v_mfma_f32_16x16x4bf16_1k GPU Native ISA AMD Matrix

Vector Packed Arithmetic

v_mfma_f32_16x16x4bf16_1k

Multiply the 16x4 matrix in the first input by the 4x16 matrix in the second input and add the 16x16 matrix in the third input using fused multiply…

Also written as v_mfma_f32_16x16x4_4b_bf16, v_mfma_f32_16x16x4bf16. AMD's machine-readable ISA specification lists these names for the same instruction.

Encoding

Binary Layout (AMD CDNA 2)
BLGP
63:61
ACC
60:59
SRC2
58:50
SRC1
49:41
SRC0
40:32
110100111
31:23
1100100
22:16
ACC_CD
15
ABID
14:11
CBSZ
10:8
VDST
7:0
 

The VOP3P_MFMA layout from AMD's machine-readable ISA specification (AMD CDNA 2). Opcode 100 in OP rebuilds 0x00000000D3E40000, an identifier AMD lists for this encoding. No other generation in the specification defines this instruction.

Format VOP3P
Width 64 bits
Opcode 100
Identifier 0x00000000D3E40000

Operands

  • VDST
    Written. N/A Data format: N/A (OPR_VGPR_OR_ACCVGPR, FMT_NUM_PK16_F32)
  • SRC0
    Read. N/A Data format: N/A (OPR_SRC_VGPR_OR_ACCVGPR, FMT_NUM_PK4_BF16)
  • SRC1
    Read. N/A Data format: N/A (OPR_SRC_VGPR_OR_ACCVGPR, FMT_NUM_PK4_BF16)
  • SRC2
    Read. N/A Data format: N/A (OPR_SRC_VGPR_OR_ACCVGPR_OR_CONST, FMT_NUM_PK16_F32)

In AMD's order, as its machine-readable ISA specification lists them for the VOP3P_MFMA encoding (AMD CDNA 2).

GFX Target Compatibility

Per-target GFX compatibility has not yet been verified for this instruction.

Related

More in Vector Packed Arithmetic

Reference

AMDGPU / GFX ISA

Description

Multiply the 16x4 matrix in the first input by the 4x16 matrix in the second input and add the 16x16 matrix in the third input using fused multiply add. Store the resulting matrix into vector registers.

Example

v_mfma_f32_16x16x4bf16_1k v[0:15], v[0:1], v[2:3], -2.0

A real instruction accepted by the LLVM assembler, taken verbatim from LLVM's own AMDGPU MC test suite (mai-gfx90a.s). Not from an AMD document, and not authored here.

Sources