diff options
author | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-05-15 03:35:49 +0000 |
---|---|---|
committer | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-05-15 03:35:49 +0000 |
commit | d8bbc7858622b6d9c278469aab701ca0b609cddf (patch) | |
tree | eff41dc61d9f714852212739e6b3738b82a2af87 /third_party/rust/suggest/src/benchmarks/README.md | |
parent | Releasing progress-linux version 125.0.3-1~progress7.99u1. (diff) | |
download | firefox-d8bbc7858622b6d9c278469aab701ca0b609cddf.tar.xz firefox-d8bbc7858622b6d9c278469aab701ca0b609cddf.zip |
Merging upstream version 126.0.
Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>
Diffstat (limited to '')
-rw-r--r-- | third_party/rust/suggest/src/benchmarks/README.md | 28 |
1 files changed, 28 insertions, 0 deletions
diff --git a/third_party/rust/suggest/src/benchmarks/README.md b/third_party/rust/suggest/src/benchmarks/README.md new file mode 100644 index 0000000000..45150d8413 --- /dev/null +++ b/third_party/rust/suggest/src/benchmarks/README.md @@ -0,0 +1,28 @@ +# Suggest benchmarking code + +Use `cargo suggest-bench` to run these benchmarks. + +The main benchmarking code lives here, while the criterion integration code lives in the `benches/` +directory. + +## Benchmarks + +### ingest-[provider-type] + +Time it takes to ingest all suggestions for a provider type on an empty database. +The bechmark downloads network resources in advance in order to exclude the network request time +from these measurements. + +### Benchmarks it would be nice to have + +- Ingestion with synthetic data. This would isolate the benchmark from changes to the RS database. +- Fetching suggestions + +## cargo suggest-debug-ingestion-sizes + +Run this to get row counts for all database tables. This can be very useful for improving +benchmarks, since targeting the tables with the largest number of rows will usually lead to the +largest improvements. + +The command also prints out the size of all remote-settings attachments, which can be good to +optimize on its own since it represents the amount of data user's need to download. |