Flipwind: reduce roundoff in phase winding a[n] calc (#534) #319

Job	Run time
prepare	2s
Build with Pip (ubuntu-22.04, 3.8, llvm, -march=native)	1m 6s
Build with Pip (ubuntu-22.04, 3.8, llvm, -march=x86-64)	1m 12s
Build with Pip (ubuntu-22.04, 3.8, gcc, -march=native)	52s
Build with Pip (ubuntu-22.04, 3.8, gcc, -march=x86-64)	51s
Build with Pip (ubuntu-22.04, 3.11, llvm, -march=native)	1m 12s
Build with Pip (ubuntu-22.04, 3.11, llvm, -march=x86-64)	1m 7s
Build with Pip (ubuntu-22.04, 3.11, gcc, -march=native)	1m 58s
Build with Pip (ubuntu-22.04, 3.11, gcc, -march=x86-64)	54s
Build with Pip (windows-2022, 3.8, msvc, /arch:AVX2)	11m 36s
Build with Pip (windows-2022, 3.8, msvc, /arch:SSE2)	11m 23s
Build with Pip (windows-2022, 3.8, llvm, /arch:AVX2)	5m 2s
Build with Pip (windows-2022, 3.8, llvm, /arch:SSE2)	4m 32s
Build with Pip (windows-2022, 3.11, msvc, /arch:AVX2)	11m 58s
Build with Pip (windows-2022, 3.11, msvc, /arch:SSE2)	12m 31s
Build with Pip (windows-2022, 3.11, llvm, /arch:AVX2)	3m 58s
Build with Pip (windows-2022, 3.11, llvm, /arch:SSE2)	4m 13s
Build with Pip (macos-13, 3.8, llvm, -march=native)	2m 50s
Build with Pip (macos-13, 3.8, llvm, -march=x86-64)	2m 26s
Build with Pip (macos-13, 3.8, gcc-14, -march=native)	1m 32s
Build with Pip (macos-13, 3.8, gcc-14, -march=x86-64)	1m 59s
Build with Pip (macos-13, 3.11, llvm, -march=native)	4m 28s
Build with Pip (macos-13, 3.11, llvm, -march=x86-64)	5m 1s
Build with Pip (macos-13, 3.11, gcc-14, -march=native)	1m 37s
Build with Pip (macos-13, 3.11, gcc-14, -march=x86-64)	1m 44s
	1h 36m 4s

Provide feedback