Test of SIMD for ADC
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
VectorTests.xcodeproj
.gitignore
CompilerExplorer.cpp
MathOps_AVX.h
MathOps_AVX512.h
MathOps_Common.h
MathOps_Neon.h
MathOps_SSE.h
MathOps_Scalar.h
Prefix_FMA.h
README.md
VecDispatchTest.cpp
VecDispatchTest.s
VectorTests.sln
VectorTests.vcxproj
VectorTests.vcxproj.filters
VectorTests.vcxproj.user
XBenchmark.h
XDSP.h
XDispatch_AVX2_XBasicAmp.cpp
XDispatch_FMA_XBasicAmp.cpp
XHelpers.h
XIOAdapter.h
XStopWatch.h
YBasicAmp.h
YEQFilter.h
YFilterLadder.h
YPannerAmp.h
YSawOsc.h
iacaMarks.h
target_build.xcconfig

README.md

###############

Copyright 2012-2017, Angus F. Hewlett for FXpansion and ROLI

This code should be considered a demo / proof-of-concept, not a library.

Templated SIMD audio signal processing examples for scalar, NEON, SSE(2,4.2), AVX(1,2), AVX512.

Demonstrates an approach to exploiting data level (SIMD) fine-grained parallelism & ILP for parallel audio engines

e.g. synthesizer voice architectures or mixing consoles.

"Write once, run everywhere", re-optimise the same code for different hardware feature-sets, voice counts, cache conditions,

data layouts etc. - fast thanks to extensive compile time code generation (C++ templates).

Recommended tools:

* Compiler Explorer

* IACA

* A good profiler - VTune, Instruments, Glowcode etc.

Drop me a line if interested in licensing a fuller featured version.

###############