Regressions in System.Security.Cryptography.Tests.Perf_Rfc2898DeriveBytes

[performanceautofiler]

Run Information

Architecture	x64
OS	alpine 3.12
Baseline	ae24a8f8dd745e24384f19c22f85b28aa1f33cd6
Compare	1615de6ccc1290be168c1de290429489c489efde
Diff	Diff

Regressions in System.Security.Cryptography.Tests.Perf_Rfc2898DeriveBytes

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio	Baseline ETL	Compare ETL
DeriveBytes - Duration of single invocation	4.70 ms	5.72 ms	1.22	0.02	True

Test Report

Repro

git clone https://github.com/dotnet/performance.git
py .\performance\scripts\benchmarks_ci.py -f net6.0 --filter 'System.Security.Cryptography.Tests.Perf_Rfc2898DeriveBytes*'

Payloads

Baseline
Compare

Histogram

System.Security.Cryptography.Tests.Perf_Rfc2898DeriveBytes.DeriveBytes

Description of detection logic

IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsRegressionBase: Marked as regression because the compare was 5% greater than the baseline, and the value was not too small.
IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsRegressionWindowed: Marked as regression because 5.717147306089743 > 4.941921660661057.
IsChangePoint: Marked as a change because one of 2/14/2022 7:55:17 PM, 2/22/2022 12:12:47 AM falls between 2/13/2022 12:11:49 PM and 2/22/2022 12:12:47 AM.
IsRegressionStdDev: Marked as regression because -51.537669114894946 (T) = (0 -5621127.050861655) / Math.Sqrt((575541557.8977764 / (22)) + (13139154971.515217 / (45))) is less than -1.997137908391408 = MathNet.Numerics.Distributions.StudentT.InvCDF(0, 1, (22) + (45) - 2, .025) and -0.1955079817183199 = (4701873.293043458 - 5621127.050861655) / 4701873.293043458 is less than -0.05.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsChangeEdgeDetector: Marked as regression because Edge Detector said so.

Docs

Profiling workflow for dotnet/runtime repository
Benchmarking workflow for dotnet/runtime repository

[dotnet-issue-labeler]

I couldn't figure out the best area label to add to this issue. If you have write-permissions please help me learn by adding exactly one area label.

[kunalspathak]

The commit range is https://github.com/dotnet/runtime/compare/3a4ade30d6ae45cf8e884692efe6a7f39062e1ab...45a9548644f34264ce2e1a8a586e8e234d6f61e4...Most likely the suspect is 45a9548 bartonjs

Tagging subscribers to this area: @dotnet/area-system-security, @vcsjones
See info in area-owners.md if you want to be subscribed.

Issue Details

---

Run Information

Architecture	x64
OS	alpine 3.12
Baseline	ae24a8f8dd745e24384f19c22f85b28aa1f33cd6
Compare	1615de6ccc1290be168c1de290429489c489efde
Diff	Diff

Regressions in System.Security.Cryptography.Tests.Perf_Rfc2898DeriveBytes

Benchmark	Baseline	Test	Test/Base	Test Quality	Edge Detector	Baseline IR	Compare IR	IR Ratio	Baseline ETL	Compare ETL
DeriveBytes - Duration of single invocation	4.70 ms	5.72 ms	1.22	0.02	True

Test Report

Repro

git clone https://github.com/dotnet/performance.git
py .\performance\scripts\benchmarks_ci.py -f net6.0 --filter 'System.Security.Cryptography.Tests.Perf_Rfc2898DeriveBytes*'

Payloads

Baseline
Compare

Histogram

System.Security.Cryptography.Tests.Perf_Rfc2898DeriveBytes.DeriveBytes

Description of detection logic

IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsRegressionBase: Marked as regression because the compare was 5% greater than the baseline, and the value was not too small.
IsRegressionChecked: Marked as regression because the three check build points were 0.05 greater than the baseline.
IsRegressionWindowed: Marked as regression because 5.717147306089743 > 4.941921660661057.
IsChangePoint: Marked as a change because one of 2/14/2022 7:55:17 PM, 2/22/2022 12:12:47 AM falls between 2/13/2022 12:11:49 PM and 2/22/2022 12:12:47 AM.
IsRegressionStdDev: Marked as regression because -51.537669114894946 (T) = (0 -5621127.050861655) / Math.Sqrt((575541557.8977764 / (22)) + (13139154971.515217 / (45))) is less than -1.997137908391408 = MathNet.Numerics.Distributions.StudentT.InvCDF(0, 1, (22) + (45) - 2, .025) and -0.1955079817183199 = (4701873.293043458 - 5621127.050861655) / 4701873.293043458 is less than -0.05.
IsImprovementBase: Marked as not an improvement because the compare was not 5% less than the baseline, or the value was too small.
IsChangeEdgeDetector: Marked as regression because Edge Detector said so.

Docs

Profiling workflow for dotnet/runtime repository
Benchmarking workflow for dotnet/runtime repository

Author:	performanceautofiler[bot]
Assignees:	-
Labels:	area-System.Security, untriaged, refs/heads/main, x64, Regression, RunKind=micro, CoreClr, alpine 3.12
Milestone:	-

[vcsjones]

Well, that's 10,000 HMAC invocations.

https://github.com/dotnet/performance/blob/d7dac8a7ca12a28d099192f8a901cf8e30361384/src/benchmarks/micro/libraries/System.Security.Cryptography/Perf.Rfc2898DeriveBytes.cs#L13

So 20,000 additional calls to ERR_clear_error. Each iteration does CryptoNative_HmacUpdate and CryptoNative_HmacFinal. If it regressed by about 1ms, then that means each invocation of ERR_clear_error added about 50ns of overhead, assuming that is the culprit. @bartonjs does that line up with your expectations?

[bartonjs]

Yeah. I'm not really worried about that differential in isolation.

According to
https://pvscmdupload.blob.core.windows.net/reports/allTestHistory%2frefs%2fheads%2fmain_x64_ubuntu%2018.04%2fSystem.Security.Cryptography.Tests.Perf_Rfc2898DeriveBytes.DeriveBytes.html vs
https://pvscmdupload.blob.core.windows.net/reports/allTestHistory%2frefs%2fheads%2fmain_x64_Windows%2010.0.18362%2fSystem.Security.Cryptography.Tests.Perf_Rfc2898DeriveBytes.DeriveBytes.html

the operation is still faster on Linux than Windows (with the same hardware).

The scale of the perf degradation, combined with PBKDF2 itself not being used in a loop (except by password breakers) makes me feel like this is an acceptable consequence of keeping the error queue better managed.