daniel-adam-tfs commented on PR #654:
URL: https://github.com/apache/arrow-go/pull/654#issuecomment-3835880748
Benchmarks comparison on my Intel macbook:
```
goos: darwin
goarch: amd64
pkg: github.com/apache/arrow-go/v18/parquet/internal/encoding
cpu: Intel(R) Core(TM) i5-1038NG7 CPU @ 2.00GHz
│
bench/BenchmarkDecodeByteStreamSplitBatchWidth4.main.txt │
bench/BenchmarkDecodeByteStreamSplitBatchWidth4.V2.txt │
bench/BenchmarkDecodeByteStreamSplitBatchWidth4.AVX2.txt │
│
sec/op │ sec/op vs base
│ sec/op vs base │
DecodeByteStreamSplitBatchWidth4/nValues=8-8
204.65n ± 0% 12.78n ± 0% -93.76%
(p=0.000 n=10) 11.50n ± 1% -94.38% (p=0.000 n=10)
DecodeByteStreamSplitBatchWidth4/nValues=64-8
395.550n ± 1% 77.810n ± 1% -80.33%
(p=0.000 n=10) 7.710n ± 1% -98.05% (p=0.000 n=10)
DecodeByteStreamSplitBatchWidth4/nValues=512-8
1425.50n ± 1% 516.15n ± 1% -63.79%
(p=0.000 n=10) 34.59n ± 2% -97.57% (p=0.000 n=10)
DecodeByteStreamSplitBatchWidth4/nValues=4096-8
9459.5n ± 2% 4073.0n ± 1% -56.94%
(p=0.000 n=10) 226.8n ± 2% -97.60% (p=0.000 n=10)
DecodeByteStreamSplitBatchWidth4/nValues=32768-8
73.885µ ± 0% 32.318µ ± 1% -56.26%
(p=0.000 n=10) 2.688µ ± 1% -96.36% (p=0.000 n=10)
DecodeByteStreamSplitBatchWidth4/nValues=262144-8
593.79µ ± 1% 261.89µ ± 1% -55.89%
(p=0.000 n=10) 37.92µ ± 2% -93.61% (p=0.000 n=10)
geomean
6.026µ 1.614µ -73.21%
203.4n -96.62%
│
bench/BenchmarkDecodeByteStreamSplitBatchWidth4.main.txt │
bench/BenchmarkDecodeByteStreamSplitBatchWidth4.V2.txt │
bench/BenchmarkDecodeByteStreamSplitBatchWidth4.AVX2.txt │
│
B/s │ B/s vs base
│ B/s vs base │
DecodeByteStreamSplitBatchWidth4/nValues=8-8
149.1Mi ± 0% 2387.9Mi ± 0% +1501.65%
(p=0.000 n=10) 2653.2Mi ± 1% +1679.58% (p=0.000 n=10)
DecodeByteStreamSplitBatchWidth4/nValues=64-8
617.2Mi ± 1% 3137.6Mi ± 1% +408.36%
(p=0.000 n=10) 31664.9Mi ± 1% +5030.38% (p=0.000 n=10)
DecodeByteStreamSplitBatchWidth4/nValues=512-8
1.338Gi ± 1% 3.695Gi ± 1% +176.19%
(p=0.000 n=10) 55.147Gi ± 2% +4021.74% (p=0.000 n=10)
DecodeByteStreamSplitBatchWidth4/nValues=4096-8
1.613Gi ± 2% 3.747Gi ± 1% +132.26%
(p=0.000 n=10) 67.269Gi ± 2% +4070.21% (p=0.000 n=10)
DecodeByteStreamSplitBatchWidth4/nValues=32768-8
1.652Gi ± 0% 3.777Gi ± 1% +128.62%
(p=0.000 n=10) 45.410Gi ± 1% +2648.52% (p=0.000 n=10)
DecodeByteStreamSplitBatchWidth4/nValues=262144-8
1.645Gi ± 1% 3.729Gi ± 1% +126.73%
(p=0.000 n=10) 25.752Gi ± 2% +1465.82% (p=0.000 n=10)
geomean
916.7Mi 3.342Gi +273.33%
26.52Gi +2862.04%
│
bench/BenchmarkDecodeByteStreamSplitBatchWidth4.main.txt │
bench/BenchmarkDecodeByteStreamSplitBatchWidth4.V2.txt │
bench/BenchmarkDecodeByteStreamSplitBatchWidth4.AVX2.txt │
│
B/op │ B/op vs base
│ B/op vs base │
DecodeByteStreamSplitBatchWidth4/nValues=8-8
80.00 ± 0% 0.00 ± 0% -100.00% (p=0.000
n=10) 0.00 ± 0% -100.00% (p=0.000 n=10)
DecodeByteStreamSplitBatchWidth4/nValues=64-8
96.00 ± 0% 0.00 ± 0% -100.00% (p=0.000
n=10) 0.00 ± 0% -100.00% (p=0.000 n=10)
DecodeByteStreamSplitBatchWidth4/nValues=512-8
112.0 ± 0% 0.0 ± 0% -100.00% (p=0.000
n=10) 0.0 ± 0% -100.00% (p=0.000 n=10)
DecodeByteStreamSplitBatchWidth4/nValues=4096-8
112.0 ± 0% 0.0 ± 0% -100.00% (p=0.000
n=10) 0.0 ± 0% -100.00% (p=0.000 n=10)
DecodeByteStreamSplitBatchWidth4/nValues=32768-8
112.0 ± 0% 0.0 ± 0% -100.00% (p=0.000
n=10) 0.0 ± 0% -100.00% (p=0.000 n=10)
DecodeByteStreamSplitBatchWidth4/nValues=262144-8
112.0 ± 0% 0.0 ± 0% -100.00% (p=0.000
n=10) 0.0 ± 0% -100.00% (p=0.000 n=10)
geomean
103.2 ?
¹ ² ? ¹ ²
¹ summaries must be >0 to compute geomean
² ratios must be >0 to compute geomean
│
bench/BenchmarkDecodeByteStreamSplitBatchWidth4.main.txt │
bench/BenchmarkDecodeByteStreamSplitBatchWidth4.V2.txt │
bench/BenchmarkDecodeByteStreamSplitBatchWidth4.AVX2.txt │
│
allocs/op │ allocs/op vs base
│ allocs/op vs base │
DecodeByteStreamSplitBatchWidth4/nValues=8-8
1.000 ± 0% 0.000 ± 0% -100.00% (p=0.000
n=10) 0.000 ± 0% -100.00% (p=0.000 n=10)
DecodeByteStreamSplitBatchWidth4/nValues=64-8
3.000 ± 0% 0.000 ± 0% -100.00% (p=0.000
n=10) 0.000 ± 0% -100.00% (p=0.000 n=10)
DecodeByteStreamSplitBatchWidth4/nValues=512-8
3.000 ± 0% 0.000 ± 0% -100.00% (p=0.000
n=10) 0.000 ± 0% -100.00% (p=0.000 n=10)
DecodeByteStreamSplitBatchWidth4/nValues=4096-8
3.000 ± 0% 0.000 ± 0% -100.00% (p=0.000
n=10) 0.000 ± 0% -100.00% (p=0.000 n=10)
DecodeByteStreamSplitBatchWidth4/nValues=32768-8
3.000 ± 0% 0.000 ± 0% -100.00% (p=0.000
n=10) 0.000 ± 0% -100.00% (p=0.000 n=10)
DecodeByteStreamSplitBatchWidth4/nValues=262144-8
3.000 ± 0% 0.000 ± 0% -100.00% (p=0.000
n=10) 0.000 ± 0% -100.00% (p=0.000 n=10)
geomean
2.498 ?
¹ ² ? ¹ ²
```
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]