Return op name shard.all_gather as a bitstring.

shard.all_gather - All-gather over a device grid.

Attributes

• grid - Single, FlatSymbolRefAttr, flat symbol reference attribute
• grid_axes - Single, Shard_GridAxesAttr, i16 dense array attribute
• gather_axis - Single, IndexAttr, index attribute

Operands

• input - Single, anonymous/composite constraint, memref of any non-token type values or ranked tensor of any non-token type values

Results

• result - Single, anonymous/composite constraint, memref of any non-token type values or ranked tensor of any non-token type values

Description

Concatenates all tensor slices from a device group defined by grid_axes along the tensor dimension gather_axis and replicates the result across all devices in the group.

Example:

shard.grid @grid0(shape = 2x2)
...
%1 = shard.all_gather %0 on @grid0 grid_axes = [1] gather_axis = 1
  : tensor<2x2xi8> -> tensor<2x4xi8>

Input:

                 +-------+-------+
device (0, 0) -> |  1  2 |  5  6 | <- device (0, 1)
                 |  3  4 |  7  8 |
                 +-------+-------+
device (1, 0) -> |  9 10 | 13 14 | <- device (1, 1)
                 | 11 12 | 15 16 |
                 +-------+-------+

Result:

gather tensor
axis 1
------------>
+-------------+
|  1  2  5  6 | <- devices (0, 0) and (0, 1)
|  3  4  7  8 |
+-------------+
|  9 10 13 14 | <- devices (1, 0) and (1, 1)
| 11 12 15 16 |
+-------------+

Return op name shard.all_reduce as a bitstring.

shard.all_reduce - All-reduce over a device grid.

Attributes

• grid - Single, FlatSymbolRefAttr, flat symbol reference attribute
• grid_axes - Single, Shard_GridAxesAttr, i16 dense array attribute
• reduction - Single, Shard_ReductionKindAttr, Reduction of an iterator/grid dimension.

Operands

• input - Single, anonymous/composite constraint, memref of any non-token type values or ranked tensor of any non-token type values

Results

• result - Single, anonymous/composite constraint, memref of any non-token type values or ranked tensor of any non-token type values

Description

Reduces the input tensor across all devices within the groups defined by grid_axes, using the specified reduction method. The operation performs an element-wise reduction over the tensor slices from all devices in each group. Each device in a group receives a replicated copy of the reduction result. The accumulation element type is determined by the result type and does not need to match the input element type. Before performing the reduction, each input element is converted to the result element type.

Attributes: reduction: Indicates the reduction method.

Example:

%1 = shard.all_reduce %0 on @grid0 grid_axes = [1, 0] reduction = <max>
  : tensor<3x4xf32> -> tensor<3x4xf64>

Return op name shard.all_slice as a bitstring.

shard.all_slice - All-slice over a device grid.

Attributes

• grid - Single, FlatSymbolRefAttr, flat symbol reference attribute
• grid_axes - Single, Shard_GridAxesAttr, i16 dense array attribute
• slice_axis - Single, IndexAttr, index attribute

Operands

• input - Single, AnyNon0RankedTensor, non-0-ranked.tensor of any non-token type values

Results

• result - Single, AnyNon0RankedTensor, non-0-ranked.tensor of any non-token type values

Description

Within each device group defined by grid_axes, slices the input tensor along the slice_axis dimension. It can be viewed as the inverse of an all-gather if the input data is replicated along the slice_axis. Each process simply crops its local data to the slice corresponding to its in-group device index. Notice: AllSliceOp does not involve any communication between devices and

    devices within a group may not have replicated input data.

Example:

shard.grid @grid0(shape = 2x2)
...
%1 = shard.all_slice %0 on @grid0 grid_axes = [1] slice_axis = 1
  : tensor<2x4xi8> -> tensor<2x2xi8>

Input:

+-------------+
|  1  2  5  6 | <- devices (0, 0) and (0, 1)
|  3  4  7  8 |
+-------------+
|  9 10 13 14 | <- devices (1, 0) and (1, 1)
| 11 12 15 16 |
+-------------+

Result:

slice tensor
axis 1
------------>
                 +-------+-------+
device (0, 0) -> |  1  2 |  5  6 | <- device (0, 1)
                 |  3  4 |  7  8 |
                 +-------+-------+
device (1, 0) -> |  9 10 | 13 14 | <- device (1, 1)
                 | 11 12 | 15 16 |
                 +-------+-------+

Return op name shard.all_to_all as a bitstring.

shard.all_to_all - All-to-all over a device grid.

Attributes

• grid - Single, FlatSymbolRefAttr, flat symbol reference attribute
• grid_axes - Single, Shard_GridAxesAttr, i16 dense array attribute
• split_axis - Single, IndexAttr, index attribute
• concat_axis - Single, IndexAttr, index attribute

Operands

• input - Single, AnyNon0RankedTensor, non-0-ranked.tensor of any non-token type values

Results

• result - Single, AnyNon0RankedTensor, non-0-ranked.tensor of any non-token type values

Description

Each participant logically splits its input along split_axis, then scatters the resulting pieces across the group defined by grid_axes. After receiving data pieces from other participants' scatters, it concatenates them along concat_axis to produce the final result.

Example:

shard.grid @grid0(shape = 3)
...
%1 = shard.all_to_all %0 on @grid0 grid_axes = [0]
  split_axis = 0 concat_axis = 0
  : tensor<3x2xi8> -> tensor<3x2xi8>

Input:

 device  device  device
 (0)     (1)     (2)
+-------+-------+-------+  | split and concat along
| 11 12 | 21 22 | 31 32 |  | tensor axis 0
| 13 14 | 23 24 | 33 34 |  ↓
| 15 16 | 25 26 | 35 36 |
+-------+-------+-------+

Result:

 device  device  device
 (0)     (1)     (2)
+-------+-------+-------+
| 11 12 | 13 14 | 15 16 |
| 21 22 | 23 24 | 25 26 |
| 31 32 | 33 34 | 35 36 |
+-------+-------+-------+

Return op name shard.broadcast as a bitstring.

shard.broadcast - Broadcast over a device grid.

Attributes

• grid - Single, FlatSymbolRefAttr, flat symbol reference attribute
• grid_axes - Single, Shard_GridAxesAttr, i16 dense array attribute
• root - Single, DenseI64ArrayAttr, i64 dense array attribute

Operands

• input - Single, AnyRankedTensor, ranked tensor of any non-token type values
• root_dynamic - Variadic, Index, variadic of index

Results

• result - Single, AnyRankedTensor, ranked tensor of any non-token type values

Description

Copies the input tensor on root to all devices in each group defined by grid_axes. The root device is defined by its in-group multi-index. The contents of input tensors on non-root devices are ignored.

Example:

shard.grid @grid0(shape = 2x2)

%1 = shard.broadcast %0 on @grid0
  grid_axes = [0]
  root = [0]
  : (tensor<2xi8>) -> tensor<2xi8>

Input:

                 +-------+-------+                   | broadcast
device (0, 0) -> |  1  2 |  3  4 | <- device (0, 1)  | along axis 0
                 +-------+-------+                   ↓
device (1, 0) -> |  *  * |  *  * | <- device (1, 1)
                 +-------+-------+

Output:

                 +-------+-------+
device (0, 0) -> |  1  2 |  3  4 | <- device (0, 1)
                 +-------+-------+
device (1, 0) -> |  1  2 |  3  4 | <- device (1, 1)
                 +-------+-------+

Return op name shard.gather as a bitstring.

shard.gather - Gather over a device grid.

Attributes

• grid - Single, FlatSymbolRefAttr, flat symbol reference attribute
• grid_axes - Single, Shard_GridAxesAttr, i16 dense array attribute
• gather_axis - Single, IndexAttr, index attribute
• root - Single, DenseI64ArrayAttr, i64 dense array attribute

Operands

• input - Single, AnyNon0RankedTensor, non-0-ranked.tensor of any non-token type values
• root_dynamic - Variadic, Index, variadic of index

Results

• result - Single, AnyNon0RankedTensor, non-0-ranked.tensor of any non-token type values

Description

Concatenates all tensor slices from a device group defined by grid_axes along the tensor dimension gather_axis and returns the resulting tensor on each root device. The result on all other (non-root) devices is undefined. The root device is defined by its in-group multi-index.

Example:

shard.grid @grid0(shape = 2x2)
...
%1 = shard.gather %0 on @grid0 grid_axes = [1]
  gather_axis = 1 root = [1]
  : (tensor<2x2xi8>) -> tensor<2x4xi8>

Input:

                  gather tensor
                  axis 1
                  ------------>
                 +-------+-------+
device (0, 0) -> |  1  2 |  5  6 | <- device (0, 1)
                 |  3  4 |  7  8 |
                 +-------+-------+
device (1, 0) -> |  9 10 | 13 14 | <- device (1, 1)
                 | 11 12 | 15 16 |
                 +-------+-------+

Result:

+-------------+
|  1  2  5  6 | <- devices (0, 1)
|  3  4  7  8 |
+-------------+
|  9 10 13 14 | <- devices (1, 1)
| 11 12 15 16 |
+-------------+

Devices (0, 0) and (1, 0) have undefined result.

Return op name shard.get_sharding as a bitstring.

shard.get_sharding - Get the sharding of the given tensor.

This op has support for result type inference.

Operands

• source - Single, AnyRankedTensor, ranked tensor of any non-token type values

Results

• result - Single, Shard_Sharding, sharding definition

Description

This operation returns the sharding of the given tensor as a Sharding.

Return op name shard.grid as a bitstring.

shard.grid - Description of a device/process grid.

Attributes

• sym_name - Single, SymbolNameAttr, string attribute
• shape - Single, DenseI64ArrayAttr, i64 dense array attribute

Description

The shard.grid operation is a symbol operation that identifies a specific grid. The operation has three attributes:

1. sym_name: This attribute uniquely identifies the name of the grid. This name serves as a symbolic reference to the grid throughout the MLIR module, allowing for consistent referencing and easier debugging.

2. shape: This attribute represents the shape of the device grid. It uses the same notation as a tensor shape. Also allowing for dynamic dimensions. This flexibility allows for dynamic device assignment or configurations where the exact number of devices might not be determined during compile time. For example 2x?x4.

Example:

// A device grid with 3 axes, the total device number is 4 * 8 * 12
// The dimension sizes are 4, 8, 12 
shard.grid @grid0(shape = 4x8x12)

// A device grid with 2 axes, the total device number is unknown
// The first dimension size is 4 and the second is unknown
shard.grid @grid1(shape = 4x?)

// A device grid with 2 axes, the total device number is unknown
// The first dimension size is unknown and the second is 4
shard.grid @grid2(shape = ?x4)

// A device grid with 2 axes, the number of devices along both axes
// is unknown
shard.grid @grid3(shape = ?x?)

Return op name shard.grid_shape as a bitstring.

shard.grid_shape - Get the shape of the grid.

Attributes

• grid - Single, FlatSymbolRefAttr, flat symbol reference attribute
• axes - Single, Shard_GridAxesAttr, i16 dense array attribute

Results

• result - Variadic, Index, variadic of index

Return op name shard.neighbors_linear_indices as a bitstring.

shard.neighbors_linear_indices - For given grid index get the linear indices of the direct neighbor processes along the given split.

This op has support for result type inference.

Attributes

• grid - Single, FlatSymbolRefAttr, flat symbol reference attribute
• split_axes - Single, Shard_GridAxesAttr, i16 dense array attribute

Operands

• device - Variadic, Index, variadic of index

Results

• neighbor_down - Single, Index, index
• neighbor_up - Single, Index, index

Description

Example:

shard.grid @grid0(shape = 10x20x30)
%c1 = arith.constant 1 : index
%c2 = arith.constant 2 : index
%c3 = arith.constant 3 : index
%idx = shard.neighbors_linear_indices on @grid[%c1, %c2, %c3] split_axes = [1] : index

The above returns two indices, 633 and 693, which correspond to the index of the previous process (1, 1, 3), and the next process (1, 3, 3) along the split axis 1.

A negative value is returned if there is no neighbor in the respective direction along the given split_axes.

Return op name shard.process_linear_index as a bitstring.

shard.process_linear_index - Get the linear index of the current device.

This op has support for result type inference.

Attributes

• grid - Single, FlatSymbolRefAttr, flat symbol reference attribute

Results

• result - Single, Index, index

Description

Example:

%idx = shard.process_linear_index on @grid : index

if @grid has shape (10, 20, 30), a device with multi index (1, 2, 3) will have linear index 3 + 30*2 + 20*30*1.

Return op name shard.process_multi_index as a bitstring.

shard.process_multi_index - Get the multi index of current device along specified grid axes.

Attributes

• grid - Single, FlatSymbolRefAttr, flat symbol reference attribute
• axes - Single, Shard_GridAxesAttr, i16 dense array attribute

Results

• result - Variadic, Index, variadic of index

Description

It is used in the SPMD format of IR. The axes mush be non-negative and less than the total number of grid axes. If the axes are empty then get the index along all axes.

Return op name shard.recv as a bitstring.

shard.recv - Send over a device grid.

Attributes

• grid - Single, FlatSymbolRefAttr, flat symbol reference attribute
• grid_axes - Single, Shard_GridAxesAttr, i16 dense array attribute
• source - Optional, DenseI64ArrayAttr, i64 dense array attribute

Operands

• input - Single, AnyNon0RankedTensor, non-0-ranked.tensor of any non-token type values
• source_dynamic - Variadic, Index, variadic of index

Results

• result - Single, AnyRankedTensor, ranked tensor of any non-token type values

Description

Receive tensor from device source, which is defined by its in-group multi-index. The groups are defined by grid_axes. The content of input tensor is ignored.

Return op name shard.reduce as a bitstring.

shard.reduce - Reduce over a device grid.

Attributes

• grid - Single, FlatSymbolRefAttr, flat symbol reference attribute
• grid_axes - Single, Shard_GridAxesAttr, i16 dense array attribute
• reduction - Single, Shard_ReductionKindAttr, Reduction of an iterator/grid dimension.
• root - Single, DenseI64ArrayAttr, i64 dense array attribute

Operands

• input - Single, AnyRankedTensor, ranked tensor of any non-token type values
• root_dynamic - Variadic, Index, variadic of index

Results

• result - Single, AnyRankedTensor, ranked tensor of any non-token type values

Description

Reduces the input tensor across all devices within the groups defined by grid_axes, using the specified reduction method. The operation performs an element-wise reduction over the tensor slices from all devices in each group. The reduction result will be returned on the root device of each group. It is undefined on all other (non-root) devices. The root device is defined by its in-group multi-index. The accumulation element type is determined by the result type and does not need to match the input element type. Before performing the reduction, each input element is converted to the result element type.

Attributes: reduction: Indicates the reduction method.

Example:

%1 = shard.reduce %0 on @grid0 grid_axes = [1, 0]
  reduction = <max> root = [2, 3]
  : (tensor<3x4xf32>) -> tensor<3x4xf64>

Return op name shard.reduce_scatter as a bitstring.

shard.reduce_scatter - Reduce-scatter over a device grid.

Attributes

• grid - Single, FlatSymbolRefAttr, flat symbol reference attribute
• grid_axes - Single, Shard_GridAxesAttr, i16 dense array attribute
• reduction - Single, Shard_ReductionKindAttr, Reduction of an iterator/grid dimension.
• scatter_dim - Single, IndexAttr, index attribute

Operands

• input - Single, anonymous/composite constraint, memref of any non-token type values or ranked tensor of any non-token type values

Results

• result - Single, anonymous/composite constraint, memref of any non-token type values or ranked tensor of any non-token type values

Description

Reduces the input tensor across all devices within the groups defined by grid_axes using the specified reduction method. The reduction is performed element-wise across the tensor pieces from all devices in the group. After reduction, the reduction result is scattered (split and distributed) across the device group along scatter_dim. Example:

shard.grid @grid0(shape = 2x2)
...
%1 = shard.reduce_scatter %0 on @grid0 grid_axes = [1]
  reduction = <max> scatter_dim = 0
  : tensor<2x2xf32> -> tensor<1x2xf64>

Input:

                          device
                          (0, 1)
                             ↓
                 +-------+-------+  | scatter tensor
device (0, 0) -> |  1  2 |  5  6 |  | axis 0
                 |  3  4 |  7  8 |  ↓
                 +-------+-------+
device (1, 0) -> |  9 10 | 13 14 |
                 | 11 12 | 15 16 |
                 +-------+-------+
                            ↑
                          device
                          (1, 1)

Result:

+-------+
|  5  6 | <- devices (0, 0)
+-------+
|  7  8 | <- devices (0, 1)
+-------+
| 13 14 | <- devices (1, 0)
+-------+
| 15 16 | <- devices (1, 1)
+-------+

Return op name shard.scatter as a bitstring.

shard.scatter - Scatter over a device grid.

Attributes

• grid - Single, FlatSymbolRefAttr, flat symbol reference attribute
• grid_axes - Single, Shard_GridAxesAttr, i16 dense array attribute
• scatter_dim - Single, IndexAttr, index attribute
• root - Single, DenseI64ArrayAttr, i64 dense array attribute

Operands

• input - Single, AnyNon0RankedTensor, non-0-ranked.tensor of any non-token type values
• root_dynamic - Variadic, Index, variadic of index

Results

• result - Single, AnyRankedTensor, ranked tensor of any non-token type values

Description

For each device group defined by grid_axes, the input tensor on the root device is split along axis scatter_dim and distributed across the group. The content of the input on all other (non-root) devices is ignored. The root device is defined by its in-group multi-index.

Example:

shard.grid @grid0(shape = 2x2)
%1 = shard.scatter %0 on @grid0 grid_axes = [0]
  scatter_dim = 0
  root = [1]
  : (tensor<2x2xi8>) -> tensor<1x2xi8>

Input:

                          device
                          (0, 1)
                             ↓
                 +-------+-------+  | scatter tensor
device (0, 0) -> |  *  * |  *  * |  | axis 0
                 |  *  * |  *  * |  ↓
                 +-------+-------+
device (1, 0) -> |  1  2 |  5  6 |
                 |  3  4 |  7  8 |
                 +-------+-------+
                            ↑
                          device
                          (1, 1)

Result:

                          device
                          (0, 1)
                             ↓
                 +-------+-------+
device (0, 0) -> |  1  2 |  5  6 |
                 +-------+-------+ 
device (1, 0) -> |  3  4 |  7  8 |
                 +-------+-------+
                            ↑
                          device
                          (1, 1)

Return op name shard.send as a bitstring.

shard.send - Send over a device grid.

Attributes

• grid - Single, FlatSymbolRefAttr, flat symbol reference attribute
• grid_axes - Single, Shard_GridAxesAttr, i16 dense array attribute
• destination - Single, DenseI64ArrayAttr, i64 dense array attribute

Operands

• input - Single, AnyNon0RankedTensor, non-0-ranked.tensor of any non-token type values
• destination_dynamic - Variadic, Index, variadic of index

Results

• result - Single, AnyRankedTensor, ranked tensor of any non-token type values

Description

Send input tensor to device destination, which is defined by its in-group multi-index. The groups are defined by grid_axes.

Return op name shard.shard as a bitstring.

shard.shard - Annotate on how a tensor is sharded across a shard.

This op has support for result type inference.

Attributes

• annotate_for_users - Optional, UnitAttr, unit attribute

Operands

• src - Single, AnyRankedTensor, ranked tensor of any non-token type values
• sharding - Single, Shard_Sharding, sharding definition

Results

• result - Single, AnyRankedTensor, ranked tensor of any non-token type values

Description

The shard.shard operation is designed to specify and guide the sharding behavior of a tensor value across a grid topology. This operation has two operands and two optional attributes:

1. input: This operand represents the tensor value that needs to be annotated for sharding.

2. sharding: This attribute is type of ShardingType, which is the core data structure to represent distribution of a tensor on a shard. it is typically defined by an shard.sharding operation.

3. annotate_for_users: A unit attribute addressing the scenario when a tensor's sharding annotation differs based on its context of use (either as a result or an operand). If specified, the sharding pertains to specific users of the tensor value, indicating how it should be considered when used as an operand in subsequent operations. If not, the sharding applies to the operation that defines the tensor value.

Example:

func.func @only_result_annotated(%arg0 : tensor<4x8xf32>) -> () {
    %sharding = shard.sharding @grid0 split_axes = [[0]] : !shard.sharding
    %0 = shard.shard %arg0 to %sharding : tensor<4x8xf32>
    ...
  }

  func.func @only_operand_annotated(%arg0 : tensor<4x8xf32>) -> () {
    %sharding = shard.sharding @grid0 split_axes = [[0]] : !shard.sharding
    %0 = shard.shard %arg0 to %sharding annotate_for_users : tensor<4x8xf32>
    ...
  }

  func.func @two_operands_annotated(%arg0 : tensor<4x8xf32>, %arg1 : tensor<16x8xf32>) -> () {
    %sharding = shard.sharding @grid0 split_axes = [[0]] : !shard.sharding
    %0 = shard.shard %arg0 to %sharding annotate_for_users : tensor<4x8xf32>
    %1 = shard.shard %arg1 to %sharding annotate_for_users : tensor<16x8xf32>
    ...
  }

  // The first shard.shard op applies to %arg0, the second shard.shard op
  // applies for the operand of op0, the third shard.shard op applies for the
  // operand of op2
  func.func @both_result_and_multi_operands_annotated(
      %arg0 : tensor<4x8xf32>) -> () {
    %sharding = shard.sharding @grid0 split_axes = [[0]] : !shard.sharding
    %0 = shard.shard %arg0 to %sharding : tensor<4x8xf32>
    %sharding1 = shard.sharding @grid0 split_axes = [[1]] : !shard.sharding
    %1 = shard.shard %0 to %sharding1 annotate_for_users : tensor<4x8xf32>
    %sharding2 = shard.sharding @grid0 split_axes = [[2]] : !shard.sharding
    %2 = shard.shard %0 to %sharding2 annotate_for_users : tensor<4x8xf32>
    "op0"(%1) : ...
    "op1"(%2) : ...
    ...
  }

The following usages are undefined:

  func.func @annotate_on_same_result_with_different_sharding(
      %arg0 : tensor<4x8xf32>) -> () {
    %sharding1 = shard.sharding @grid0 split_axes = [[0]] : !shard.sharding
    %sharding2 = shard.sharding @grid0 split_axes = [[1]] : !shard.sharding
    %0 = shard.shard %arg0 to $sharding1 : tensor<4x8xf32>
    %1 = shard.shard %0 to sharding2 : tensor<4x8xf32>
    ...
  }

  func.func @annotate_on_same_result_same_value_with_different_sharding(
      %arg0 : tensor<4x8xf32>) -> () {
    %sharding1 = shard.sharding @grid0 split_axes = [[0]] : !shard.sharding
    %sharding2 = shard.sharding @grid0 split_axes = [[1]] : !shard.sharding
    %0 = shard.shard %arg0 to %sharding1 : tensor<4x8xf32>
    %1 = shard.shard %arg0 to %sharding2 : tensor<4x8xf32>
    ...
  }

  func.func @annotate_on_same_operand_with_different_sharding(
      %arg0 : tensor<4x8xf32>) -> () {
    %sharding1 = shard.sharding @grid0 split_axes = [[0]] : !shard.sharding
    %sharding2 = shard.sharding @grid0 split_axes = [[1]] : !shard.sharding
    %0 = shard.shard %arg0 to %sharding1 annotate_for_users : tensor<4x8xf32>
    %1 = shard.shard %0 to %sharding2 annotate_for_users : tensor<4x8xf32>
    ...
  }

  func.func @result_annotated_after_operand(
      %arg0 : tensor<4x8xf32>) -> () {
    %sharding1 = shard.sharding @grid0 split_axes = [[0]] : !shard.sharding
    %sharding2 = shard.sharding @grid0 split_axes = [[1]] : !shard.sharding
    %0 = shard.shard %arg0 to %sharding1 annotate_for_users : tensor<4x8xf32>
    %1 = shard.shard %0 to %sharding2 : tensor<4x8xf32>
    ...
  }

Return op name shard.shard_shape as a bitstring.

shard.shard_shape - Get the shard shape for a given process/device.

Attributes

• dims - Single, DenseI64ArrayAttr, i64 dense array attribute
• device - Single, DenseI64ArrayAttr, i64 dense array attribute

Operands

• dims_dynamic - Variadic, Index, variadic of index
• sharding - Single, Shard_Sharding, sharding definition
• device_dynamic - Variadic, Index, variadic of index

Results

• result - Variadic, Index, variadic of index

Description

The device/process id is a multi-index of the device/process in the shard. This operation might be used during partition when the shard shape depends on (non-constant) values used in shard.sharding.

Return op name shard.sharding as a bitstring.

shard.sharding - Define a sharding of a tensor.

This op has support for result type inference.

Attributes

• grid - Single, FlatSymbolRefAttr, flat symbol reference attribute
• split_axes - Single, Shard_GridAxesArrayAttr,
• static_sharded_dims_offsets - Single, DenseI64ArrayAttr, i64 dense array attribute
• static_halo_sizes - Single, DenseI64ArrayAttr, i64 dense array attribute

Operands

• dynamic_sharded_dims_offsets - Variadic, I64, variadic of 64-bit signless integer
• dynamic_halo_sizes - Variadic, I64, variadic of 64-bit signless integer

Results

• result - Single, Shard_Sharding, sharding definition

Description

The Sharding specifies how a tensor is sharded and distributed across the process shard. It is typically used in a shard.shard operation. The operation has the following attributes and operands:

1. grid: this attribute is a FlatSymbolRefAttr that refers to the device grid where the distributed tensor is placed. The symbol must resolve to a shard.grid operation.

2. split_axes: is an array composed of int64_t sub-arrays. The outer array's maximum size is the rank of the related tensor. For the i-th sub-array, if its value is [x, y], it indicates that the tensor's i-th dimension is splitted along the x and y axes of the device grid.

3. [Optional] Sizes of halos to be added for each sharded tensor dimension. halo_sizes is provided as a flattened 1d array of i64s, 2 values for each sharded dimension. halo_sizes = [1, 2] means that the first sharded dimension gets an additional halo of size 1 at the start of the first dimension and a halo size is 2 at its end. halo_sizes = [1, 2, 2, 3] defines halos for the first 2 sharded dimensions e.g. the first sharded dimension gets [1,2] halos and the seconds gets [2,3] halos. ? indicates dynamic halo sizes.

4. [Optional] Offsets for each shard and sharded tensor dimension. sharded_dims_offsets is provided as a flattened 1d array of i64s. For each sharded tensor dimension the offsets (starting index) of all shards in that dimension and an additional value for the end of the last shard are provided. For a 1d sharding this means that position i has the exclusive prefix sum for shard i, and since only contiguous sharding is supported, its inclusive prefix sum is at position 'i+1'.

Assuming a 3d-tensor of shape 32x32x32 with the first 2 dimensions being sharded, sharded_dims_offsets = [0, 24, 32, 0, 20, 32] means that the first device of the device-grid will get a shard of shape 24x20x32 and the second device will get a shard of shape 8x12x32. ? indicates dynamic shard dimensions.

halo_sizes and sharded_dims_offsets are mutually exclusive.

Examples:

shard.grid @grid0(shape = 2x2x4)
shard.grid @grid1d_4(shape = 4)

// The tensor is fully replicated on @grid0.
// Currently, there must be at least one sub-array present in axes, even
// if it's empty. Otherwise, a parsing error will occur.
%sharding0 = shard.sharding @grid0 split_axes = [[]]

// The tensor is sharded on the first dimension along axis 0 of @grid0
%sharding1 = shard.sharding @grid0 split_axes = [[0]]

// Could be used for a shard.shard op
%sharded0 = shard.shard %arg0 to %sharding3 : tensor<4x8xf32>

// The tensor is sharded on its first dimension along axis 0 of @grid0 and
// and it has halo-sizes of 1 and 2 on the sharded dim.
%halo_sharding = shard.sharding @grid0 split_axes = [[0]] halo_sizes = [1, 2]
%sharded1 = shard.shard %arg0 to %halo_sharding : tensor<4x8xf32>

// The tensor is sharded on its second dimension along axis 0 of @grid1d_4
// and it has pre-defined shard sizes. The shards of the devices will have
// the following shapes: [4x2, 4x3, 4x4, 4x5]
%sharding4 = shard.sharding @grid1d_4 split_axes = [[], [0]] sharded_dims_offsets = [0, 2, 5, 9, 14]
%sharded2 = shard.shard %arg0 to %sharding4 : tensor<4x14xf32>

Return op name shard.shift as a bitstring.

shard.shift - Shift over a device grid.

Attributes

• grid - Single, FlatSymbolRefAttr, flat symbol reference attribute
• grid_axes - Single, Shard_GridAxesAttr, i16 dense array attribute
• shift_axis - Single, IndexAttr, index attribute
• offset - Single, I64Attr, 64-bit signless integer attribute
• rotate - Optional, UnitAttr, unit attribute

Operands

• input - Single, AnyNon0RankedTensor, non-0-ranked.tensor of any non-token type values

Results

• result - Single, AnyRankedTensor, ranked tensor of any non-token type values

Description

Within each device group defined by grid_axes, shifts input tensors along the device grid's axis shift_axis by the specified offset. The shift_axis must be one of the grid_axes. If the rotate attribute is set, the shift is circular. That is, the offset wraps around according to the group size along shift_axis. Otherwise, the results on devices without a corresponding source are undefined.

Example:

shard.grid @grid0(shape = 2x4)
%1 = shard.shift on @grid0 grid_axes = [1]
  shift_axis = 1 offset = 2 rotate
  : tensor<2xi8> -> tensor<2xi8>

Input:

grid axis 1
----------->

+----+----+----+----+
|  1 |  2 |  3 |  4 |
+----+----+----+----+
|  5 |  6 |  7 |  8 |
+----+----+----+----+

Result:

+----+----+----+----+
|  3 |  4 |  1 |  2 |
+----+----+----+----+
|  7 |  8 |  5 |  6 |
+----+----+----+----+

Return op name shard.update_halo as a bitstring.

shard.update_halo - Update halo data.

Attributes

• grid - Single, FlatSymbolRefAttr, flat symbol reference attribute
• split_axes - Single, Shard_GridAxesArrayAttr,
• static_halo_sizes - Single, DenseI64ArrayAttr, i64 dense array attribute

Operands

• destination - Single, anonymous/composite constraint, non-0-ranked.memref of any non-token type values or non-0-ranked.tensor of any non-token type values
• halo_sizes - Variadic, I64, variadic of 64-bit signless integer

Results

• result - Single, anonymous/composite constraint, non-0-ranked.memref of any non-token type values or non-0-ranked.tensor of any non-token type values

Description

This operation updates halo regions of shards, e.g. if their sharding specified halos and the actual tensor/memref data might have changed on the remote devices. Changes might be caused by mutating operations and/or if the new halo regions are larger than the existing ones.

Destination is supposed to be initialized with the local data (not halos).

Assumes all devices hold tensors with same-sized halo data as specified by source_halo_sizes/static_source_halo_sizes and destination_halo_sizes/static_destination_halo_sizes in source shard and destination/result shard.

split_axes specifies for each tensor axis along which grid axes its halo data is updated.