.56ms ± 4% ~ (p=0.707 n=16+19) BM_LoadAdsDescriptor_Upb<WithLayout> 6.20ms ± 3% 6.17ms ± 2% ~ (p=0.424 n=18+18) BM_LoadAdsDescriptor_Proto2<NoLayout> 11.8ms ± 7% 11.7ms ± 5% ~ (p=0.297 n=16+18) BM_LoadAdsDescriptor_Proto2<WithLayout> 11.9ms ± 3% 11.9ms ± 3% ~ (p=0.351 n=18+17) BM_Parse_Upb_FileDesc<UseArena, Copy> 12.3µs ± 4% 12.3µs ± 4% ~ (p=1.000 n=17+18) BM_Parse_Upb_FileDesc<UseArena, Alias> 11.3µs ± 6% 11.3µs ± 3% ~ (p=0.845 n=16+17) BM_Parse_Upb_FileDesc<InitBlock, Copy> 12.1µs ± 4% 12.1µs ± 3% ~ (p=0.542 n=18+18) BM_Parse_Upb_FileDesc<InitBlock, Alias> 11.1µs ± 4% 11.2µs ± 2% ~ (p=0.330 n=18+16) BM_Parse_Proto2<FileDesc, NoArena, Copy> 24.2µs ± 3% 25.7µs ±17% ~ (p=0.167 n=17+20) BM_Parse_Proto2<FileDesc, UseArena, Copy> 11.6µs ± 3% 11.7µs ± 3% ~ (p=0.232 n=17+18) BM_Parse_Proto2<FileDesc, InitBlock, Copy> 11.5µs ± 7% 11.4µs ± 4% ~ (p=0.799 n=18+19) BM_Parse_Proto2<FileDescSV, InitBlock, Alias> 12.8µs ± 5% 13.0µs ±14% ~ (p=0.807 n=18+17) BM_SerializeDescriptor_Proto2 5.71µs ± 5% 5.78µs ± 6% ~ (p=0.143 n=18+18) BM_SerializeDescriptor_Upb 10.2µs ± 4% 10.2µs ± 3% ~ (p=0.613 n=18+17) name old allocs/op new allocs/op delta BM_ArenaOneAlloc 1.00 ± 0% 1.00 ± 0% ~ (all samples are equal) BM_ArenaFuseUnbalanced/2 2.00 ± 0% 2.00 ± 0% ~ (all samples are equal) BM_ArenaFuseUnbalanced/8 8.00 ± 0% 8.00 ± 0% ~ (all samples are equal) BM_ArenaFuseUnbalanced/64 64.0 ± 0% 64.0 ± 0% ~ (all samples are equal) BM_ArenaFuseUnbalanced/128 128 ± 0% 128 ± 0% ~ (all samples are equal) BM_ArenaFuseBalanced/2 2.00 ± 0% 2.00 ± 0% ~ (all samples are equal) BM_ArenaFuseBalanced/8 8.00 ± 0% 8.00 ± 0% ~ (all samples are equal) BM_ArenaFuseBalanced/64 64.0 ± 0% 64.0 ± 0% ~ (all samples are equal) BM_ArenaFuseBalanced/128 128 ± 0% 128 ± 0% ~ (all samples are equal) BM_LoadAdsDescriptor_Upb<NoLayout> 6.05k ± 0% 6.05k ± 0% ~ (all samples are equal) BM_LoadAdsDescriptor_Upb<WithLayout> 6.36k ± 0% 6.36k ± 0% ~ (all samples are equal) BM_LoadAdsDescriptor_Proto2<NoLayout> 83.4k ± 0% 83.4k ± 0% ~ (all samples are equal) BM_LoadAdsDescriptor_Proto2<WithLayout> 84.4k ± 0% 84.4k ± 0% -0.00% (p=0.013 n=19+20) BM_Parse_Upb_FileDesc<UseArena, Copy> 7.00 ± 0% 7.00 ± 0% ~ (all samples are equal) BM_Parse_Upb_FileDesc<UseArena, Alias> 7.00 ± 0% 7.00 ± 0% ~ (all samples are equal) BM_Parse_Proto2<FileDesc, NoArena, Copy> 765 ± 0% 765 ± 0% ~ (all samples are equal) BM_Parse_Proto2<FileDesc, UseArena, Copy> 8.00 ± 0% 8.00 ± 0% ~ (all samples are equal) name old peak-mem(Bytes)/op new peak-mem(Bytes)/op delta BM_ArenaOneAlloc 336 ± 0% 328 ± 0% -2.38% (p=0.000 n=20+20) BM_ArenaFuseUnbalanced/2 672 ± 0% 656 ± 0% -2.38% (p=0.000 n=20+20) BM_ArenaFuseUnbalanced/8 2.69k ± 0% 2.62k ± 0% -2.38% (p=0.000 n=20+20) BM_ArenaFuseUnbalanced/64 21.5k ± 0% 21.0k ± 0% -2.38% (p=0.000 n=20+20) BM_ArenaFuseUnbalanced/128 43.0k ± 0% 42.0k ± 0% -2.38% (p=0.000 n=20+20) BM_ArenaFuseBalanced/2 672 ± 0% 656 ± 0% -2.38% (p=0.000 n=20+20) BM_ArenaFuseBalanced/8 2.69k ± 0% 2.62k ± 0% -2.38% (p=0.000 n=20+20) BM_ArenaFuseBalanced/64 21.5k ± 0% 21.0k ± 0% -2.38% (p=0.000 n=20+20) BM_ArenaFuseBalanced/128 43.0k ± 0% 42.0k ± 0% -2.38% (p=0.000 n=20+20) BM_LoadAdsDescriptor_Upb<NoLayout> 10.0M ± 0% 9.9M ± 0% -0.05% (p=0.000 n=20+20) BM_LoadAdsDescriptor_Upb<WithLayout> 10.0M ± 0% 10.0M ± 0% -0.05% (p=0.000 n=20+20) BM_LoadAdsDescriptor_Proto2<NoLayout> 6.62M ± 0% 6.62M ± 0% ~ (all samples are equal) BM_LoadAdsDescriptor_Proto2<WithLayout> 6.66M ± 0% 6.66M ± 0% -0.01% (p=0.013 n=19+20) BM_Parse_Upb_FileDesc<UseArena, Copy> 36.5k ± 0% 36.5k ± 0% -0.02% (p=0.000 n=20+20) BM_Parse_Upb_FileDesc<UseArena, Alias> 36.5k ± 0% 36.5k ± 0% -0.02% (p=0.000 n=20+20) BM_Parse_Proto2<FileDesc, NoArena, Copy> 35.8k ± 0% 35.8k ± 0% ~ (all samples are equal) BM_Parse_Proto2<FileDesc, UseArena, Copy> 65.3k ± 0% 65.3k ± 0% ~ (all samples are equal) name old speed new speed delta BM_LoadAdsDescriptor_Upb<NoLayout> 137MB/s ± 2% 137MB/s ± 4% ~ (p=0.707 n=16+19) BM_LoadAdsDescriptor_Upb<WithLayout> 122MB/s ± 3% 123MB/s ± 3% ~ (p=0.501 n=18+18) BM_LoadAdsDescriptor_Proto2<NoLayout> 64.2MB/s ± 7% 64.7MB/s ± 5% ~ (p=0.330 n=16+18) BM_LoadAdsDescriptor_Proto2<WithLayout> 63.6MB/s ± 3% 63.9MB/s ± 3% ~ (p=0.303 n=18+17) BM_Parse_Upb_FileDesc<UseArena, Copy> 614MB/s ± 4% 613MB/s ± 4% ~ (p=0.935 n=17+18) BM_Parse_Upb_FileDesc<UseArena, Alias> 665MB/s ± 6% 667MB/s ± 3% ~ (p=0.873 n=16+17) BM_Parse_Upb_FileDesc<InitBlock, Copy> 624MB/s ± 4% 622MB/s ± 3% ~ (p=0.501 n=18+18) BM_Parse_Upb_FileDesc<InitBlock, Alias> 681MB/s ± 4% 675MB/s ± 2% ~ (p=0.297 n=18+16) BM_Parse_Proto2<FileDesc, NoArena, Copy> 311MB/s ± 3% 296MB/s ±15% ~ (p=0.177 n=17+20) BM_Parse_Proto2<FileDesc, UseArena, Copy> 649MB/s ± 3% 644MB/s ± 3% ~ (p=0.232 n=17+18) BM_Parse_Proto2<FileDesc, InitBlock, Copy> 656MB/s ± 7% 659MB/s ± 4% ~ (p=0.707 n=18+19) BM_Parse_Proto2<FileDescSV, InitBlock, Alias> 587MB/s ± 5% 576MB/s ±16% ~ (p=0.584 n=18+18) BM_SerializeDescriptor_Proto2 1.32GB/s ± 5% 1.31GB/s ± 7% ~ (p=0.143 n=18+18) BM_SerializeDescriptor_Upb 737MB/s ± 4% 737MB/s ± 7% ~ (p=0.839 n=18+18) ``` PiperOrigin-RevId: 520452349 |
2 years ago |
|
Joshua Haberman
|
73489a9339 |
Removed the cleanup list
It no longer has any users. If we need it later, we can add it back. This saves one pointer of memory from `sizeof(upb_Arena)`. Also, we now allow fuses if if the block allocators differ. This is made possible by cl/520144430, but was not take advantage of in that CL. PiperOrigin-RevId: 520174588 |
2 years ago |
|
Joshua Haberman
|
a73852630f |
Changed Arena representation so that fusing links arenas together instead of blocks.
Previously when fusing, we would concatenate all blocks into a single list that lived in the arena root. From then on, all arenas would add their blocks to this single unified list. After this CL, arenas keep their distinct list of blocks even after being fused. Instead of unifying the block list, fuse now puts the arenas themselves into a list, so all arenas in the fused group can be iterated over at any time. This design makes it easier to keep each individual arena thread-compatible, because fuse and free are now the only mutating operations that touch state that is shared with the entire group. Read-only operations like `SpaceAllocated()` also iterate the list of arenas, but in a read-only fashion. (Note: we need tests for SpaceAllocated(), both single-threaded for correctness and multi-threaded for resilience to crashes and data races). Performance of fuse regresses by 5-20%. This is somewhat expected as we are performing more atomic operations during a fuse. ``` name old cpu/op new cpu/op delta BM_ArenaOneAlloc 18.4ns ± 6% 18.7ns ± 4% +2.00% (p=0.016 n=18+18) BM_ArenaInitialBlockOneAlloc 5.50ns ± 4% 6.57ns ± 4% +19.42% (p=0.000 n=16+17) BM_ArenaFuseUnbalanced/2 59.3ns ±10% 68.7ns ± 4% +15.85% (p=0.000 n=19+19) BM_ArenaFuseUnbalanced/8 479ns ± 5% 540ns ± 8% +12.57% (p=0.000 n=18+19) BM_ArenaFuseUnbalanced/64 4.50µs ± 4% 4.93µs ± 8% +9.59% (p=0.000 n=17+17) BM_ArenaFuseUnbalanced/128 9.24µs ± 3% 9.96µs ± 3% +7.81% (p=0.000 n=17+17) BM_ArenaFuseBalanced/2 63.3ns ±18% 71.0ns ± 4% +12.14% (p=0.000 n=19+18) BM_ArenaFuseBalanced/8 484ns ± 9% 543ns ±10% +12.11% (p=0.000 n=17+16) BM_ArenaFuseBalanced/64 4.50µs ± 6% 4.94µs ± 4% +9.62% (p=0.000 n=19+17) BM_ArenaFuseBalanced/128 9.20µs ± 4% 9.95µs ± 4% +8.12% (p=0.000 n=16+19) BM_LoadAdsDescriptor_Upb<NoLayout> 5.50ms ± 8% 5.69ms ±17% ~ (p=0.189 n=18+19) BM_LoadAdsDescriptor_Upb<WithLayout> 6.10ms ± 5% 6.05ms ± 4% ~ (p=0.258 n=17+18) BM_LoadAdsDescriptor_Proto2<NoLayout> 11.9ms ±15% 11.6ms ± 5% ~ (p=0.589 n=19+16) BM_LoadAdsDescriptor_Proto2<WithLayout> 11.8ms ± 5% 12.4ms ±17% ~ (p=0.604 n=16+20) BM_Parse_Upb_FileDesc<UseArena, Copy> 12.1µs ± 8% 12.1µs ± 4% ~ (p=1.000 n=18+18) BM_Parse_Upb_FileDesc<UseArena, Alias> 11.8µs ±17% 11.1µs ± 4% ~ (p=0.104 n=20+17) BM_Parse_Upb_FileDesc<InitBlock, Copy> 12.0µs ± 5% 11.9µs ± 4% ~ (p=0.134 n=18+19) BM_Parse_Upb_FileDesc<InitBlock, Alias> 10.9µs ± 7% 11.0µs ± 4% ~ (p=0.195 n=17+18) BM_Parse_Proto2<FileDesc, NoArena, Copy> 24.2µs ± 4% 24.4µs ± 7% ~ (p=0.767 n=18+18) BM_Parse_Proto2<FileDesc, UseArena, Copy> 11.6µs ± 5% 11.6µs ± 4% ~ (p=0.621 n=18+16) BM_Parse_Proto2<FileDesc, InitBlock, Copy> 11.3µs ± 3% 11.3µs ± 3% ~ (p=0.743 n=18+18) BM_Parse_Proto2<FileDescSV, InitBlock, Alias> 12.7µs ± 8% 12.7µs ± 4% ~ (p=0.988 n=18+19) BM_SerializeDescriptor_Proto2 5.77µs ± 5% 5.71µs ± 5% ~ (p=0.433 n=17+17) BM_SerializeDescriptor_Upb 10.0µs ± 5% 10.1µs ± 7% ~ (p=0.102 n=19+16) name old time/op new time/op delta BM_ArenaOneAlloc 18.4ns ± 6% 18.8ns ± 4% +1.98% (p=0.019 n=18+18) BM_ArenaInitialBlockOneAlloc 5.51ns ± 4% 6.58ns ± 4% +19.42% (p=0.000 n=16+17) BM_ArenaFuseUnbalanced/2 59.5ns ±10% 68.9ns ± 4% +15.83% (p=0.000 n=19+19) BM_ArenaFuseUnbalanced/8 481ns ± 5% 541ns ± 8% +12.54% (p=0.000 n=18+19) BM_ArenaFuseUnbalanced/64 4.51µs ± 4% 4.94µs ± 8% +9.53% (p=0.000 n=17+17) BM_ArenaFuseUnbalanced/128 9.26µs ± 3% 9.98µs ± 3% +7.79% (p=0.000 n=17+17) BM_ArenaFuseBalanced/2 63.5ns ±19% 71.1ns ± 3% +12.07% (p=0.000 n=19+18) BM_ArenaFuseBalanced/8 485ns ± 9% 551ns ±20% +13.47% (p=0.000 n=17+17) BM_ArenaFuseBalanced/64 4.51µs ± 6% 4.95µs ± 4% +9.62% (p=0.000 n=19+17) BM_ArenaFuseBalanced/128 9.22µs ± 4% 9.97µs ± 4% +8.12% (p=0.000 n=16+19) BM_LoadAdsDescriptor_Upb<NoLayout> 5.52ms ± 8% 5.72ms ±18% ~ (p=0.199 n=18+19) BM_LoadAdsDescriptor_Upb<WithLayout> 6.12ms ± 5% 6.07ms ± 4% ~ (p=0.273 n=17+18) BM_LoadAdsDescriptor_Proto2<NoLayout> 11.9ms ±15% 11.6ms ± 5% ~ (p=0.589 n=19+16) BM_LoadAdsDescriptor_Proto2<WithLayout> 11.9ms ± 5% 12.5ms ±18% ~ (p=0.582 n=16+20) BM_Parse_Upb_FileDesc<UseArena, Copy> 12.2µs ± 8% 12.1µs ± 3% ~ (p=0.963 n=18+18) BM_Parse_Upb_FileDesc<UseArena, Alias> 11.8µs ±17% 11.1µs ± 4% ~ (p=0.104 n=20+17) BM_Parse_Upb_FileDesc<InitBlock, Copy> 12.0µs ± 5% 11.9µs ± 4% ~ (p=0.126 n=18+19) BM_Parse_Upb_FileDesc<InitBlock, Alias> 11.0µs ± 6% 11.1µs ± 4% ~ (p=0.195 n=17+18) BM_Parse_Proto2<FileDesc, NoArena, Copy> 24.3µs ± 4% 24.5µs ± 6% ~ (p=0.743 n=18+18) BM_Parse_Proto2<FileDesc, UseArena, Copy> 11.7µs ± 5% 11.6µs ± 4% ~ (p=0.574 n=18+16) BM_Parse_Proto2<FileDesc, InitBlock, Copy> 11.3µs ± 3% 11.3µs ± 3% ~ (p=0.743 n=18+18) BM_Parse_Proto2<FileDescSV, InitBlock, Alias> 12.7µs ± 8% 12.7µs ± 4% ~ (p=0.988 n=18+19) BM_SerializeDescriptor_Proto2 5.78µs ± 5% 5.73µs ± 5% ~ (p=0.357 n=17+17) BM_SerializeDescriptor_Upb 10.0µs ± 5% 10.1µs ± 7% ~ (p=0.117 n=19+16) name old allocs/op new allocs/op delta BM_ArenaOneAlloc 1.00 ± 0% 1.00 ± 0% ~ (all samples are equal) BM_ArenaFuseUnbalanced/2 2.00 ± 0% 2.00 ± 0% ~ (all samples are equal) BM_ArenaFuseUnbalanced/8 8.00 ± 0% 8.00 ± 0% ~ (all samples are equal) BM_ArenaFuseUnbalanced/64 64.0 ± 0% 64.0 ± 0% ~ (all samples are equal) BM_ArenaFuseUnbalanced/128 128 ± 0% 128 ± 0% ~ (all samples are equal) BM_ArenaFuseBalanced/2 2.00 ± 0% 2.00 ± 0% ~ (all samples are equal) BM_ArenaFuseBalanced/8 8.00 ± 0% 8.00 ± 0% ~ (all samples are equal) BM_ArenaFuseBalanced/64 64.0 ± 0% 64.0 ± 0% ~ (all samples are equal) BM_ArenaFuseBalanced/128 128 ± 0% 128 ± 0% ~ (all samples are equal) BM_LoadAdsDescriptor_Upb<NoLayout> 6.08k ± 0% 6.05k ± 0% -0.54% (p=0.000 n=20+20) BM_LoadAdsDescriptor_Upb<WithLayout> 6.39k ± 0% 6.36k ± 0% -0.55% (p=0.000 n=20+20) BM_LoadAdsDescriptor_Proto2<NoLayout> 83.4k ± 0% 83.4k ± 0% ~ (p=0.800 n=20+20) BM_LoadAdsDescriptor_Proto2<WithLayout> 84.4k ± 0% 84.4k ± 0% ~ (p=0.752 n=20+20) BM_Parse_Upb_FileDesc<UseArena, Copy> 7.00 ± 0% 7.00 ± 0% ~ (all samples are equal) BM_Parse_Upb_FileDesc<UseArena, Alias> 7.00 ± 0% 7.00 ± 0% ~ (all samples are equal) BM_Parse_Proto2<FileDesc, NoArena, Copy> 765 ± 0% 765 ± 0% ~ (all samples are equal) BM_Parse_Proto2<FileDesc, UseArena, Copy> 8.00 ± 0% 8.00 ± 0% ~ (all samples are equal) name old peak-mem(Bytes)/op new peak-mem(Bytes)/op delta BM_ArenaOneAlloc 336 ± 0% 336 ± 0% ~ (all samples are equal) BM_ArenaFuseUnbalanced/2 672 ± 0% 672 ± 0% ~ (all samples are equal) BM_ArenaFuseUnbalanced/8 2.69k ± 0% 2.69k ± 0% ~ (all samples are equal) BM_ArenaFuseUnbalanced/64 21.5k ± 0% 21.5k ± 0% ~ (all samples are equal) BM_ArenaFuseUnbalanced/128 43.0k ± 0% 43.0k ± 0% ~ (all samples are equal) BM_ArenaFuseBalanced/2 672 ± 0% 672 ± 0% ~ (all samples are equal) BM_ArenaFuseBalanced/8 2.69k ± 0% 2.69k ± 0% ~ (all samples are equal) BM_ArenaFuseBalanced/64 21.5k ± 0% 21.5k ± 0% ~ (all samples are equal) BM_ArenaFuseBalanced/128 43.0k ± 0% 43.0k ± 0% ~ (all samples are equal) BM_LoadAdsDescriptor_Upb<NoLayout> 9.89M ± 0% 9.95M ± 0% +0.65% (p=0.000 n=20+20) BM_LoadAdsDescriptor_Upb<WithLayout> 9.95M ± 0% 10.02M ± 0% +0.70% (p=0.000 n=20+20) BM_LoadAdsDescriptor_Proto2<NoLayout> 6.62M ± 0% 6.62M ± 0% ~ (p=0.800 n=20+20) BM_LoadAdsDescriptor_Proto2<WithLayout> 6.66M ± 0% 6.66M ± 0% ~ (p=0.752 n=20+20) BM_Parse_Upb_FileDesc<UseArena, Copy> 36.5k ± 0% 36.5k ± 0% ~ (all samples are equal) BM_Parse_Upb_FileDesc<UseArena, Alias> 36.5k ± 0% 36.5k ± 0% ~ (all samples are equal) BM_Parse_Proto2<FileDesc, NoArena, Copy> 35.8k ± 0% 35.8k ± 0% ~ (all samples are equal) BM_Parse_Proto2<FileDesc, UseArena, Copy> 65.3k ± 0% 65.3k ± 0% ~ (all samples are equal) name old speed new speed delta BM_LoadAdsDescriptor_Upb<NoLayout> 138MB/s ± 7% 132MB/s ±15% ~ (p=0.126 n=18+20) BM_LoadAdsDescriptor_Upb<WithLayout> 124MB/s ± 5% 125MB/s ± 4% ~ (p=0.258 n=17+18) BM_LoadAdsDescriptor_Proto2<NoLayout> 63.9MB/s ±13% 65.2MB/s ± 5% ~ (p=0.589 n=19+16) BM_LoadAdsDescriptor_Proto2<WithLayout> 64.0MB/s ± 5% 61.3MB/s ±15% ~ (p=0.604 n=16+20) BM_Parse_Upb_FileDesc<UseArena, Copy> 620MB/s ± 8% 622MB/s ± 4% ~ (p=1.000 n=18+18) BM_Parse_Upb_FileDesc<UseArena, Alias> 644MB/s ±15% 679MB/s ± 4% ~ (p=0.104 n=20+17) BM_Parse_Upb_FileDesc<InitBlock, Copy> 627MB/s ± 4% 633MB/s ± 4% ~ (p=0.134 n=18+19) BM_Parse_Upb_FileDesc<InitBlock, Alias> 688MB/s ± 6% 682MB/s ± 4% ~ (p=0.195 n=17+18) BM_Parse_Proto2<FileDesc, NoArena, Copy> 310MB/s ± 4% 309MB/s ± 6% ~ (p=0.767 n=18+18) BM_Parse_Proto2<FileDesc, UseArena, Copy> 646MB/s ± 4% 649MB/s ± 4% ~ (p=0.621 n=18+16) BM_Parse_Proto2<FileDesc, InitBlock, Copy> 666MB/s ± 3% 666MB/s ± 3% ~ (p=0.743 n=18+18) BM_Parse_Proto2<FileDescSV, InitBlock, Alias> 592MB/s ± 7% 593MB/s ± 4% ~ (p=0.988 n=18+19) BM_SerializeDescriptor_Proto2 1.30GB/s ± 5% 1.32GB/s ± 5% ~ (p=0.433 n=17+17) BM_SerializeDescriptor_Upb 756MB/s ± 5% 745MB/s ± 6% ~ (p=0.102 n=19+16) ``` PiperOrigin-RevId: 520144430 |
2 years ago |
|
Joshua Haberman
|
9c0bcc0e90 |
Revamped portable atomic layer
Two main improvements: 1. Memory order is an explicit parameter. 2. Functions are type-generic. PiperOrigin-RevId: 520074974 |
2 years ago |
|
Joshua Haberman
|
6d6ab90ece |
Enable arena test (it was previously #ifdef'd away accidentally)
PiperOrigin-RevId: 519782179 |
2 years ago |
|
Joshua Haberman
|
d450990631 |
Allow for fuse/free races in `upb_Arena`.
Implementation is by kfm@, I only added the portability code around it. `upb_Arena` was designed to be only thread-compatible. However, fusing of arenas muddies the waters somewhat, because two distinct `upb_Arena` objects will end up sharing state when fused. This causes a `upb_Arena_Free(a)` to interfere with `upb_Arena_Fuse(b, c)` if `a` and `b` were previously fused. It turns out that we can use atomics to fix this with about a 35% regression in fuse performance (see below). Arena create+free does not regress, thanks to special-case logic in Free(). `upb_Arena` is still a thread-compatible type, and it is still never safe to call `upb_Arena_xxx(a)` and `upb_Arena_yyy(a)` in parallel. However you can at least now call `upb_Arena_Free(a)` and `upb_Arena_Fuse(b, c)` in parallel, even if `a` and `b` were previously fused. Note that `upb_Arena_Fuse(a, b)` and `upb_Arena_Fuse(c, d)` is still not allowed if `b` and `c` were previously fused. In practice this means that fuses must still be single-threaded within a single fused group. Performance results: ``` name old cpu/op new cpu/op delta BM_ArenaOneAlloc 18.6ns ± 1% 18.6ns ± 1% ~ (p=0.726 n=18+17) BM_ArenaInitialBlockOneAlloc 6.28ns ± 1% 5.73ns ± 1% -8.68% (p=0.000 n=17+20) BM_ArenaFuseUnbalanced/2 44.1ns ± 2% 60.4ns ± 1% +37.05% (p=0.000 n=18+19) BM_ArenaFuseUnbalanced/8 370ns ± 2% 500ns ± 1% +35.12% (p=0.000 n=19+20) BM_ArenaFuseUnbalanced/64 3.52µs ± 1% 4.71µs ± 1% +33.80% (p=0.000 n=18+19) BM_ArenaFuseUnbalanced/128 7.20µs ± 1% 9.72µs ± 2% +34.93% (p=0.000 n=16+19) BM_ArenaFuseBalanced/2 44.4ns ± 2% 61.4ns ± 1% +38.23% (p=0.000 n=20+17) BM_ArenaFuseBalanced/8 373ns ± 2% 509ns ± 1% +36.57% (p=0.000 n=19+17) BM_ArenaFuseBalanced/64 3.55µs ± 2% 4.79µs ± 1% +34.80% (p=0.000 n=19+19) BM_ArenaFuseBalanced/128 7.26µs ± 1% 9.76µs ± 1% +34.45% (p=0.000 n=17+19) BM_LoadAdsDescriptor_Upb<NoLayout> 5.66ms ± 1% 5.69ms ± 1% +0.57% (p=0.013 n=18+20) BM_LoadAdsDescriptor_Upb<WithLayout> 6.30ms ± 1% 6.36ms ± 1% +0.90% (p=0.000 n=19+18) BM_LoadAdsDescriptor_Proto2<NoLayout> 12.1ms ± 1% 12.1ms ± 1% ~ (p=0.118 n=18+18) BM_LoadAdsDescriptor_Proto2<WithLayout> 12.2ms ± 1% 12.3ms ± 1% +0.50% (p=0.006 n=18+18) BM_Parse_Upb_FileDesc<UseArena, Copy> 12.7µs ± 1% 12.7µs ± 1% ~ (p=0.194 n=20+19) BM_Parse_Upb_FileDesc<UseArena, Alias> 11.6µs ± 1% 11.6µs ± 1% ~ (p=0.192 n=20+20) BM_Parse_Upb_FileDesc<InitBlock, Copy> 12.5µs ± 1% 12.5µs ± 0% ~ (p=0.750 n=18+14) BM_Parse_Upb_FileDesc<InitBlock, Alias> 11.4µs ± 1% 11.3µs ± 1% -0.34% (p=0.046 n=19+19) BM_Parse_Proto2<FileDesc, NoArena, Copy> 25.4µs ± 1% 25.7µs ± 2% +1.37% (p=0.000 n=18+18) BM_Parse_Proto2<FileDesc, UseArena, Copy> 12.1µs ± 2% 12.1µs ± 1% ~ (p=0.143 n=18+18) BM_Parse_Proto2<FileDesc, InitBlock, Copy> 11.9µs ± 3% 11.9µs ± 1% ~ (p=0.076 n=17+19) BM_Parse_Proto2<FileDescSV, InitBlock, Alias> 13.2µs ± 1% 13.2µs ± 1% ~ (p=0.053 n=19+19) BM_SerializeDescriptor_Proto2 5.97µs ± 4% 5.90µs ± 4% ~ (p=0.093 n=17+19) BM_SerializeDescriptor_Upb 10.4µs ± 1% 10.4µs ± 1% ~ (p=0.909 n=17+18) name old time/op new time/op delta BM_ArenaOneAlloc 18.7ns ± 2% 18.6ns ± 0% ~ (p=0.607 n=18+17) BM_ArenaInitialBlockOneAlloc 6.29ns ± 1% 5.74ns ± 1% -8.71% (p=0.000 n=17+19) BM_ArenaFuseUnbalanced/2 44.1ns ± 1% 60.6ns ± 1% +37.21% (p=0.000 n=17+19) BM_ArenaFuseUnbalanced/8 371ns ± 2% 500ns ± 1% +35.02% (p=0.000 n=19+16) BM_ArenaFuseUnbalanced/64 3.53µs ± 1% 4.72µs ± 1% +33.85% (p=0.000 n=18+19) BM_ArenaFuseUnbalanced/128 7.22µs ± 1% 9.73µs ± 2% +34.87% (p=0.000 n=16+19) BM_ArenaFuseBalanced/2 44.5ns ± 2% 61.5ns ± 1% +38.22% (p=0.000 n=20+17) BM_ArenaFuseBalanced/8 373ns ± 2% 510ns ± 1% +36.58% (p=0.000 n=19+16) BM_ArenaFuseBalanced/64 3.56µs ± 2% 4.80µs ± 1% +34.87% (p=0.000 n=19+19) BM_ArenaFuseBalanced/128 7.27µs ± 1% 9.77µs ± 1% +34.40% (p=0.000 n=17+19) BM_LoadAdsDescriptor_Upb<NoLayout> 5.67ms ± 1% 5.71ms ± 1% +0.60% (p=0.011 n=18+20) BM_LoadAdsDescriptor_Upb<WithLayout> 6.32ms ± 1% 6.37ms ± 1% +0.87% (p=0.000 n=19+18) BM_LoadAdsDescriptor_Proto2<NoLayout> 12.1ms ± 1% 12.2ms ± 1% ~ (p=0.126 n=18+19) BM_LoadAdsDescriptor_Proto2<WithLayout> 12.2ms ± 1% 12.3ms ± 1% +0.51% (p=0.002 n=18+18) BM_Parse_Upb_FileDesc<UseArena, Copy> 12.7µs ± 1% 12.7µs ± 1% ~ (p=0.149 n=20+19) BM_Parse_Upb_FileDesc<UseArena, Alias> 11.6µs ± 1% 11.6µs ± 1% ~ (p=0.211 n=20+20) BM_Parse_Upb_FileDesc<InitBlock, Copy> 12.5µs ± 1% 12.5µs ± 1% ~ (p=0.986 n=18+15) BM_Parse_Upb_FileDesc<InitBlock, Alias> 11.4µs ± 1% 11.3µs ± 1% ~ (p=0.081 n=19+18) BM_Parse_Proto2<FileDesc, NoArena, Copy> 25.4µs ± 1% 25.8µs ± 2% +1.41% (p=0.000 n=18+18) BM_Parse_Proto2<FileDesc, UseArena, Copy> 12.1µs ± 2% 12.1µs ± 1% ~ (p=0.558 n=19+18) BM_Parse_Proto2<FileDesc, InitBlock, Copy> 12.0µs ± 3% 11.9µs ± 1% ~ (p=0.165 n=17+19) BM_Parse_Proto2<FileDescSV, InitBlock, Alias> 13.2µs ± 1% 13.2µs ± 1% ~ (p=0.070 n=19+19) BM_SerializeDescriptor_Proto2 5.98µs ± 4% 5.92µs ± 3% ~ (p=0.138 n=17+19) BM_SerializeDescriptor_Upb 10.4µs ± 1% 10.4µs ± 1% ~ (p=0.858 n=17+18) ``` PiperOrigin-RevId: 518573683 |
2 years ago |
|
Matt Kulukundis
|
5bc3cae2d7 |
Add threading tests for arenas
PiperOrigin-RevId: 518009578 |
2 years ago |
|
Eric Salo
|
f458e05718 |
add some UPB_API / UPB_API_INTERNAL declarations
PiperOrigin-RevId: 492551734 |
2 years ago |
|
Eric Salo
|
e70b102387 |
delete upb_Arena_Alloc()
PiperOrigin-RevId: 491682231 |
2 years ago |
|
Eric Salo
|
e932f7931a |
clean up the arena code a bit more:
_upb_Arena_FastMalloc() has been directly inlined Arena::allocator() has been removed mem_block is now _upb_MemBlock some calls to upb_Arena_Alloc() (which is going away) have been changed to use upb_Arena_Malloc() instead PiperOrigin-RevId: 489874377 |
2 years ago |
|
Eric Salo
|
f6307877d3 |
move portability stuff into upb/port/
Also delete redundant system #includes that are already pulled in by port/def.inc PiperOrigin-RevId: 486398989 |
2 years ago |
|
Eric Salo
|
f409c9985b |
create mem/ subdir for all arena code
PiperOrigin-RevId: 485884460 |
2 years ago |