SVE Instruction List by Dougall Johnson
BFMLALB (vectors): BFloat16 floating-point multiply-add long to single-precision (bottom)
BFMLALB Zda.S, Zn.H, Zm.H (SVE+BF16 (SME+BF16
svfloat32_t svbfmlalb[_f32](svfloat32_t op1, svbfloat16_t op2, svbfloat16_t op3)
128-bit SVE
For each even BFloat16 calculate (1) * (2), and add that to the 32-bit float from (3), then set (4) to the result.
256-bit SVE
For each even BFloat16 calculate (1) * (2), and add that to the 32-bit float from (3), then set (4) to the result.
512-bit SVE
For each even BFloat16 calculate (1) * (2), and add that to the 32-bit float from (3), then set (4) to the result.
Larger sizes
1024-bit SVE
For each even BFloat16 calculate (1) * (2), and add that to the 32-bit float from (3), then set (4) to the result.
2048-bit SVE
For each even BFloat16 calculate (1) * (2), and add that to the 32-bit float from (3), then set (4) to the result.
Report mistakes or give feedback
Inspired by and based on the x86/x64 SIMD Instruction List by Daytime.