Commit Graph

9 Commits

Author SHA1 Message Date
Craig Tiller 4e666c740b Internal change
PiperOrigin-RevId: 590780731
2023-12-13 20:06:30 -08:00
Tanvi Jagtap 8019903a44 Internal configuration cleanup
PiperOrigin-RevId: 590528536
2023-12-14 03:42:30 +00:00
Craig Tiller b7077f4bbf
[hpack] Rollforward huffman read optimization (#33657)
Rollforward in first commit, fixes in subsequent.
2023-07-11 15:26:54 -07:00
Craig Tiller 57c697d8ae
Revert "[hpack] Huffman read optimization" (#33655)
Reverts grpc/grpc#33269
2023-07-11 12:58:49 -07:00
Craig Tiller 4ce51fe45d
[hpack] Huffman read optimization (#33269)
In real services most of our time ends up in the `Read1()` function,
which populates one byte into the bit buffer.
Change this to read in as many as possible bytes at a time into that
buffer.

Additionally, generate all possible (to some depth) parser geometries,
and add a benchmark for them. Run that benchmark and select the best
geometry for decoding base64 strings (since this is the main use-case).

(gives about a 30% speed boost parsing base64 then huffman encoded
random binary strings)

---------

Co-authored-by: ctiller <ctiller@users.noreply.github.com>
2023-07-11 08:41:15 -07:00
Yash Tibrewal 5f6c357e74
Move some targets from //: to //src/core: (#31402)
* temp

* temp

* temp

* top BUILD builds

* everything builds

* buildify

* temp fix_build_deps change

* fix gpr_atm

* additional changes

* fix build files

* nofixdeps

* Generate projects

* json_args

* fix context

* Generate projects

* Sprinkle some avoid_dep tags

* Run fix_build_deps

* Remove additional avoid_dep

* Revert some targets

* grpc_mock_cel
2022-10-20 20:12:53 -07:00
Craig Tiller 92f58c18a8
Revert "Revert "[chttp2] Improve huffman decode efficiency (#30479)" (#30961)" (#30962)
This reverts commit bfea47093f.
2022-09-13 13:16:55 -07:00
Craig Tiller bfea47093f
Revert "[chttp2] Improve huffman decode efficiency (#30479)" (#30961)
This reverts commit 6c7f921f5f.
2022-09-13 12:25:00 -07:00
Craig Tiller 6c7f921f5f
[chttp2] Improve huffman decode efficiency (#30479)
* first pass

* refinement #1

* fix

* opt

* 8

* 5

* 12

* 10

* 7

* 15

* 8n

* 15n

* 8a

* 16a

* 15a

* 5a

* 10a

* 7a

* 9a

* 12a

* 11a

* 11b

* 11b

* 7b

* 15b

* 14b

* 15i

* 12i

* 9i

* 10i

* finalizing

* fix

* tweak

* 10f

* 10r

* 7r

* 8r

* 9r

* more

* rething

* 8s

* 9s

* 12s

* 7s

* 10s

* 15snr

* 12snr

* 14snr

* 13snr

* 11snr

* 20snr

* 15snr

* 18snr

* 7snr

* 7snr

* 12-2stp

* 13-2stp

* 14-2stp

* 11-2stp

* 10-2stp

* 9-3stp

* 8-3stp

* 7-3stp

* 8-2stp

* auto-tune1

* max-depth-2

* max-depth-3

* max-depth-2

* abbreviate

* working version

* add benchmark

* fix

* fix

* fix

* ditch define

* better code layout

* static fns

* fix

* Automated change: Fix sanity tests

* Automated change: Fix sanity tests

* fix build

* cleanup and comment

* fmt

* fix

* test

* tag

* Automated change: Fix sanity tests

* fix

* fix

* Automated change: Fix sanity tests

* fix

* fix

* fix

* progress

* progress

* progress

* fixes

* cleanup

* Automated change: Fix sanity tests

* Automated change: Fix sanity tests

* fmt

* fix

* fix

* Automated change: Fix sanity tests

* fix

* fix

* Automated change: Fix sanity tests

* fix

* add encode/decode fuzzer

* Automated change: Fix sanity tests

* fix

* fix

* fix

* review feedback

Co-authored-by: ctiller <ctiller@users.noreply.github.com>
2022-09-13 11:24:35 -07:00