TROWSUM

Instruction Diagram

Introduction

Reduces each row by summing its columns.

Mathematical Semantics

Assume R = src.GetValidRow() and C = src.GetValidCol(). For 0 <= i < R:

$$\mathrm{dst}_{i,0} = \sum_{j=0}^{C-1} \mathrm{src}_{i,j}$$

Assembly Syntax

Synchronous form:

%dst = trowsum %src : !pto.tile<...> -> !pto.tile<...>

An internal temporary tile may be introduced during reduction; the C++ built-in API requires an explicit tmp operand.

AS Level 1 (SSA)

%dst = pto.trowsum %src, %tmp : (!pto.tile<...>, !pto.tile<...>) -> !pto.tile<...>

AS Level 2 (DPS)

pto.trowsum ins(%src, %tmp : !pto.tile_buf<...>, !pto.tile_buf<...>) outs(%dst : !pto.tile_buf<...>)

C++ Built-in APIs

Declared in include/pto/common/pto_instr.hpp:

The public include header is <pto/pto-inst.hpp>, and the internal declaration is located in pto/common/pto_instr.hpp.

template <typename TileDataOut, typename TileDataIn, typename TileDataTmp, typename... WaitEvents>
PTO_INST RecordEvent TROWSUM(TileDataOut &dst, TileDataIn &src, TileDataTmp &tmp, WaitEvents &... events);

Constraints

General Constraints or Checks

• dst and src must both be TileType::Vec.
• src must use the standard ND layout: row-major and non-fractal (BLayout::RowMajor, SLayout::NoneBox).
• dst must be a non-fractal layout (!isBoxedLayout) and use one of the following two layouts:
  • ND layout (BLayout::RowMajor, SLayout::NoneBox), or
  • DN layout with the number of columns strictly equal to 1 (BLayout::ColMajor, SLayout::NoneBox, Cols == 1).
• dst and src must have the same element type.
• Runtime valid region checks:
  • src.GetValidRow() != 0
  • src.GetValidCol() != 0
  • src.GetValidRow() == dst.GetValidRow()
• The built-in API signature requires explicitly passing the tmp operand.

Implementation Check for Atlas A2/A3 Training Products/Atlas A2/A3 Inference Products

• Supported element types (Atlas A2/A3 training products/Atlas A2/A3 inference products): half, float, int32_t, int16_t.
• Supported element types (Ascend 950PR/Ascend 950DT): half, float, int32_t, int16_t.
• The implementation accepts both ND output and DN output with Cols == 1, not only DN output.
• Runtime checks follow the shared row reduction check path:
  • src.GetValidRow() != 0
  • src.GetValidCol() != 0
  • src.GetValidRow() == dst.GetValidRow()

Temporary Space

Atlas A2/A3 Training Products/Atlas A2/A3 Inference Products

tmp is used as scratch storage for row reduction.

• For integer types (int32_t, int16_t): tmp is used as a row-wise accumulator buffer (1 block). For each row, tmp is initialized to 0, and then the blocks of src are accumulated through vadd. The final sum result is read from tmp in scalar mode.
  • tmp size: at least 1 row and BLOCK_BYTE_SIZE / sizeof(T) columns (8 for int32_t, 16 for int16_t).
• For floating-point types (float, half): tmp is used for binary tree reduction through vcadd/vcgadd.
  • Safe default settings: set tmp to the same shape as src.

Ascend 950PR/Ascend 950DT

tmp is accepted by the API but is not used by the Ascend 950PR/Ascend 950DT implementation. The Ascend 950PR/Ascend 950DT backend uses vector-register-based reduction (the vcadd instruction) and does not require scratch tile storage. tmp is retained in the C++ built-in API signature only for API compatibility with the Atlas A2/A3 training products/Atlas A2/A3 inference products.

Examples

Automatic

#include <pto/pto-inst.hpp>

using namespace pto;

void example_auto() {
  using SrcT = Tile<TileType::Vec, float, 16, 16>;
  using DstT = Tile<TileType::Vec, float, 16, 1, BLayout::ColMajor>;
  using TmpT = Tile<TileType::Vec, float, 16, 16>;
  SrcT src;
  DstT dst;
  TmpT tmp;
  TROWSUM(dst, src, tmp);
}

Manual

#include <pto/pto-inst.hpp>

using namespace pto;

void example_manual() {
  using SrcT = Tile<TileType::Vec, float, 16, 16>;
  using DstT = Tile<TileType::Vec, float, 16, 1, BLayout::ColMajor>;
  using TmpT = Tile<TileType::Vec, float, 16, 16>;
  SrcT src;
  DstT dst;
  TmpT tmp;
  TASSIGN(src, 0x1000);
  TASSIGN(dst, 0x2000);
  TASSIGN(tmp, 0x3000);
  TROWSUM(dst, src, tmp);
}

ASM Examples

Automatic Mode

# Automatic mode: the compiler/runtime is responsible for resource placement and scheduling.
%dst = pto.trowsum %src, %tmp : (!pto.tile<...>, !pto.tile<...>) -> !pto.tile<...>

Manual Mode

# Manual mode: explicitly bind resources first, then issue the instruction.
# Optional (when the instruction contains tile operands):
# pto.tassign %arg0, @tile(0x1000)
# pto.tassign %arg1, @tile(0x2000)
%dst = pto.trowsum %src, %tmp : (!pto.tile<...>, !pto.tile<...>) -> !pto.tile<...>

PTO Assembly Form

%dst = trowsum %src : !pto.tile<...> -> !pto.tile<...>
# AS Level 2 (DPS)
pto.trowsum ins(%src, %tmp : !pto.tile_buf<...>, !pto.tile_buf<...>) outs(%dst : !pto.tile_buf<...>)