tannergooding
Follow
Pinned
1,364 contributions in the last year
We're celebrating 100 million developers!
Activity overview
Contributed to
dotnet/runtime,
dotnet/ClangSharp,
terrafx/terrafx.interop.windows
and 43 other
repositories
Contribution activity
January 2023
Created 26 commits in 4 repositories
Created a pull request in dotnet/runtime that received 19 comments
Change the default loop unroll limit to 4
Today, the JIT defaults to 1 unless the upper limit is a Vector###<T>.Count property in which case it will try to unroll provided the method body i…
+6
−18
•
19
comments
Opened 17 other pull requests in 3 repositories
dotnet/runtime
1
open
14
merged
- Fixing a comment attached to DEFAULT_UNROLL_LOOP_MAX_ITERATION_COUNT
- Add Lerp to IFloatingPointIeee754
-
Update the JIT to track
Span.LengthandReadOnlySpan.Lengthas "never negative" - Handle Span<T>.Length and ReadOnlySpan<T>.Length as intrinsic
- Ensure various scalar cross platform helper APIs are handled directly as intrinsic
- Refactor lookupNamedIntrinsic to avoid common duplicate checks
- Update our inline observations to special case a few more intrinsics
- Fix some low-hanging HWIntrinsic issues
- Have SIMD Load/Store use GT_IND and GT_ASG where possible
- Add a debug only config knob to control the maximum force inline depth
- Fix a build error around the ConstantExpected analyzer
- Consistently use fgMakeMultiUse in the gtNewSimd*Node APIs
- Annotate platform specific hardware intrinsics with the ConstExpected attribute
- SimdAsHWIntrinsic improvements and cleanup
- Rewrite how Matrix3x2 and Matrix4x4 are implemented
terrafx/terrafx.interop.mimalloc
1
merged
terrafx/terrafx
1
merged
Reviewed 34 pull requests in 4 repositories
dotnet/runtime
25 pull requests
- Sprinkle some more readonly around
- Change the default loop unroll limit to 4
- Initial support for zmm in .NET
-
Eliminate a bounds check in
NumberBuffer - Vectorize ProbabilisticMap.IndexOfAny
- Implement DivRem intrinsic for X86
- Updating JitInterface to support Vector512.
- Adding the Vector512 and Vector512<T> types
- Arm64: Implement VectorTableLookup intrinsinsics + Consecutive registers support
- Add missing case for upper save/restore
-
Update the JIT to track
Span.LengthandReadOnlySpan.Lengthas "never negative" - [JIT] Fixed assertion for targetType and op1 type not being the same
- Handle Span<T>.Length and ReadOnlySpan<T>.Length as intrinsic
- Add MemoryExtensions.Count
- Reduce cost of checking OptimizationEnabled
-
Fold
static readonly VectorXto GT_CNS_VEC - Fold some expressions early in importer
- Ensure various scalar cross platform helper APIs are handled directly as intrinsic
- Fix overwriting of superpmi log file
- Increase superpmi-replay x64 parallelism
- Update our inline observations to special case a few more intrinsics
- Consistently use fgMakeMultiUse in the gtNewSimd*Node APIs
- Have SIMD Load/Store use GT_IND and GT_ASG where possible
- Enable AVX512 Additional 16 SIMD Registers
- JIT: Fall back to a canonical SIMD handle in gtGetStructHandleForSimdOrHW
- Some pull request reviews not shown.
dotnet/docs
2 pull requests
dakersnar/runtime
1 pull request
dotnet/machinelearning
1 pull request
Created an issue in dotnet/runtime that received 5 comments
tracing/eventpipe/eventsourceerror/eventsourceerror/eventsourceerror failure
Build Information Build: https://dev.azure.com/dnceng-public/cbb18261-c48f-4abb-8651-8cdcb5474649/_build/results?buildId=137937 Build error leg or …
5
comments
Opened 6 other issues in 3 repositories
dotnet/runtime
1
closed
2
open
EgorBo/Disasmo
2
open
dotnet/jitutils
1
closed
25
contributions
in private repositories
Jan 20 – Jan 26