ENH: Use AVX for float32 implementation of np.sin & np.cos#13368

Merged

mattip merged 11 commits intonumpy:masterfrom

r-devulap:sincos-simd

Aug 18, 2019

Commits on Aug 3, 2019

ENH: Use AVX for float32 implementation of np.sin & np.cos
Raghuveer Devulapalli
committed
BUG: fixing NAN handling and adding tests for sin/cos
Raghuveer Devulapalli
committed
BUG: sin and cos cast float16 to float32
Raghuveer Devulapalli
committed
TEST: adding tests to validate AVX sin/cos implementation
Raghuveer Devulapalli
committed
BUG: use strides and process strided arrays using AVX
Raghuveer Devulapalli
committed
BUG: AVX2 impl of sin/cos requires an FMA
Raghuveer Devulapalli
committed
TEST: Rounding max tolerable ulp error to an int
Raghuveer Devulapalli
committed
BUG: eliminate unsed variables warning in cpuid
Raghuveer Devulapalli
committed
MAINT: using an enum to switch between sin/cos
Raghuveer Devulapalli
committed
TEST: improving test coverage for sin/cos for input > 117435.992f
Raghuveer Devulapalli
committed
BUG: rename avx2_scalef_ps to fma_scalef_ps
Raghuveer Devulapalli
committed