GF2P8MULB — Galois Field Multiply Bytes

Opcode/Instruction Op/En 64/32 bit Mode Support CPUID Feature Flag Description
66 0F38 CF /r GF2P8MULB xmm1, xmm2/m128 A V/V GFNI Multiplies elements in the finite field GF(2^8).

Instruction Operand Encoding

Op/En Tuple Operand 1 Operand 2 Operand 3 Operand 4
A NA ModRM:reg (r, w) ModRM:r/m (r) NA NA


The instruction multiplies elements in the finite field GF(28), operating on a byte (field element) in the first source operand and the corresponding byte in a second source operand. The field GF(28) is represented in polynomial representation with the reduction polynomial x8 + x4 + x3 + x + 1.

This instruction does not support broadcasting.

The SSE encoded forms of the instruction require16B alignment on their memory operations.


define gf2p8mul_byte(src1byte, src2byte):
    tword ← 0
    FOR i ← 0 to 7:
        IF src2byte.bit[i]:
            tword ← tword XOR (src1byte<< i)
        * carry out polynomial reduction by the characteristic polynomial p*
    FOR i ← 14 downto 8:
        p ← 0x11B << (i-8) *0x11B = 0000_0001_0001_1011 in binary*
        IF tword.bit[i]:
            tword ← tword XOR p
return tword.byte[0]

GF2P8MULB srcdest, src1 (128b SSE encoded version)

FOR j ← 0 TO 15:
    SRCDEST.byte[j] ← gf2p8mul_byte(SRCDEST.byte[j], SRC1.byte[j])

Intel C/C++ Compiler Intrinsic Equivalent

VGF2P8MULB __m128i _mm_gf2p8mul_epi8(__m128i, __m128i);
VGF2P8MULB __m128i _mm_mask_gf2p8mul_epi8(__m128i, __mmask16, __m128i, __m128i);
VGF2P8MULB __m128i _mm_maskz_gf2p8mul_epi8(__mmask16, __m128i, __m128i);

SIMD Floating-Point Exceptions


Other Exceptions

Legacy-encoded: Exceptions Type 4.