SVE Instruction List by Dougall Johnson
FMAXNMP: Floating-point maximum number pairwise
FMAXNMP Zdn.H, Pg/M, Zdn.H, Zm.H (SVE2 (SME
svfloat16_t svmaxnmp[_f16]_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
128-bit SVE
Compare adjacent pairs of 16-bit float elements and set even elements of (3) to the maximum number (i.e. non-QNaN) across pairs from (1), and the odd elements to the maximum number across pairs from (2).
256-bit SVE
Compare adjacent pairs of 16-bit float elements and set even elements of (3) to the maximum number (i.e. non-QNaN) across pairs from (1), and the odd elements to the maximum number across pairs from (2).
512-bit SVE
Compare adjacent pairs of 16-bit float elements and set even elements of (3) to the maximum number (i.e. non-QNaN) across pairs from (1), and the odd elements to the maximum number across pairs from (2).
Larger sizes
1024-bit SVE
Compare adjacent pairs of 16-bit float elements and set even elements of (3) to the maximum number (i.e. non-QNaN) across pairs from (1), and the odd elements to the maximum number across pairs from (2).
2048-bit SVE
Compare adjacent pairs of 16-bit float elements and set even elements of (3) to the maximum number (i.e. non-QNaN) across pairs from (1), and the odd elements to the maximum number across pairs from (2).
Report mistakes or give feedback
Inspired by and based on the x86/x64 SIMD Instruction List by Daytime.