Issues

• NEON u16 / u32 use MAX_INDEX = u8::MAX, and short chunks make NEON 32-bit argmin / argmax faster

  jvdd/argminmax#123 · Filter by author jvdd opened on Oct 16, 2026

• perf: group the vectors of the masked SIMD core (as in #88)

  enhancement
  jvdd/argminmax#109 · Filter by author jvdd opened on Oct 8, 2026

• Use SSE for 64-bit dtypes on x86 CPUs without AVX2?

  enhancement
  jvdd/argminmax#108 · Filter by author jvdd opened on Oct 8, 2026

• Benchmarks: control the alignment of the data

  jvdd/argminmax#95 · Filter by author jvdd opened on Oct 6, 2026

• Support axis-wise argmin / argmax for n-D ndarray arrays

  enhancement
  jvdd/argminmax#89 · Filter by author jvdd opened on Oct 6, 2026

• Look into using SVE & SVE2

  enhancement
  jvdd/argminmax#59 · Filter by author jvdd opened on Jun 7, 2023

• Make return type Option<usize>

  unsure
  jvdd/argminmax#27 · Filter by author jvdd opened on Feb 17, 2023

• ♻️ use portable SIMD to simplify code

  enhancement
  jvdd/argminmax#8 · Filter by author jvdd opened on Dec 28, 2022