Benchmarking implementation #2813
Conversation
| let indexName = IndexName.op_Implicit("benchmark-reports") | ||
| let typeName = TypeName.op_Implicit("benchmarkreport") | ||
|
|
||
| type Memory(gen0Collections:int, gen1Collections: int, gen2Collections: int, totalOperations:int64, bytesAllocatedPerOperation:int64) = |
There was a problem hiding this comment.
Do we have plans to create a custom Elasticssearch BenchMarkDotNet exporter?
https://github.com/dotnet/BenchmarkDotNet/blob/master/docs/guide/Configs/Exporters.md
That would be such a cool nuget package
There was a problem hiding this comment.
While this works and I know we have these types because the json type provider garbled certain types I rather in the long run we don't maintain these POCO's as part of our build script.
There was a problem hiding this comment.
I completely agree, it's really not ideal - I will switch to just using the raw JSON output at some point.
There was a problem hiding this comment.
In terms of the exporter, it might be a nicer way for us to get the data into ES than using the F# script :)
There was a problem hiding this comment.
++ I think this is good to merge now for progression over perfection and build an exporter afterwards
| @@ -70,12 +72,8 @@ module Build = | |||
| compileCore incremental | |||
|
|
|||
| let Clean() = | |||
| match (quickBuild, getBuildParam "target" = "clean") with | |||
There was a problem hiding this comment.
Why was this removed? this section helped when running tests from the command line to rely on incremental builds. Incremental builds are not that fast on .NET core 1.1 but they still help a great deal.
| DotNetCli.RunCommand(fun p -> | ||
| { p with | ||
| WorkingDir = testsProjectDirectory | ||
| }) "run -f net46 -c Release Benchmark" |
There was a problem hiding this comment.
Should we hardcode net46 here? The fake command line parser snoops -all off target names so we can e.g do built test and build test-all or build integrate-all which would tests all the TFM's we support.
Should we do something similar for benchmark ?
There was a problem hiding this comment.
Discussed offline. The net46 framework is only the framework under which to run Tests.exe and the benchmark switcher. The framework used for benchmarking tests is chose through the benchmark configuration for the chosen benchmark.
|
Appveyor is not happy because it can not find the DocGenerator https://ci.appveyor.com/project/Mpdreamz/elasticsearch-net/build/1.0.245 Output path changes not completely in sync with: #2806 |
Broken with the move to new new csproj
Move definitions for index and type to top of file so they are easier to read.
codebrain
force-pushed
the
feature/master-profiling-benchmarking
branch
from
4fa7dae to
2485f1b
Compare
Renamed TestMain to Main to allow for benchmarking, if you'd like to run tests in VS rename back to TestMain.
Added benchmarking capabilities; - Control over run counts - Collects memory diagnostics - Store results in an ES instance (cherry picked from commit eca4335)
Added benchmarking capabilities; - Control over run counts - Collects memory diagnostics - Store results in an ES instance (cherry picked from commit eca4335)
Added benchmarking capabilities; - Control over run counts - Collects memory diagnostics - Store results in an ES instance
First PR!
Benchmarking