agenium-scale/nsimd

Issues

• Windows build is broken

  #108 opened 5 months ago by rafaelspring
  0

• There is not ‘&’ bitwise operator for ’packl‘ type and no conversion of 'pack' and 'packl' type

  #107 opened 3 years ago by Panhaolin2001
  0

• NSIMD generates too much assembly

  #106 opened 3 years ago by raphaelthegreat
  5

• NSIMD SVE intrinsics generates movprfx

  #86 opened 3 years ago by gquintin
  3

• Provide += operators etc.

  #88 opened 3 years ago by eschnett
  0

• Provide fabs, fmax, fmin

  #90 opened 3 years ago by eschnett
  2

• Use Sleef math functions to complete NSIMD

  #91 opened 3 years ago by gquintin
  0

• Document "andnot"

  #96 opened 3 years ago by eschnett
  0

• Provide a C11 API for core

  #85 opened 3 years ago by gquintin
  0

• Support Reg Size /2

  #103 opened 4 years ago by ustcsq
  1

• Vector<nsimd::pack<T>>

  #102 opened 4 years ago by ustcsq
  4

• Performance different on X86 and ARM service

  #100 opened 4 years ago by luoming17
  2

• Support for WebAssembly SIMD

  #81 opened 4 years ago by omnisip
  35

• Feature request: allow logical arguments for if_else1

  #99 opened 4 years ago by eschnett
  2

• Calculate i32 horizontal sum in nsimd.

  #97 opened 4 years ago by luoming17
  3

• How to use SSE2 128-bit pack in my project?

  #75 opened 4 years ago by luoming17
  2

• Compile nsimd-2.2 on the ARM server

  #95 opened 4 years ago by Pancrase
  2

• Provide a constexpr size function

  #94 opened 4 years ago by eschnett
  4

• Provide more "inline" operators such as copysign, isfinite, etc...

  #92 opened 4 years ago by eschnett
  2

• Implement a `flipsign` function

  #93 opened 4 years ago by eschnett
  1

• Add "copysign" intrinsic

  #32 opened 4 years ago by eschnett
  1

• mask_for_loop_tail produces scalar code

  #89 opened 4 years ago by eschnett
  2

• nsimd defines `i64` etc. in global namespace

  #84 opened 4 years ago by eschnett
  3

• Provide cbrt

  #87 opened 4 years ago by eschnett
  1

• Self-test errors on macOS

  #78 opened 4 years ago by eschnett
  19

• [setup.bat] change NSTOOLS_URL in setup.sh/setup.bat

  #80 opened 4 years ago by ThomasRetornaz
  1

• Add support for CMake (CMakeLists.txt and FindNsimd.cmake as a start)

  #82 opened 4 years ago by NK-Nikunj
  6

• Cannot build 2.0 release

  #83 opened 4 years ago by eschnett
  6

• Too many arguments for format string

  #67 opened 4 years ago by eschnett
  0

• Three tests failing on MacOS (Intel, AVX2)

  #68 opened 4 years ago by eschnett
  1

• Masked vector store functions?

  #23 opened 4 years ago by eschnett
  3

• When regenerating files, only write those files that changed

  #29 opened 4 years ago by eschnett
  4

• Don't use unions for reinterpreting data

  #34 opened 4 years ago by eschnett
  6

• Documentation on building/running tests is probably out of date.

  #49 opened 4 years ago by DenisYaroshevskiy
  3

• Add documentation for to_pack, to_pack_interleave, get_pack and scoped_aligned_mem

  #64 opened 4 years ago by vianney-st
  0

• Debug output in test cases?

  #70 opened 4 years ago by eschnett
  2

• Compilation issue with msvc 2015 on 32 bits target

  #76 opened 4 years ago by fabienhe
  10

• Autodetect CPU architecture

  #26 opened 5 years ago by eschnett
  1

• Usage of SVE pack with vector

  #66 opened 5 years ago by NK-Nikunj
  3

• Citing NSIMD in a research paper

  #72 opened 5 years ago by NK-Nikunj
  3

• Error while building Nsimd with AARCH64

  #65 opened 5 years ago by NK-Nikunj
  2

• Right shift operation not working

  #62 opened 5 years ago by NK-Nikunj
  2

• Add is_aligned for both C and C++ APIs

  #58 opened 5 years ago by vianney-st
  0

• Support all NSIMD operators for the fixed_point module

  #47 opened 5 years ago by aarnaud1
  0

• Tests should special case rsqrt8

  #39 opened 5 years ago by eschnett
  2

• Strange intrinsic name "round_to_even"

  #40 opened 5 years ago by eschnett
  2

• Separate human-maintained and autogenerated code

  #27 opened 5 years ago by eschnett
  0

• Why rounding to zero when converting from f32 to f16?

  #38 opened 5 years ago by eschnett
  1

• Speed up Python code generation

  #36 opened 5 years ago by eschnett
  0

• Installs into "include/include"

  #22 opened 5 years ago by eschnett
  3