dotnet / dotnet/machinelearning
Perf: Optimization for DoubleParser (VTune)
Open
area-Core
Priority:2
tenet-performance
- Dominant language
- C#
- Stars
- 9.4k
- Forks
- 2k
- Avg merge
- 2d 20h
- Merged PRs (30d)
- 11
Description
Benchmarking using VTune has found several bottlenecks in parsing Doubles from string in text reader
Perhaps we should consider optimizing this method:
For example, for training a FM it takes almost 37% of all clock cycles:
Function | Module | CPU Time
-- | -- | --
Microsoft::ML::Internal::Utilities::DoubleParser::TryParseCore | Microsoft.ML.Core.dll | 27.390s
CalculateGradientAndUpdateNative | factorizationmachinenative.dll | 22.609s
HelperImpl::FetchNextField | Microsoft.ML.Data.dll | 13.826s
CalculateIntermediateVariablesNative | factorizationmachinenative.dll | 12.201s
Microsoft::ML::Internal::Utilities::DoubleParser::TryParse | Microsoft.ML.Core.dll | 9.666s
Contributor guide
Assessment
This issue has not been assessed yet.