BSTART.TGEMV

1// PTO-INSTRUCTION: {"assembly":["BSTART.TGEMV DataType"],"block":[],"catalog_indices":[34],"catalog_records":[{"asm":"BSTART.TGEMV DataType","constraints":[{"field":"DataType","operator":"one-of","values":[0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,16,17,18,19,20,24,25,26,27,28]}],"encoding":[{"index":0,"mask":"0x07ffffff","match":"0x01031181","width_bits":32}],"encoding_kind":"L32","fields":[{"name":"DataType","pieces":[{"instruction_lsb":27,"value_lsb":0,"width":5}],"signedness":"encoding-defined","width":5}],"form_id":"bstart_tgemv_32_ae19f5b678f5","length_bits":32,"mnemonic":"BSTART.TGEMV","semantic_family":"CMD","semantic_group":"Bundle Split","semantic_handler":"ExecuteBundleStart","semantic_summary":"Starts CUBE Function 16 for the TGEMV M=1 Matrix-vector complete-bundle operation.","status":"accepted"}],"classification":["execution"],"contract":{"block_composition":["BSTART.TGEMV AType","B.DATR BType, RMode, Sat (optional; BType defaults to AType)","B.FPATR PreQuantMode, ReluMode, GroupNCode, RowMaxEn, GroupMaxEn, RowMaxInit, MaxAbsEn, TransA, TransB, CScaleEn (exactly one)","B.DIM LB0 M (optional, default 1; TGEMV permits only M=1)","B.DIM LB1 N (optional, default 1)","B.DIM LB2 K (optional, default 1)","B.IOT ordered Local mathematical sources: A CUBE_M16/M32 primary, B CUBE_N8 primary","B.IOT D matching A's CUBE_M16/M32 layout, optional RowMaxOut, optional GroupMaxOut destinations","B.IOT/B.IOR postprocess operands selected by B.FPATR","BSTOP or the next BSTART completion boundary"],"canonical_assembly":["BSTART.TGEMV DataType"],"defaults":["Encoded DataType is always AType. Omitted B.DATR preserves AType as BType, selects RNE, and disables saturation.","Omitted LB0, LB1, and LB2 default M, N, and K independently to one; TGEMV fixes M to one.","Exactly one all-zero B.FPATR selects no conversion, activation, or reduction; B.IOR and auxiliary B.IOT operands exist only when a selected postprocess mode requires them.","Local A uses persistent CUBE_M16 or CUBE_M32, Local B uses persistent CUBE_N8, and D is newly allocated in A's M layout. M is fixed to one; N and K are arbitrary positive values independent of per-PE TSize.","TransA=0 and TransB=0 select no logical transpose. TGEMV requires both controls to remain zero."],"encoding_class":"standalone-encoded","examples":["BSTART.TGEMV AType; B.DATR BType, RMode, Sat (optional; BType defaults to AType); B.FPATR PreQuantMode, ReluMode, GroupNCode, RowMaxEn, GroupMaxEn, RowMaxInit, MaxAbsEn, TransA, TransB, CScaleEn (exactly one); B.DIM LB0 M (optional, default 1; TGEMV permits only M=1); B.DIM LB1 N (optional, default 1); B.DIM LB2 K (optional, default 1); B.IOT ordered Local mathematical sources: A CUBE_M16/M32 primary, B CUBE_N8 primary; B.IOT D matching A's CUBE_M16/M32 layout, optional RowMaxOut, optional GroupMaxOut destinations; B.IOT/B.IOR postprocess operands selected by B.FPATR; BSTOP or the next BSTART completion boundary"],"exceptions":["A reserved DataType or fixed-bit mismatch raises Fault_IllegalInstruction before block state changes.","Missing, duplicate, or non-Matrix B.FPATR use raises Fault_BundleControl before allocation or payload effects.","Illegal types, dimensions, masks, binding streams, descriptors, shapes, capacities, aliases, readiness, or postprocess values raise Fault_TileLegality before source snapshots and effects."],"field_contracts":{"DataType":{"ref":"PTO-FIELD-BLOCK-DATATYPE"}},"field_zero_meanings":{"DataType":"Encoded zero selects FP64."},"legality":["The carrier selects exactly CUBE Function 16 and TileOperation_TGEMV.","Local A uses persistent CUBE_M16 or CUBE_M32, Local B uses persistent CUBE_N8, and D is newly allocated in A's M layout. M is fixed to one; N and K are arbitrary positive values independent of per-PE TSize.","TGEMV is Local-only: TransA and TransB are zero and every effective Shared binding rejects before effects.","AType and BType must be supported ordinary Matrix types from one numeric class. M is fixed to one and every Shared binding is illegal.","Every common nonzero four-bit PE_MASK is legal; all four PEs complete cooperative Shared readiness while only selected PEs allocate and publish. Mask zero is a strict no-op before descriptor reads, faults, allocation, readiness checks, or lifetime effects.","B.DATR permits only BType, RMode, and Sat. Exactly one B.FPATR is mandatory and closes the conditional postprocess schema."],"memory_effects":["none"],"operands":[{"field":"DataType","role":"tile element data type selector"}],"ordering":["Complete schema, field, type, dimension, descriptor, shape, capacity, readiness, alias, and allocation preflight precedes every source snapshot and destination effect.","D and every enabled reduction output publish as one atomic group; rejection publishes none and successful sources persist."],"standalone_opcode":true,"state_effects":["Start a CUBE Function 16 descriptor with encoded DataType preserved as AType.","At block completion execute TileOperation_TGEMV using the resolved M, N, K, input types, mathematical operands, and B.FPATR postprocess schema.","Publish the complete output group atomically after successful preflight and computation; do not consume mathematical or postprocess sources.","For Local execution, publish D with A's CUBE_M16 or CUBE_M32 layout and final output dtype; ordinary Bias, MX scales, and enabled reduction auxiliaries keep their operation-owned layouts."]},"depends_on":["PTO-BLOCK-MODEL-SCHEMA-PROFILE-ENCODING"],"id":"PTO-BLOCK-BSTART-TGEMV","mnemonic":"BSTART.TGEMV","summary":"Starts CUBE Function 16 for the TGEMV M=1 Matrix-vector complete-bundle operation.","surface":"block"}2// PTO-REVIEW: {"review_method":"formal-definition-read","outcome":"FORMAL-COMPLETE","reviewed_fields":["assembly","encoding","defaults","operation","state","memory","ordering","faults","reserved"]}3// NDF-BEGIN: PTO-BSTART-TGEMV-CONTRACT-0014// ndf: kind=contract level=L1 layer=block status=accepted5// BSTART.TGEMV MUST select CUBE Function 16 and TileOperation_TGEMV.6// Its encoded DataType is AType; omitted B.DATR MUST preserve AType as7// BType, omitted LB0/LB1/LB2 MUST default M/N/K independently to one,8// Local N and K MUST be positive and independent of per-PE TSize; M9// MUST equal one.10// AType and BType must be supported ordinary Matrix types from one numeric11// class. M is fixed to one and every Shared binding is illegal.12// TransA and TransB MUST remain zero and every Shared binding MUST reject.13// Every common nonzero Local PE mask is legal; mask zero is a strict no-op.14// Exactly one B.FPATR MUST close postprocess defaults and operands.15// Complete schema, type, shape, capacity, definedness, readiness, alias,16// and allocation preflight MUST precede source snapshots and effects;17// successful outputs MUST publish atomically and sources MUST persist.18// NDF-END: PTO-BSTART-TGEMV-CONTRACT-00119
20
21// DOC-BEGIN: decode22readonly func InstructionContractMatches_BSTART_TGEMV(23    operation: CommandOperation) => boolean24begin25    return operation ==26        CommandOperation_bstart_tgemv_32_ae19f5b678f5;27end;28// DOC-END: decode29// DOC-BEGIN: operation30readonly func InstructionContractTileOperation_BSTART_TGEMV()31    => TileOperation32begin33    return TileOperation_TGEMV;34end;35
36readonly func InstructionContractCubeFunction_BSTART_TGEMV()37    => integer {0..31}38begin39    return 16;40end;41
42readonly func InstructionContractSharedOperandsAllowed_BSTART_TGEMV()43    => boolean44begin45    return FALSE;46end;47
48readonly func InstructionContractHandler_BSTART_TGEMV()49    => CommandSemanticHandler50begin51    return CommandHandler_ExecuteBundleStart;52end;53// DOC-END: operation54