apitech/loki - loki - Gitea: Git with a cup of tea

Commit Graph

Author	SHA1	Message	Date
Owen Diehl	a4db1a4359	Headblock interop (#3995 ) * merge feature/unordered-replay * interoperable head chunks	4 years ago
Cyril Tovena	4a8f62ba00	Improve head chunk allocations when reading samples. (#3968 ) * Improve head chunk allocations when reading samples. Reading sample is the worst because we are not limited and must return all of them. Re-using memory will helps ingesters read path. ``` ❯ benchcmp before.txt after.txt benchmark old ns/op new ns/op delta BenchmarkHeadBlockSampleIterator/Size_20000-16 2935102 1568140 -46.57% BenchmarkHeadBlockSampleIterator/Size_10000-16 1441980 750856 -47.93% BenchmarkHeadBlockSampleIterator/Size_8000-16 1146807 602796 -47.44% BenchmarkHeadBlockSampleIterator/Size_5000-16 698638 370655 -46.95% benchmark old allocs new allocs delta BenchmarkHeadBlockSampleIterator/Size_20000-16 32 16 -50.00% BenchmarkHeadBlockSampleIterator/Size_10000-16 29 15 -48.28% BenchmarkHeadBlockSampleIterator/Size_8000-16 28 15 -46.43% BenchmarkHeadBlockSampleIterator/Size_5000-16 26 15 -42.31% benchmark old bytes new bytes delta BenchmarkHeadBlockSampleIterator/Size_20000-16 2269664 484068 -78.67% BenchmarkHeadBlockSampleIterator/Size_10000-16 1073582 984 -99.91% BenchmarkHeadBlockSampleIterator/Size_8000-16 827825 819 -99.90% BenchmarkHeadBlockSampleIterator/Size_5000-16 475557 786 -99.83% ``` Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Improve unorderedHeadBlock too. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com>	4 years ago
Owen Diehl	23ad7f2e61	Unordered head block (#3957 ) * speccing out unordered head block * testware & unordered serialise * common utils for iter & sampleIter * more generic forEntries * more efficient unordedHeadChunk serialise (no internal re-casting) * roundtripping unordered head block, exit headchunk iteration early, add constant for current default chunk version * adds head block write benchmarks for ordered & unordered writes * fixes bench * removes unused initializer * gofmt * linting	5 years ago

Author

SHA1

Message

Date

Owen Diehl

a4db1a4359

Headblock interop (#3995 )

* merge feature/unordered-replay

* interoperable head chunks

4 years ago

Cyril Tovena

4a8f62ba00

Improve head chunk allocations when reading samples. (#3968 )

* Improve head chunk allocations when reading samples.

Reading sample is the worst because we are not limited and must return all of them.
 Re-using memory will helps ingesters read path.

```
❯ benchcmp  before.txt after.txt
benchmark                                          old ns/op     new ns/op     delta
BenchmarkHeadBlockSampleIterator/Size_20000-16     2935102       1568140       -46.57%
BenchmarkHeadBlockSampleIterator/Size_10000-16     1441980       750856        -47.93%
BenchmarkHeadBlockSampleIterator/Size_8000-16      1146807       602796        -47.44%
BenchmarkHeadBlockSampleIterator/Size_5000-16      698638        370655        -46.95%

benchmark                                          old allocs     new allocs     delta
BenchmarkHeadBlockSampleIterator/Size_20000-16     32             16             -50.00%
BenchmarkHeadBlockSampleIterator/Size_10000-16     29             15             -48.28%
BenchmarkHeadBlockSampleIterator/Size_8000-16      28             15             -46.43%
BenchmarkHeadBlockSampleIterator/Size_5000-16      26             15             -42.31%

benchmark                                          old bytes     new bytes     delta
BenchmarkHeadBlockSampleIterator/Size_20000-16     2269664       484068        -78.67%
BenchmarkHeadBlockSampleIterator/Size_10000-16     1073582       984           -99.91%
BenchmarkHeadBlockSampleIterator/Size_8000-16      827825        819           -99.90%
BenchmarkHeadBlockSampleIterator/Size_5000-16      475557        786           -99.83%
```

Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com>

* Improve unorderedHeadBlock too.

Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com>

4 years ago

Owen Diehl

23ad7f2e61

Unordered head block (#3957 )

* speccing out unordered head block

* testware & unordered serialise

* common utils for iter & sampleIter

* more generic forEntries

* more efficient unordedHeadChunk serialise (no internal re-casting)

* roundtripping unordered head block, exit headchunk iteration early, add constant for current default chunk version

* adds head block write benchmarks for ordered & unordered writes

* fixes bench

* removes unused initializer

* gofmt

* linting

5 years ago

3 Commits (4c2feb32b55dc0431aff659607b67e4ac078325e)