Phần sau mô tả ngữ nghĩa của các hoạt động được xác định trong giao diện XlaBuilder
. Thông thường, các hoạt động này ánh xạ 1-1 với các hoạt động được xác định trong giao diện RPC trong xla_data.proto
.
Lưu ý về danh pháp: kiểu dữ liệu tổng quát mà XLA xử lý là mảng N chiều chứa các phần tử của một số kiểu thống nhất (chẳng hạn như float 32 bit). Trong toàn bộ tài liệu, mảng được sử dụng để biểu thị một mảng có chiều tùy ý. Để thuận tiện, các trường hợp đặc biệt có tên cụ thể và quen thuộc hơn; ví dụ một vectơ là mảng 1 chiều và ma trận là mảng 2 chiều.
AfterAll
Xem thêm XlaBuilder::AfterAll
.
AfterAll lấy một số lượng khác nhau của mã thông báo và tạo ra một mã thông báo duy nhất. Các mã thông báo là các loại nguyên thủy có thể được xâu chuỗi giữa các hoạt động tác động bên để thực thi thứ tự. AfterAll
có thể được sử dụng như một kết hợp của các mã thông báo để sắp xếp một hoạt động sau một hoạt động đã định.
AfterAll(operands)
Tranh luận | Gõ phím | Ngữ nghĩa học |
---|---|---|
operands | XlaOp | số lượng mã thông báo khác nhau |
AllGather
Xem thêm XlaBuilder::AllGather
.
Thực hiện nối giữa các bản sao.
AllGather(operand, all_gather_dim, shard_count, replica_group_ids, channel_id)
Tranh luận | Gõ phím | Ngữ nghĩa học |
---|---|---|
operand | XlaOp | Mảng để nối giữa các bản sao. |
all_gather_dim | int64 | Chiều liên kết. |
replica_groups | vectơ của vectơ int64 | Các nhóm mà quá trình nối được thực hiện. |
channel_id | int64 tùy chọn | ID kênh tùy chọn cho giao tiếp giữa các mô-đun. |
-
replica_groups
là danh sách các nhóm bản sao mà quá trình nối được thực hiện (id bản sao cho bản sao hiện tại có thể được truy xuất bằng cách sử dụngReplicaId
). Thứ tự của các bản sao trong mỗi nhóm xác định thứ tự mà các đầu vào của chúng nằm trong kết quả.replica_groups
phải trống (trong trường hợp đó tất cả các bản sao đều thuộc một nhóm duy nhất, được sắp xếp từ0
đếnN - 1
) hoặc chứa cùng số phần tử với số lượng bản sao. Ví dụ:replica_groups = {0, 2}, {1, 3}
thực hiện phép ghép giữa các bản sao0
và2
, và1
và3
. -
shard_count
là kích thước của mỗi nhóm bản sao. Chúng ta cần điều này trong trường hợpreplica_groups
trống. -
channel_id
được sử dụng cho giao tiếp giữa các mô-đun: chỉall-gather
với cùng mộtchannel_id
mới có thể giao tiếp với nhau.
Hình dạng đầu ra là hình dạng đầu vào với all_gather_dim
được tạo thành shard_count
lớn hơn lần. Ví dụ: nếu có hai bản sao và toán hạng có giá trị [1.0, 2.5]
và [3.0, 5.25]
tương ứng trên hai bản sao, thì giá trị đầu ra từ op này trong đó all_gather_dim
là 0
sẽ là [1.0, 2.5, 3.0, 5.25]
trên cả hai bản sao.
Giảm tất cả
Xem thêm XlaBuilder::AllReduce
.
Thực hiện tính toán tùy chỉnh trên các bản sao.
AllReduce(operand, computation, replica_group_ids, channel_id)
Tranh luận | Gõ phím | Ngữ nghĩa học |
---|---|---|
operand | XlaOp | Mảng hoặc một bộ mảng không trống để giảm trên các bản sao. |
computation | XlaComputation | Giảm tính toán |
replica_groups | vectơ của vectơ int64 | Các nhóm mà việc cắt giảm được thực hiện |
channel_id | int64 tùy chọn | ID kênh tùy chọn cho giao tiếp giữa các mô-đun |
- Khi
operand
là một bộ nhiều mảng, việc giảm tất cả được thực hiện trên mỗi phần tử của bộ giá trị. -
replica_groups
là danh sách các nhóm bản sao mà quá trình giảm được thực hiện (id bản sao cho bản sao hiện tại có thể được truy xuất bằng cách sử dụngReplicaId
).replica_groups
phải trống (trong trường hợp này tất cả các bản sao đều thuộc về một nhóm duy nhất) hoặc chứa cùng số phần tử với số lượng bản sao. Ví dụ:replica_groups = {0, 2}, {1, 3}
thực hiện giảm giữa các bản sao0
và2
và1
và3
. -
channel_id
được sử dụng cho giao tiếp giữa các mô-đun: chỉall-reduce
với cùng mộtchannel_id
mới có thể giao tiếp với nhau.
Hình dạng đầu ra giống với hình dạng đầu vào. Ví dụ: nếu có hai bản sao và toán hạng có giá trị [1.0, 2.5]
và [3.0, 5.25]
tương ứng trên hai bản sao, thì giá trị đầu ra từ phép tính op và tổng này sẽ là [4.0, 7.75]
trên cả hai bản sao. Nếu đầu vào là một bộ, thì đầu ra cũng là một bộ.
Việc tính toán kết quả của AllReduce
yêu cầu phải có một đầu vào từ mỗi bản sao, vì vậy nếu một bản sao thực thi một nút AllReduce
nhiều lần hơn một nút khác, thì bản sao cũ sẽ đợi mãi mãi. Vì các bản sao đều đang chạy cùng một chương trình, không có nhiều cách để điều đó xảy ra, nhưng có thể xảy ra khi điều kiện của vòng lặp while phụ thuộc vào dữ liệu từ nguồn nạp và dữ liệu được đưa vào khiến vòng lặp while lặp lại nhiều lần hơn trên một bản sao hơn một bản sao khác.
Tất cả
Xem thêm XlaBuilder::AllToAll
.
AllToAll là một hoạt động tập thể gửi dữ liệu từ tất cả các lõi đến tất cả các lõi. Nó có hai giai đoạn:
- Giai đoạn phân tán. Trên mỗi lõi, toán hạng được chia thành số khối
split_count
dọc theosplit_dimensions
và các khối được phân tán cho tất cả các lõi, ví dụ: khối thứ i được gửi đến lõi thứ i. - Giai đoạn tập hợp. Mỗi lõi nối các khối đã nhận được dọc theo
concat_dimension
.
Các lõi tham gia có thể được định cấu hình bằng:
-
replica_groups
: mỗi ReplicaGroup chứa danh sách id bản sao tham gia tính toán (id bản sao cho bản sao hiện tại có thể được truy xuất bằng cách sử dụngReplicaId
). AllToAll sẽ được áp dụng trong các nhóm con theo thứ tự được chỉ định. Ví dụ:replica_groups = { {1,2,3}, {4,5,0} }
có nghĩa là AllToAll sẽ được áp dụng trong các bản sao{1, 2, 3}
và trong giai đoạn tập hợp và các khối đã nhận sẽ được nối theo cùng một thứ tự 1, 2, 3. Sau đó, một AllToAll khác sẽ được áp dụng trong các bản sao 4, 5, 0 và thứ tự nối cũng là 4, 5, 0. Nếureplica_groups
trống, tất cả các bản sao đều thuộc về một nhóm, theo thứ tự nối của sự xuất hiện của chúng.
Điều kiện tiên quyết:
- Kích thước thứ nguyên của toán hạng trên
split_dimension
chia hết chosplit_count
. - Hình dạng của toán hạng không phải là bộ ba.
AllToAll(operand, split_dimension, concat_dimension, split_count, replica_groups)
Tranh luận | Gõ phím | Ngữ nghĩa học |
---|---|---|
operand | XlaOp | mảng đầu vào n chiều |
split_dimension | int64 | Một giá trị trong khoảng [0, n) đặt tên cho thứ nguyên mà toán hạng được phân chia |
concat_dimension | int64 | một giá trị trong khoảng [0, n) đặt tên cho thứ nguyên mà các khối tách được nối với nhau |
split_count | int64 | số lượng lõi tham gia hoạt động này. Nếu replica_groups trống, đây sẽ là số lượng các bản sao; nếu không, con số này phải bằng số lượng bản sao trong mỗi nhóm. |
replica_groups | ReplicaGroup | mỗi nhóm chứa một danh sách id bản sao. |
Dưới đây là một ví dụ về Alltoall.
XlaBuilder b("alltoall");
auto x = Parameter(&b, 0, ShapeUtil::MakeShape(F32, {4, 16}), "x");
AllToAll(x, /*split_dimension=*/1, /*concat_dimension=*/0, /*split_count=*/4);

Trong ví dụ này, có 4 lõi tham gia Alltoall. Trên mỗi lõi, toán hạng được chia thành 4 phần dọc theo chiều 0, vì vậy mỗi phần có hình dạng f32 [4,4]. 4 phần nằm rải rác đến tất cả các lõi. Sau đó, mỗi lõi nối các phần đã nhận dọc theo chiều 1, theo thứ tự hoặc lõi 0-4. Vì vậy đầu ra trên mỗi lõi có hình dạng f32 [16,4].
BatchNormGrad
Xem thêm XlaBuilder::BatchNormGrad
và bài báo chuẩn hóa hàng loạt ban đầu để biết mô tả chi tiết về thuật toán.
Tính toán độ dốc của định mức hàng loạt.
BatchNormGrad(operand, scale, mean, variance, grad_output, epsilon, feature_index)
Tranh luận | Gõ phím | Ngữ nghĩa học |
---|---|---|
operand | XlaOp | mảng n chiều được chuẩn hóa (x) |
scale | XlaOp | Mảng 1 chiều (\(\gamma\)) |
mean | XlaOp | Mảng 1 chiều (\(\mu\)) |
variance | XlaOp | Mảng 1 chiều (\(\sigma^2\)) |
grad_output | XlaOp | Gradients được chuyển đến BatchNormTraining (\( \nabla y\)) |
epsilon | float | Giá trị Epsilon (\(\epsilon\)) |
feature_index | int64 | Lập chỉ mục cho thứ nguyên đặc trưng trong operand |
Đối với mỗi đối tượng trong thứ nguyên đối tượng ( feature_index
là chỉ số cho thứ nguyên đối tượng trong operand
), thao tác tính toán độ dốc liên quan đến operand
, offset
và scale
trên tất cả các thứ nguyên khác. feature_index
phải là một chỉ mục hợp lệ cho thứ nguyên tính năng trong operand
.
Ba gradient được xác định bằng các công thức sau (giả sử mảng 4 chiều là operand
và với chỉ số kích thước đặc trưng l
, kích thước lô m
và kích thước không gian w
và h
):
\[ \begin{split} c_l&= \frac{1}{mwh}\sum_{i=1}^m\sum_{j=1}^w\sum_{k=1}^h \left( \nabla y_{ijkl} \frac{x_{ijkl} - \mu_l}{\sigma^2_l+\epsilon} \right) \\\\ \nabla x_{ijkl} &= \frac{\gamma_{l} }{\sqrt{\sigma^2_{l}+\epsilon} } \left( \nabla y_{ijkl} - \mathrm{mean}(\nabla y) - c_l (x_{ijkl} - \mu_{l}) \right) \\\\ \nabla \gamma_l &= \sum_{i=1}^m\sum_{j=1}^w\sum_{k=1}^h \left( \nabla y_{ijkl} \frac{x_{ijkl} - \mu_l}{\sqrt{\sigma^2_{l}+\epsilon} } \right) \\\\\ \nabla \beta_l &= \sum_{i=1}^m\sum_{j=1}^w\sum_{k=1}^h \nabla y_{ijkl} \end{split} \]
Giá trị đầu vào mean
và variance
đại diện cho giá trị mô men qua các kích thước theo lô và không gian.
Loại đầu ra là một bộ ba tay cầm:
Kết quả đầu ra | Gõ phím | Ngữ nghĩa học |
---|---|---|
grad_operand | XlaOp | gradient liên quan đến operand đầu vào (\( \nabla x\)) |
grad_scale | XlaOp | gradient đối với scale đầu vào (\( \nabla \gamma\)) |
grad_offset | XlaOp | gradient liên quan đến offset đầu vào (\( \nabla \beta\)) |
BatchNormInference
Xem thêm XlaBuilder::BatchNormInference
và bài báo chuẩn hóa hàng loạt ban đầu để biết mô tả chi tiết về thuật toán.
Chuẩn hóa một mảng trên các kích thước hàng loạt và không gian.
BatchNormInference(operand, scale, offset, mean, variance, epsilon, feature_index)
Tranh luận | Gõ phím | Ngữ nghĩa học |
---|---|---|
operand | XlaOp | mảng n chiều được chuẩn hóa |
scale | XlaOp | Mảng 1 chiều |
offset | XlaOp | Mảng 1 chiều |
mean | XlaOp | Mảng 1 chiều |
variance | XlaOp | Mảng 1 chiều |
epsilon | float | Giá trị Epsilon |
feature_index | int64 | Lập chỉ mục cho thứ nguyên đặc trưng trong operand |
Đối với mỗi đối tượng trong thứ nguyên đối tượng ( feature_index
là chỉ số cho thứ nguyên đối tượng trong operand
), thao tác tính giá trị trung bình và phương sai trên tất cả các thứ nguyên khác, đồng thời sử dụng giá trị trung bình và phương sai để chuẩn hóa từng phần tử trong operand
. feature_index
phải là một chỉ mục hợp lệ cho thứ nguyên tính năng trong operand
.
BatchNormInference
tương đương với việc gọi BatchNormTraining
mà không cần tính toán mean
và variance
cho mỗi lô. Nó sử dụng giá mean
đầu vào và variance
thay vì làm giá trị ước tính. Mục đích của op này là để giảm độ trễ trong suy luận, do đó có tên là BatchNormInference
.
Đầu ra là một mảng chuẩn hóa n chiều có cùng hình dạng với operand
đầu vào.
BatchNormTraining
Xem thêm XlaBuilder::BatchNormTraining
và the original batch normalization paper
để biết mô tả chi tiết về thuật toán.
Chuẩn hóa một mảng trên các kích thước hàng loạt và không gian.
BatchNormTraining(operand, scale, offset, epsilon, feature_index)
Tranh luận | Gõ phím | Ngữ nghĩa học |
---|---|---|
operand | XlaOp | mảng n chiều được chuẩn hóa (x) |
scale | XlaOp | Mảng 1 chiều (\(\gamma\)) |
offset | XlaOp | Mảng 1 chiều (\(\beta\)) |
epsilon | float | Giá trị Epsilon (\(\epsilon\)) |
feature_index | int64 | Lập chỉ mục cho thứ nguyên đặc trưng trong operand |
Đối với mỗi đối tượng trong thứ nguyên đối tượng ( feature_index
là chỉ số cho thứ nguyên đối tượng trong operand
), thao tác tính giá trị trung bình và phương sai trên tất cả các thứ nguyên khác, đồng thời sử dụng giá trị trung bình và phương sai để chuẩn hóa từng phần tử trong operand
. feature_index
phải là một chỉ mục hợp lệ cho thứ nguyên tính năng trong operand
.
Thuật toán diễn ra như sau cho mỗi lô trong operand
\(x\) chứa m
phần tử với w
và h
là kích thước của kích thước không gian (giả sử operand
là một mảng 4 chiều):
Tính toán trung bình hàng loạt \(\mu_l\) cho mỗi đối tượng
l
trong thứ nguyên đặc điểm:\(\mu_l=\frac{1}{mwh}\sum_{i=1}^m\sum_{j=1}^w\sum_{k=1}^h x_{ijkl}\)Tính toán phương sai lô \(\sigma^2_l\):\(\sigma^2_l=\frac{1}{mwh}\sum_{i=1}^m\sum_{j=1}^w\sum_{k=1}^h (x_{ijkl} - \mu_l)^2\)
Chuẩn hóa, chia tỷ lệ và thay đổi:\(y_{ijkl}=\frac{\gamma_l(x_{ijkl}-\mu_l)}{\sqrt[2]{\sigma^2_l+\epsilon} }+\beta_l\)
Giá trị epsilon, thường là một số nhỏ, được thêm vào để tránh lỗi chia cho không.
Loại đầu ra là một bộ ba XlaOp
s:
Kết quả đầu ra | Gõ phím | Ngữ nghĩa học |
---|---|---|
output | XlaOp | mảng n chiều có cùng hình dạng với operand đầu vào (y) |
batch_mean | XlaOp | Mảng 1 chiều (\(\mu\)) |
batch_var | XlaOp | Mảng 1 chiều (\(\sigma^2\)) |
batch_mean
và batch_var
là các khoảnh khắc được tính toán trên các kích thước không gian và lô bằng cách sử dụng các công thức ở trên.
BitcastConvertType
Xem thêm XlaBuilder::BitcastConvertType
.
Tương tự như tf.bitcast
trong TensorFlow, thực hiện hoạt động bitcast thông minh phần tử từ hình dạng dữ liệu đến hình dạng đích. Kích thước đầu vào và đầu ra phải phù hợp: ví dụ: phần tử s32
trở thành phần tử f32
thông qua quy trình bitcast và một phần tử s32
sẽ trở thành bốn phần tử s8
. Bitcast được thực hiện như một đúc cấp thấp, vì vậy các máy có các biểu diễn dấu phẩy động khác nhau sẽ cho các kết quả khác nhau.
BitcastConvertType(operand, new_element_type)
Tranh luận | Gõ phím | Ngữ nghĩa học |
---|---|---|
operand | XlaOp | mảng kiểu T với độ mờ D |
new_element_type | PrimitiveType | gõ U |
Kích thước của toán hạng và hình dạng đích phải khớp với nhau, ngoại trừ kích thước cuối cùng sẽ thay đổi theo tỷ lệ của kích thước ban đầu trước và sau khi chuyển đổi.
Các loại phần tử nguồn và đích không được là bộ giá trị.
Chuyển đổi bit sang kiểu nguyên thủy có chiều rộng khác nhau
BitcastConvert
HLO hỗ trợ trường hợp kích thước của loại phần tử đầu ra T'
không bằng kích thước của phần tử đầu vào T
Vì toàn bộ hoạt động về mặt khái niệm là một bitcast và không thay đổi các byte bên dưới, hình dạng của phần tử đầu ra phải thay đổi. Đối với B = sizeof(T), B' = sizeof(T')
, có hai trường hợp có thể xảy ra.
Đầu tiên, khi B > B'
, hình dạng đầu ra nhận được một kích thước nhỏ nhất mới là kích thước B/B'
. Ví dụ:
f16[10,2]{1,0} %output = f16[10,2]{1,0} bitcast-convert(f32[10]{0} %input)
Quy tắc vẫn giữ nguyên đối với các đại lượng vô hướng hiệu quả:
f16[2]{0} %output = f16[2]{0} bitcast-convert(f32[] %input)
Ngoài ra, đối với B' > B
, lệnh yêu cầu kích thước logic cuối cùng của hình dạng đầu vào phải bằng B'/B
và thứ nguyên này bị loại bỏ trong quá trình chuyển đổi:
f32[10]{0} %output = f32[10]{0} bitcast-convert(f16[10,2]{1,0} %input)
Lưu ý rằng chuyển đổi giữa các tốc độ bit khác nhau không phải là yếu tố.
Phát tin
Xem thêm XlaBuilder::Broadcast
.
Thêm thứ nguyên vào một mảng bằng cách sao chép dữ liệu trong mảng.
Broadcast(operand, broadcast_sizes)
Tranh luận | Gõ phím | Ngữ nghĩa học |
---|---|---|
operand | XlaOp | Mảng để nhân đôi |
broadcast_sizes | ArraySlice<int64> | Kích thước của các thứ nguyên mới |
Các kích thước mới được chèn vào bên trái, tức là nếu broadcast_sizes
có các giá trị {a0, ..., aN}
và hình dạng toán hạng có kích thước {b0, ..., bM}
thì hình dạng của đầu ra có kích thước {a0, ..., aN, b0, ..., bM}
.
Các thứ nguyên mới lập chỉ mục thành các bản sao của toán hạng, tức là
output[i0, ..., iN, j0, ..., jM] = operand[j0, ..., jM]
Ví dụ: nếu operand
là một đại lượng vô hướng f32
với giá trị 2.0f
và broadcast_sizes
là {2, 3}
, thì kết quả sẽ là một mảng có hình dạng f32[2, 3]
và tất cả các giá trị trong kết quả sẽ là 2.0f
.
BroadcastInDim
Xem thêm XlaBuilder::BroadcastInDim
.
Mở rộng kích thước và thứ hạng của một mảng bằng cách sao chép dữ liệu trong mảng.
BroadcastInDim(operand, out_dim_size, broadcast_dimensions)
Tranh luận | Gõ phím | Ngữ nghĩa học |
---|---|---|
operand | XlaOp | Mảng để nhân đôi |
out_dim_size | ArraySlice<int64> | Kích thước của các kích thước của hình dạng mục tiêu |
broadcast_dimensions | ArraySlice<int64> | Kích thước nào trong hình dạng đích mà mỗi kích thước của hình dạng toán hạng tương ứng với |
Tương tự như Broadcast, nhưng cho phép thêm thứ nguyên ở bất kỳ đâu và mở rộng thứ nguyên hiện có với kích thước 1.
operand
được phát tới hình dạng được mô tả bởi out_dim_size
. broadcast_dimensions
ánh xạ các kích thước của operand
với các kích thước của hình dạng đích, tức là kích thước thứ i của toán hạng được ánh xạ tới kích thước thứ của broadcast_dimension [i] của hình dạng đầu ra. Kích thước của operand
phải có kích thước 1 hoặc có cùng kích thước với kích thước trong hình dạng đầu ra mà chúng được ánh xạ tới. Các kích thước còn lại được lấp đầy bằng các kích thước có kích thước 1. Phát sóng theo chiều thoái hóa sau đó phát sóng dọc theo các kích thước suy giảm này để đạt được hình dạng đầu ra. Ngữ nghĩa được mô tả chi tiết trên trang phát sóng .
Gọi điện
Xem thêm XlaBuilder::Call
.
Gọi một phép tính với các đối số đã cho.
Call(computation, args...)
Tranh luận | Gõ phím | Ngữ nghĩa học |
---|---|---|
computation | XlaComputation | tính toán kiểu T_0, T_1, ..., T_{N-1} -> S với N tham số kiểu tùy ý |
args | dãy N XlaOp s | N đối số kiểu tùy ý |
Độ hiếm và loại của args
phải phù hợp với các tham số của computation
. Nó được phép không có args
.
Cholesky
Xem thêm XlaBuilder::Cholesky
.
Tính toán phân rã Cholesky của một lô ma trận xác định dương đối xứng (Hermitian).
Cholesky(a, lower)
Tranh luận | Gõ phím | Ngữ nghĩa học |
---|---|---|
a | XlaOp | một mảng thứ hạng> 2 thuộc kiểu dấu phẩy động hoặc phức. |
lower | bool | sử dụng tam giác trên hay tam giác dưới của a . |
Nếu giá trị lower
là true
, hãy tính các ma trận tam giác thấp hơn l
sao cho \( a = l
. l^T \). Nếu giá trị lower
là false
, hãy tính các ma trận tam giác trên u
sao cho \( a = u^T . u \).
Dữ liệu đầu vào chỉ được đọc từ tam giác dưới / trên của a
, tùy thuộc vào giá trị của lower
. Các giá trị từ tam giác khác bị bỏ qua. Dữ liệu đầu ra được trả về trong cùng một tam giác; các giá trị trong tam giác khác được xác định bởi triển khai và có thể là bất kỳ giá trị nào.
Nếu hạng của a
lớn hơn 2, a
được coi là một lô ma trận, trong đó tất cả ngoại trừ 2 thứ nguyên nhỏ là thứ nguyên lô.
Nếu a
không đối xứng (Hermitian) xác định dương, kết quả được xác định thực thi.
Kẹp
Xem thêm XlaBuilder::Clamp
.
Kẹp toán hạng trong phạm vi giữa giá trị nhỏ nhất và giá trị lớn nhất.
Clamp(min, operand, max)
Tranh luận | Gõ phím | Ngữ nghĩa học |
---|---|---|
min | XlaOp | mảng kiểu T |
operand | XlaOp | mảng kiểu T |
max | XlaOp | mảng kiểu T |
Cho một toán hạng và các giá trị nhỏ nhất và lớn nhất, trả về toán hạng nếu nó nằm trong phạm vi giữa giá trị tối thiểu và tối đa, còn lại trả về giá trị nhỏ nhất nếu toán hạng nằm dưới phạm vi này hoặc giá trị lớn nhất nếu toán hạng nằm trên phạm vi này. Tức là, clamp(a, x, b) = min(max(a, x), b)
.
Tất cả ba mảng phải có cùng hình dạng. Ngoài ra, là một dạng phát sóng bị hạn chế, min
và / hoặc max
có thể là một vô hướng của loại T
Ví dụ với min
và max
vô hướng:
let operand: s32[3] = {-1, 5, 9};
let min: s32 = 0;
let max: s32 = 6;
==>
Clamp(min, operand, max) = s32[3]{0, 5, 6};
Sụp đổ
Xem thêm XlaBuilder::Collapse
và hoạt động tf.reshape
.
Thu gọn các kích thước của một mảng thành một thứ nguyên.
Collapse(operand, dimensions)
Tranh luận | Gõ phím | Ngữ nghĩa học |
---|---|---|
operand | XlaOp | mảng kiểu T |
dimensions | vector int64 | theo thứ tự, tập con liên tiếp các kích thước của T. |
Thu gọn thay thế tập hợp con đã cho của các kích thước của toán hạng bằng một thứ nguyên. Các đối số đầu vào là một mảng tùy ý kiểu T và một vectơ biên dịch-thời gian-hằng số của các chỉ số thứ nguyên. Các chỉ số thứ nguyên phải là một tập hợp con theo thứ tự (số thứ nguyên từ thấp đến cao), liên tiếp của các thứ nguyên T. Do đó, {0, 1, 2}, {0, 1} hoặc {1, 2} đều là tập hợp thứ nguyên hợp lệ, nhưng {1, 0} hoặc {0, 2} thì không. Chúng được thay thế bằng một thứ nguyên mới, ở cùng vị trí trong chuỗi thứ nguyên với thứ mà chúng thay thế, với kích thước thứ nguyên mới bằng tích của các kích thước thứ nguyên ban đầu. Số thứ nguyên thấp nhất trong các thứ dimensions
là thứ nguyên thay đổi chậm nhất (chính yếu nhất) trong tổ vòng lặp thu gọn thứ nguyên này và số thứ nguyên cao nhất thay đổi nhanh nhất (nhỏ nhất). Xem toán tử tf.reshape
nếu cần thêm thứ tự thu gọn chung.
Ví dụ, cho v là một mảng gồm 24 phần tử:
let v = f32[4x2x3] { { {10, 11, 12}, {15, 16, 17} },
{ {20, 21, 22}, {25, 26, 27} },
{ {30, 31, 32}, {35, 36, 37} },
{ {40, 41, 42}, {45, 46, 47} } };
// Collapse to a single dimension, leaving one dimension.
let v012 = Collapse(v, {0,1,2});
then v012 == f32[24] {10, 11, 12, 15, 16, 17,
20, 21, 22, 25, 26, 27,
30, 31, 32, 35, 36, 37,
40, 41, 42, 45, 46, 47};
// Collapse the two lower dimensions, leaving two dimensions.
let v01 = Collapse(v, {0,1});
then v01 == f32[4x6] { {10, 11, 12, 15, 16, 17},
{20, 21, 22, 25, 26, 27},
{30, 31, 32, 35, 36, 37},
{40, 41, 42, 45, 46, 47} };
// Collapse the two higher dimensions, leaving two dimensions.
let v12 = Collapse(v, {1,2});
then v12 == f32[8x3] { {10, 11, 12},
{15, 16, 17},
{20, 21, 22},
{25, 26, 27},
{30, 31, 32},
{35, 36, 37},
{40, 41, 42},
{45, 46, 47} };
CollectivePermute
Xem thêm XlaBuilder::CollectivePermute
.
CollectivePermute là một hoạt động tập thể gửi và nhận các bản sao chéo dữ liệu.
CollectivePermute(operand, source_target_pairs)
Tranh luận | Gõ phím | Ngữ nghĩa học |
---|---|---|
operand | XlaOp | mảng đầu vào n chiều |
source_target_pairs | <int64, int64> vectơ | Danh sách các cặp (source_replica_id, target_replica_id). Đối với mỗi cặp, toán hạng được gửi từ bản sao nguồn đến bản sao đích. |
Lưu ý rằng có những hạn chế sau đối với source_target_pair
:
- Hai cặp bất kỳ không được có cùng id bản sao đích và chúng không được có cùng id bản sao nguồn.
- Nếu một id bản sao không phải là mục tiêu trong bất kỳ cặp nào, thì đầu ra trên bản sao đó là một tensor bao gồm 0 (các) có cùng hình dạng với đầu vào.
Kết hợp
Xem thêm XlaBuilder::ConcatInDim
.
Concatenate tạo một mảng từ nhiều toán hạng mảng. Mảng có cùng thứ hạng với mỗi toán hạng mảng đầu vào (phải có cùng thứ hạng với nhau) và chứa các đối số theo thứ tự mà chúng đã được chỉ định.
Concatenate(operands..., dimension)
Tranh luận | Gõ phím | Ngữ nghĩa học |
---|---|---|
operands | trình tự của N XlaOp | N mảng kiểu T có kích thước [L0, L1, ...]. Yêu cầu N> = 1. |
dimension | int64 | Một giá trị trong khoảng [0, N) đặt tên cho thứ nguyên được nối giữa các operands . |
Ngoại trừ thứ dimension
, tất cả các thứ nguyên phải giống nhau. Điều này là do XLA không hỗ trợ các mảng "rách rưới". Cũng lưu ý rằng các giá trị hạng-0 không thể được nối (vì không thể đặt tên cho thứ nguyên mà quá trình nối xảy ra).
Ví dụ 1 chiều:
Concat({ {2, 3}, {4, 5}, {6, 7} }, 0)
>>> {2, 3, 4, 5, 6, 7}
Ví dụ 2 chiều:
let a = {
{1, 2},
{3, 4},
{5, 6},
};
let b = {
{7, 8},
};
Concat({a, b}, 0)
>>> {
{1, 2},
{3, 4},
{5, 6},
{7, 8},
}
Biểu đồ:

Có điều kiện
Xem thêm XlaBuilder::Conditional
.
Conditional(pred, true_operand, true_computation, false_operand, false_computation)
Tranh luận | Gõ phím | Ngữ nghĩa học |
---|---|---|
pred | XlaOp | Vô hướng loại PRED |
true_operand | XlaOp | Đối số kiểu \(T_0\) |
true_computation | XlaComputation | XlaComputation của kiểu \(T_0 \to S\) |
false_operand | XlaOp | Đối số kiểu \(T_1\) |
false_computation | XlaComputation | XlaComputation của kiểu \(T_1 \to S\) |
Thực thi true_computation
nếu pred
là true
, false_computation
nếu pred
là false
và trả về kết quả.
true_computation
phải nhận một đối số duy nhất của kiểu \(T_0\) và sẽ được gọi với true_operand
phải cùng kiểu. false_computation
phải nhận một đối số duy nhất của kiểu \(T_1\) và sẽ được gọi với false_operand
phải cùng kiểu. Loại giá trị trả về của true_computation
và false_computation
phải giống nhau.
Lưu ý rằng chỉ một trong số true_computation
và false_computation
sẽ được thực thi tùy thuộc vào giá trị của pred
.
Conditional(branch_index, branch_computations, branch_operands)
Tranh luận | Gõ phím | Ngữ nghĩa học |
---|---|---|
branch_index | XlaOp | Vô hướng loại S32 |
branch_computations | trình tự của N XlaComputation | XlaComputations của loại \( T_0 \to S , T_1 \to S , ..., T_{N-1} \to S \) |
branch_operands | trình tự của N XlaOp | Đối số kiểu \( T_0 , T_1 , ..., T_{N-1} \) |
Thực thi branch_computations[branch_index]
và trả về kết quả. Nếu branch_index
là S32
<0 hoặc> = N, thì branch_computations[N-1]
được thực thi như là một nhánh mặc định.
Mỗi branch_computations[b]
phải nhận một đối số duy nhất của kiểu T_b
và sẽ được gọi với branch_operands[b]
phải cùng kiểu. Loại giá trị trả về của mỗi branch_computations[b]
phải giống nhau.
Lưu ý rằng chỉ một trong các branch_computations
sẽ được thực thi tùy thuộc vào giá trị của branch_index
.
Ch.đổi (tích chập)
Xem thêm XlaBuilder::Conv
.
Là ConvWithGeneralPadding, nhưng phần đệm được chỉ định một cách ngắn gọn là CÙNG hoặc HỢP LỆ. Phần đệm SAME đệm đầu vào ( lhs
) bằng các số 0 để đầu ra có cùng hình dạng với đầu vào khi không tính đến bước đệm. Đệm HỢP LỆ chỉ có nghĩa là không có đệm.
ConvWithGeneralPadding (tích chập)
Xem thêm XlaBuilder::ConvWithGeneralPadding
.
Tính toán một tích chập của loại được sử dụng trong mạng nơ-ron. Ở đây, một phép chập có thể được coi là một cửa sổ n chiều di chuyển trên một vùng cơ sở n chiều và một phép tính được thực hiện cho từng vị trí có thể có của cửa sổ.
Tranh luận | Gõ phím | Ngữ nghĩa học |
---|---|---|
lhs | XlaOp | xếp hạng n + 2 mảng đầu vào |
rhs | XlaOp | xếp hạng n + 2 mảng trọng số hạt nhân |
window_strides | ArraySlice<int64> | mảng thứ n của các bước nhân |
padding | ArraySlice< pair<int64, int64>> | mảng thứ n của khoảng đệm (thấp, cao) |
lhs_dilation | ArraySlice<int64> | mảng hệ số giãn nở lhs |
rhs_dilation | ArraySlice<int64> | mảng hệ số giãn nở nd rhs |
feature_group_count | int64 | số lượng nhóm tính năng |
batch_group_count | int64 | số lượng nhóm hàng loạt |
Gọi n là số kích thước không gian. Đối số lhs
là một mảng hạng n + 2 mô tả vùng cơ sở. Đây được gọi là đầu vào, mặc dù tất nhiên rhs cũng là đầu vào. Trong mạng nơ-ron, đây là những kích hoạt đầu vào. Thứ nguyên n + 2 theo thứ tự sau:
-
batch
: Mỗi tọa độ trong thứ nguyên này đại diện cho một đầu vào độc lập mà phép tích chập được thực hiện. -
z/depth/features
: Mỗi vị trí (y, x) trong vùng cơ sở có một vectơ được liên kết với nó, đi vào thứ nguyên này. -
spatial_dims
: Mô tản
kích thước không gian xác định vùng cơ sở mà cửa sổ di chuyển qua.
Đối số rhs
là một mảng thứ n + 2 mô tả bộ lọc chập / nhân / cửa sổ. Các kích thước, theo thứ tự sau:
-
output-z
: Thứ nguyênz
của đầu ra. -
input-z
: Kích thước của thứ nguyên này nhân vớifeature_group_count
phải bằng kích thước của thứ nguyênz
tính bằng lhs. -
spatial_dims
: Mô tản
kích thước không gian xác định cửa sổ thứ n di chuyển trên vùng cơ sở.
Đối số window_strides
chỉ định bước của cửa sổ tích chập trong các kích thước không gian. Ví dụ: nếu bước đi trong chiều không gian đầu tiên là 3, thì cửa sổ chỉ có thể được đặt tại tọa độ mà chỉ số không gian đầu tiên chia hết cho 3.
Đối số padding
chỉ định số lượng không đệm sẽ được áp dụng cho vùng cơ sở. Số lượng phần đệm có thể là số âm - giá trị tuyệt đối của phần đệm âm cho biết số phần tử cần loại bỏ khỏi thứ nguyên được chỉ định trước khi thực hiện phép chập. padding[0]
chỉ định padding cho thứ nguyên y
và padding[1]
chỉ định padding cho thứ nguyên x
. Mỗi cặp có phần đệm thấp là phần tử đầu tiên và phần đệm cao là phần tử thứ hai. Phần đệm thấp được áp dụng theo hướng của các chỉ số thấp hơn trong khi phần đệm cao được áp dụng theo hướng của các chỉ số cao hơn. Ví dụ, nếu padding[1]
là (2,3)
thì sẽ có khoảng đệm bởi 2 số 0 ở bên trái và 3 số 0 ở bên phải trong chiều không gian thứ hai. Sử dụng padding tương đương với việc chèn các giá trị 0 giống nhau đó vào đầu vào ( lhs
) trước khi thực hiện phép tích chập.
Các đối số lhs_dilation
và rhs_dilation
chỉ định hệ số giãn nở được áp dụng cho lhs và rhs, tương ứng, trong mỗi chiều không gian. Nếu hệ số giãn nở trong một chiều không gian là d, thì d-1 lỗ được đặt ngầm giữa mỗi mục trong chiều đó, làm tăng kích thước của mảng. Các lỗ được lấp đầy bằng giá trị no-op, đối với tích chập có nghĩa là số 0.
Sự giãn nở của rhs còn được gọi là tích chập bất thường. Để biết thêm chi tiết, hãy xem tf.nn.atrous_conv2d
. Sự giãn nở của lhs còn được gọi là sự tích chập chuyển vị. Để biết thêm chi tiết, hãy xem tf.nn.conv2d_transpose
.
Đối số feature_group_count
(giá trị mặc định 1) có thể được sử dụng cho các chập được nhóm. feature_group_count
cần phải là một ước của cả thứ nguyên tính năng đầu vào và đầu ra. Nếu feature_group_count
lớn hơn 1, điều đó có nghĩa là về mặt khái niệm, thứ nguyên tính năng đầu vào và đầu ra và thứ nguyên tính năng đầu ra rhs
được chia đều thành nhiều nhóm feature_group_count
, mỗi nhóm bao gồm một dãy con liên tiếp của các tính năng. Kích thước tính năng đầu vào của rhs
cần bằng với kích thước tính năng đầu vào lhs
chia cho feature_group_count
(vì vậy nó đã có kích thước của một nhóm tính năng đầu vào). Các nhóm thứ i được sử dụng cùng nhau để tính feature_group_count
nhiều phức hợp riêng biệt. Kết quả của các chập này được nối với nhau trong thứ nguyên tính năng đầu ra.
Đối với phép chập theo chiều sâu, đối số feature_group_count
sẽ được đặt thành thứ nguyên tính năng đầu vào và bộ lọc sẽ được định hình lại từ [filter_height, filter_width, in_channels, channel_multiplier]
thành [filter_height, filter_width, 1, in_channels * channel_multiplier]
. Để biết thêm chi tiết, hãy xem tf.nn.depthwise_conv2d
.
Đối batch_group_count
(giá trị mặc định 1) có thể được sử dụng cho các bộ lọc được nhóm trong quá trình nhân giống ngược. batch_group_count
cần phải là một ước số của kích thước của thứ nguyên lô lhs
(đầu vào). Nếu batch_group_count
lớn hơn 1, điều đó có nghĩa là thứ nguyên lô đầu ra phải có kích thước input batch / batch_group_count
. batch_group_count
phải là ước số của kích thước tính năng đầu ra.
Hình dạng đầu ra có các kích thước này, theo thứ tự sau:
-
batch
: Kích thước của thứ nguyên này nhân vớibatch_group_count
phải bằng kích thước của thứ nguyênbatch
tính bằng lhs. -
z
: Cùng kích thước vớioutput-z
trên kernel (rhs
). -
spatial_dims
: Một giá trị cho mỗi vị trí hợp lệ của cửa sổ tích hợp.
Các vị trí hợp lệ của cửa sổ tích chập được xác định bởi các bước và kích thước của vùng cơ sở sau khi đệm.
Để mô tả chức năng của một tích chập, hãy xem xét một tích chập 2d và chọn một số tọa độ batch
, z
, y
, x
cố định trong đầu ra. Khi đó (y,x)
là vị trí của một góc của cửa sổ trong vùng cơ sở (ví dụ: góc trên bên trái, tùy thuộc vào cách bạn diễn giải các kích thước không gian). Bây giờ chúng ta có một cửa sổ 2d, được lấy từ vùng cơ sở, trong đó mỗi điểm 2d được liên kết với một vectơ 1d, vì vậy chúng ta nhận được một hộp 3d. Từ hạt nhân chập, vì chúng ta đã cố định tọa độ đầu ra z
, chúng ta cũng có một hộp 3d. Hai hộp có cùng kích thước, vì vậy chúng ta có thể lấy tổng các tích theo phần tử giữa hai hộp (tương tự như một tích chấm). Đó là giá trị đầu ra.
Lưu ý rằng nếu output-z
là 5, thì mỗi vị trí của cửa sổ tạo ra 5 giá trị trong đầu ra vào kích thước z
của đầu ra. Các giá trị này khác nhau ở phần nào của nhân tích chập được sử dụng - có một hộp giá trị 3d riêng biệt được sử dụng cho mỗi tọa độ output-z
. Vì vậy, bạn có thể nghĩ về nó như là 5 phần chập riêng biệt với một bộ lọc khác nhau cho mỗi phần trong số chúng.
Đây là mã giả cho một phép chập 2d với phần đệm và bước:
for (b, oz, oy, ox) { // output coordinates
value = 0;
for (iz, ky, kx) { // kernel coordinates and input z
iy = oy*stride_y + ky - pad_low_y;
ix = ox*stride_x + kx - pad_low_x;
if ((iy, ix) inside the base area considered without padding) {
value += input(b, iz, iy, ix) * kernel(oz, iz, ky, kx);
}
}
output(b, oz, oy, ox) = value;
}
ConvertElementType
Xem thêm XlaBuilder::ConvertElementType
.
Tương tự như static_cast
phần tử khôn ngoan trong C ++, thực hiện thao tác chuyển đổi phần tử khôn ngoan từ hình dạng dữ liệu sang hình dạng đích. Các thứ nguyên phải phù hợp và chuyển đổi là một yếu tố khôn ngoan; ví dụ: các phần tử s32
trở thành phần tử f32
thông qua một quy trình chuyển đổi s32
-to- f32
.
ConvertElementType(operand, new_element_type)
Tranh luận | Gõ phím | Ngữ nghĩa học |
---|---|---|
operand | XlaOp | mảng kiểu T với độ mờ D |
new_element_type | PrimitiveType | gõ U |
Kích thước của toán hạng và hình dạng đích phải khớp nhau. Các loại phần tử nguồn và đích không được là bộ giá trị.
Một chuyển đổi chẳng hạn như T=s32
thành U=f32
sẽ thực hiện quy trình chuyển đổi int-to-float chuẩn hóa chẳng hạn như làm tròn đến chẵn.
let a: s32[3] = {0, 1, 2};
let b: f32[3] = convert(a, f32);
then b == f32[3]{0.0, 1.0, 2.0}
CrossReplicaSum
Thực hiện AllReduce
với một phép tính tổng kết.
Tối ưu hóaBarrier
Chặn mọi tính toán di chuyển vượt qua rào cản.
Đảm bảo rằng tất cả các đầu vào được đánh giá trước bất kỳ nhà điều hành nào phụ thuộc vào đầu ra của rào cản.
CustomCall
Xem thêm XlaBuilder::CustomCall
.
Gọi một hàm do người dùng cung cấp trong một tính toán.
CustomCall(target_name, args..., shape)
Tranh luận | Gõ phím | Ngữ nghĩa học |
---|---|---|
target_name | string | Tên của chức năng. Lệnh gọi sẽ được phát ra nhắm vào tên biểu tượng này. |
args | dãy N XlaOp s | N đối số kiểu tùy ý, sẽ được chuyển cho hàm. |
shape | Shape | Hình dạng đầu ra của hàm |
Chữ ký hàm giống nhau, bất kể mức độ hiếm có hoặc loại args:
extern "C" void target_name(void* out, void** in);
Ví dụ: nếu CustomCall được sử dụng như sau:
let x = f32[2] {1,2};
let y = f32[2x3] { {10, 20, 30}, {40, 50, 60} };
CustomCall("myfunc", {x, y}, f32[3x3])
Đây là một ví dụ về việc triển khai myfunc
:
extern "C" void myfunc(void* out, void** in) {
float (&x)[2] = *static_cast<float(*)[2]>(in[0]);
float (&y)[2][3] = *static_cast<float(*)[2][3]>(in[1]);
EXPECT_EQ(1, x[0]);
EXPECT_EQ(2, x[1]);
EXPECT_EQ(10, y[0][0]);
EXPECT_EQ(20, y[0][1]);
EXPECT_EQ(30, y[0][2]);
EXPECT_EQ(40, y[1][0]);
EXPECT_EQ(50, y[1][1]);
EXPECT_EQ(60, y[1][2]);
float (&z)[3][3] = *static_cast<float(*)[3][3]>(out);
z[0][0] = x[1] + y[1][0];
// ...
}
Hàm do người dùng cung cấp không được có tác dụng phụ và quá trình thực thi của nó phải tốt.
Chấm
Xem thêm XlaBuilder::Dot
.
Dot(lhs, rhs)
Tranh luận | Gõ phím | Ngữ nghĩa học |
---|---|---|
lhs | XlaOp | mảng kiểu T |
rhs | XlaOp | mảng kiểu T |
Ngữ nghĩa chính xác của thao tác này phụ thuộc vào cấp bậc của các toán hạng:
Đầu vào | Đầu ra | Ngữ nghĩa học |
---|---|---|
vectơ [n] dot vectơ [n] | vô hướng | sản phẩm chấm vector |
ma trận [mxk] vectơ dot [k] | vectơ [m] | phép nhân vectơ-ma trận |
ma trận [mxk] ma trận dot [kxn] | ma trận [mxn] | phép nhân ma trận |
Phép toán thực hiện tính tổng của các sản phẩm trên chiều thứ hai của lhs
(hoặc thứ nhất nếu nó có hạng 1) và chiều thứ nhất của rhs
. Đây là các kích thước "hợp đồng". Kích thước theo hợp đồng của lhs
và rhs
phải có cùng kích thước. Trong thực tế, nó có thể được sử dụng để thực hiện các tích số chấm giữa các vectơ, phép nhân vectơ / ma trận hoặc phép nhân ma trận / ma trận.
DotGeneral
Xem thêm XlaBuilder::DotGeneral
.
DotGeneral(lhs, rhs, dimension_numbers)
Tranh luận | Gõ phím | Ngữ nghĩa học |
---|---|---|
lhs | XlaOp | mảng kiểu T |
rhs | XlaOp | mảng kiểu T |
dimension_numbers | DotDimensionNumbers | hợp đồng và số thứ nguyên lô |
Là dấu chấm, nhưng cho phép chỉ định số thứ nguyên hợp đồng và lô cho cả 'lhs' và 'rhs'.
Trường DotDimensionNumbers | Gõ phím | Ngữ nghĩa học |
---|---|---|
'lhs_contract_dimensions' | int64 lặp lại | số thứ nguyên hợp đồng 'lhs' |
'rhs_contract_dimensions' | int64 lặp lại | số thứ nguyên hợp đồng 'rhs' |
'lhs_batch_dimensions' | int64 lặp lại | số thứ nguyên lô 'lhs' |
'rhs_batch_dimensions' | int64 lặp lại | số thứ nguyên lô 'rhs' |
DotGeneral thực hiện tổng số sản phẩm trên các kích thước hợp đồng được chỉ định trong 'số_thứ_lượng'.
Các số thứ nguyên trong hợp đồng được liên kết từ 'lhs' và 'rhs' không cần giống nhau nhưng phải có cùng kích thước thứ nguyên.
Ví dụ với số thứ nguyên theo hợp đồng:
lhs = { {1.0, 2.0, 3.0},
{4.0, 5.0, 6.0} }
rhs = { {1.0, 1.0, 1.0},
{2.0, 2.0, 2.0} }
DotDimensionNumbers dnums;
dnums.add_lhs_contracting_dimensions(1);
dnums.add_rhs_contracting_dimensions(1);
DotGeneral(lhs, rhs, dnums) -> { {6.0, 12.0},
{15.0, 30.0} }
Số thứ nguyên lô được liên kết từ 'lhs' và 'rhs' phải có cùng kích thước thứ nguyên.
Ví dụ với số thứ nguyên lô (kích thước lô 2, ma trận 2x2):
lhs = { { {1.0, 2.0},
{3.0, 4.0} },
{ {5.0, 6.0},
{7.0, 8.0} } }
rhs = { { {1.0, 0.0},
{0.0, 1.0} },
{ {1.0, 0.0},
{0.0, 1.0} } }
DotDimensionNumbers dnums;
dnums.add_lhs_contracting_dimensions(2);
dnums.add_rhs_contracting_dimensions(1);
dnums.add_lhs_batch_dimensions(0);
dnums.add_rhs_batch_dimensions(0);
DotGeneral(lhs, rhs, dnums) -> { { {1.0, 2.0},
{3.0, 4.0} },
{ {5.0, 6.0},
{7.0, 8.0} } }
Đầu vào | Đầu ra | Ngữ nghĩa học |
---|---|---|
[b0, m, k] dot [b0, k, n] | [b0, m, n] | matmul lô |
[b0, b1, m, k] dot [b0, b1, k, n] | [b0, b1, m, n] | matmul lô |
Theo đó, số thứ nguyên kết quả bắt đầu với thứ nguyên lô, sau đó là thứ nguyên không hợp đồng / không theo lô 'lhs' và cuối cùng là thứ nguyên không hợp đồng / không theo lô 'rhs'.
DynamicSlice
Xem thêm XlaBuilder::DynamicSlice
.
DynamicSlice trích xuất một mảng con từ mảng đầu vào tại dynamic start_indices
. Kích thước của lát trong mỗi thứ nguyên được chuyển trong size_indices
, chỉ định điểm cuối của các khoảng lát riêng trong mỗi thứ nguyên: [start, start + size). Hình dạng của start_indices
phải là rank == 1, với kích thước thứ nguyên bằng thứ hạng của operand
.
DynamicSlice(operand, start_indices, size_indices)
Tranh luận | Gõ phím | Ngữ nghĩa học |
---|---|---|
operand | XlaOp | Mảng N chiều kiểu T |
start_indices | trình tự của N XlaOp | Danh sách N số nguyên vô hướng chứa các chỉ số bắt đầu của lát cắt cho mỗi chiều. Giá trị phải lớn hơn hoặc bằng không. |
size_indices | ArraySlice<int64> | Danh sách N số nguyên chứa kích thước lát cắt cho mỗi chiều. Each value must be strictly greater than zero, and start + size must be less than or equal to the size of the dimension to avoid wrapping modulo dimension size. |
The effective slice indices are computed by applying the following transformation for each index i
in [1, N)
before performing the slice:
start_indices[i] = clamp(start_indices[i], 0, operand.dimension_size[i] - size_indices[i])
This ensures that the extracted slice is always in-bounds with respect to the operand array. If the slice is in-bounds before the transformation is applied, the transformation has no effect.
1-dimensional example:
let a = {0.0, 1.0, 2.0, 3.0, 4.0}
let s = {2}
DynamicSlice(a, s, {2}) produces:
{2.0, 3.0}
2-dimensional example:
let b =
{ {0.0, 1.0, 2.0},
{3.0, 4.0, 5.0},
{6.0, 7.0, 8.0},
{9.0, 10.0, 11.0} }
let s = {2, 1}
DynamicSlice(b, s, {2, 2}) produces:
{ { 7.0, 8.0},
{10.0, 11.0} }
DynamicUpdateSlice
See also XlaBuilder::DynamicUpdateSlice
.
DynamicUpdateSlice generates a result which is the value of the input array operand
, with a slice update
overwritten at start_indices
. The shape of update
determines the shape of the sub-array of the result which is updated. The shape of start_indices
must be rank == 1, with dimension size equal to the rank of operand
.
DynamicUpdateSlice(operand, update, start_indices)
Arguments | Type | Semantics |
---|---|---|
operand | XlaOp | N dimensional array of type T |
update | XlaOp | N dimensional array of type T containing the slice update. Each dimension of update shape must be strictly greater than zero, and start + update must be less than or equal to the operand size for each dimension to avoid generating out-of-bounds update indices. |
start_indices | sequence of N XlaOp | List of N scalar integers containing the starting indices of the slice for each dimension. Value must be greater than or equal to zero. |
The effective slice indices are computed by applying the following transformation for each index i
in [1, N)
before performing the slice:
start_indices[i] = clamp(start_indices[i], 0, operand.dimension_size[i] - update.dimension_size[i])
This ensures that the updated slice is always in-bounds with respect to the operand array. If the slice is in-bounds before the transformation is applied, the transformation has no effect.
1-dimensional example:
let a = {0.0, 1.0, 2.0, 3.0, 4.0}
let u = {5.0, 6.0}
let s = {2}
DynamicUpdateSlice(a, u, s) produces:
{0.0, 1.0, 5.0, 6.0, 4.0}
2-dimensional example:
let b =
{ {0.0, 1.0, 2.0},
{3.0, 4.0, 5.0},
{6.0, 7.0, 8.0},
{9.0, 10.0, 11.0} }
let u =
{ {12.0, 13.0},
{14.0, 15.0},
{16.0, 17.0} }
let s = {1, 1}
DynamicUpdateSlice(b, u, s) produces:
{ {0.0, 1.0, 2.0},
{3.0, 12.0, 13.0},
{6.0, 14.0, 15.0},
{9.0, 16.0, 17.0} }
Element-wise binary arithmetic operations
See also XlaBuilder::Add
.
A set of element-wise binary arithmetic operations is supported.
Op(lhs, rhs)
Where Op
is one of Add
(addition), Sub
(subtraction), Mul
(multiplication), Div
(division), Rem
(remainder), Max
(maximum), Min
(minimum), LogicalAnd
(logical AND), or LogicalOr
(logical OR).
Arguments | Type | Semantics |
---|---|---|
lhs | XlaOp | left-hand-side operand: array of type T |
rhs | XlaOp | right-hand-side operand: array of type T |
The arguments' shapes have to be either similar or compatible. See the broadcasting documentation about what it means for shapes to be compatible. The result of an operation has a shape which is the result of broadcasting the two input arrays. In this variant, operations between arrays of different ranks are not supported, unless one of the operands is a scalar.
When Op
is Rem
, the sign of the result is taken from the dividend, and the absolute value of the result is always less than the divisor's absolute value.
Integer division overflow (signed/unsigned division/remainder by zero or signed division/remainder of INT_SMIN
with -1
) produces an implementation defined value.
An alternative variant with different-rank broadcasting support exists for these operations:
Op(lhs, rhs, broadcast_dimensions)
Where Op
is the same as above. This variant of the operation should be used for arithmetic operations between arrays of different ranks (such as adding a matrix to a vector).
The additional broadcast_dimensions
operand is a slice of integers used to expand the rank of the lower-rank operand up to the rank of the higher-rank operand. broadcast_dimensions
maps the dimensions of the lower-rank shape to the dimensions of the higher-rank shape. The unmapped dimensions of the expanded shape are filled with dimensions of size one. Degenerate-dimension broadcasting then broadcasts the shapes along these degenerate dimensions to equalize the shapes of both operands. The semantics are described in detail on the broadcasting page .
Element-wise comparison operations
See also XlaBuilder::Eq
.
A set of standard element-wise binary comparison operations is supported. Note that standard IEEE 754 floating-point comparison semantics apply when comparing floating-point types.
Op(lhs, rhs)
Where Op
is one of Eq
(equal-to), Ne
(not equal-to), Ge
(greater-or-equal-than), Gt
(greater-than), Le
(less-or-equal-than), Lt
(less-than). Another set of operators, EqTotalOrder, NeTotalOrder, GeTotalOrder, GtTotalOrder, LeTotalOrder, and LtTotalOrder, provide the same functionalities, except that they additionally support a total order over the floating point numbers, by enforcing -NaN < -Inf < -Finite < -0 < +0 < +Finite < +Inf < +NaN.
Arguments | Type | Semantics |
---|---|---|
lhs | XlaOp | left-hand-side operand: array of type T |
rhs | XlaOp | right-hand-side operand: array of type T |
The arguments' shapes have to be either similar or compatible. See the broadcasting documentation about what it means for shapes to be compatible. The result of an operation has a shape which is the result of broadcasting the two input arrays with the element type PRED
. In this variant, operations between arrays of different ranks are not supported, unless one of the operands is a scalar.
An alternative variant with different-rank broadcasting support exists for these operations:
Op(lhs, rhs, broadcast_dimensions)
Where Op
is the same as above. This variant of the operation should be used for comparison operations between arrays of different ranks (such as adding a matrix to a vector).
The additional broadcast_dimensions
operand is a slice of integers specifying the dimensions to use for broadcasting the operands. The semantics are described in detail on the broadcasting page .
Element-wise unary functions
XlaBuilder supports these element-wise unary functions:
Abs(operand)
Element-wise abs x -> |x|
.
Ceil(operand)
Element-wise ceil x -> ⌈x⌉
.
Cos(operand)
Element-wise cosine x -> cos(x)
.
Exp(operand)
Element-wise natural exponential x -> e^x
.
Floor(operand)
Element-wise floor x -> ⌊x⌋
.
Imag(operand)
Element-wise imaginary part of a complex (or real) shape. x -> imag(x)
. If the operand is a floating point type, returns 0.
IsFinite(operand)
Tests whether each element of operand
is finite, ie, is not positive or negative infinity, and is not NaN
. Returns an array of PRED
values with the same shape as the input, where each element is true
if and only if the corresponding input element is finite.
Log(operand)
Element-wise natural logarithm x -> ln(x)
.
LogicalNot(operand)
Element-wise logical not x -> !(x)
.
Logistic(operand)
Element-wise logistic function computation x -> logistic(x)
.
PopulationCount(operand)
Computes the number of bits set in each element of operand
.
Neg(operand)
Element-wise negation x -> -x
.
Real(operand)
Element-wise real part of a complex (or real) shape. x -> real(x)
. If the operand is a floating point type, returns the same value.
Rsqrt(operand)
Element-wise reciprocal of square root operation x -> 1.0 / sqrt(x)
.
Sign(operand)
Element-wise sign operation x -> sgn(x)
where
\[\text{sgn}(x) = \begin{cases} -1 & x < 0\\ -0 & x = -0\\ NaN & x = NaN\\ +0 & x = +0\\ 1 & x > 0 \end{cases}\]
using the comparison operator of the element type of operand
.
Sqrt(operand)
Element-wise square root operation x -> sqrt(x)
.
Cbrt(operand)
Element-wise cubic root operation x -> cbrt(x)
.
Tanh(operand)
Element-wise hyperbolic tangent x -> tanh(x)
.
Arguments | Type | Semantics |
---|---|---|
operand | XlaOp | The operand to the function |
The function is applied to each element in the operand
array, resulting in an array with the same shape. It is allowed for operand
to be a scalar (rank 0).
Fft
The XLA FFT operation implements the forward and inverse Fourier Transforms for real and complex inputs/outputs. Multidimensional FFTs on up to 3 axes are supported.
See also XlaBuilder::Fft
.
Arguments | Type | Semantics |
---|---|---|
operand | XlaOp | The array we are Fourier transforming. |
fft_type | FftType | See the table below. |
fft_length | ArraySlice<int64> | The time-domain lengths of the axes being transformed. This is needed in particular for IRFFT to right-size the innermost axis, since RFFT(fft_length=[16]) has the same output shape as RFFT(fft_length=[17]) . |
FftType | Semantics |
---|---|
FFT | Forward complex-to-complex FFT. Shape is unchanged. |
IFFT | Inverse complex-to-complex FFT. Shape is unchanged. |
RFFT | Forward real-to-complex FFT. Shape of the innermost axis is reduced to fft_length[-1] // 2 + 1 if fft_length[-1] is a non-zero value, omitting the reversed conjugate part of the transformed signal beyond the Nyquist frequency. |
IRFFT | Inverse real-to-complex FFT (ie takes complex, returns real). Shape of the innermost axis is expanded to fft_length[-1] if fft_length[-1] is a non-zero value, inferring the part of the transformed signal beyond the Nyquist frequency from the reverse conjugate of the 1 to fft_length[-1] // 2 + 1 entries. |
Multidimensional FFT
When more than 1 fft_length
is provided, this is equivalent to applying a cascade of FFT operations to each of the innermost axes. Note that for the real->complex and complex->real cases, the innermost axis transform is (effectively) performed first (RFFT; last for IRFFT), which is why the innermost axis is the one which changes size. Other axis transforms will then be complex->complex.
Implementation details
CPU FFT is backed by Eigen's TensorFFT. GPU FFT uses cuFFT.
Gather
The XLA gather operation stitches together several slices (each slice at a potentially different runtime offset) of an input array.
General Semantics
See also XlaBuilder::Gather
. For a more intuitive description, see the "Informal Description" section below.
gather(operand, start_indices, offset_dims, collapsed_slice_dims, slice_sizes, start_index_map)
Arguments | Type | Semantics |
---|---|---|
operand | XlaOp | The array we're gathering from. |
start_indices | XlaOp | Array containing the starting indices of the slices we gather. |
index_vector_dim | int64 | The dimension in start_indices that "contains" the starting indices. See below for a detailed description. |
offset_dims | ArraySlice<int64> | The set of dimensions in the output shape that offset into an array sliced from operand. |
slice_sizes | ArraySlice<int64> | slice_sizes[i] is the bounds for the slice on dimension i . |
collapsed_slice_dims | ArraySlice<int64> | The set of dimensions in each slice that are collapsed away. These dimensions must have size 1. |
start_index_map | ArraySlice<int64> | A map that describes how to map indices in start_indices to legal indices into operand. |
indices_are_sorted | bool | Whether the indices are guaranteed to be sorted by the caller. |
unique_indices | bool | Whether the indices are guaranteed to be unique by the caller. |
For convenience, we label dimensions in the output array not in offset_dims
as batch_dims
.
The output is an array of rank batch_dims.size
+ offset_dims.size
.
The operand.rank
must equal the sum of offset_dims.size
and collapsed_slice_dims.size
. Also, slice_sizes.size
has to be equal to operand.rank
.
If index_vector_dim
is equal to start_indices.rank
we implicitly consider start_indices
to have a trailing 1
dimension (ie if start_indices
was of shape [6,7]
and index_vector_dim
is 2
then we implicitly consider the shape of start_indices
to be [6,7,1]
).
The bounds for the output array along dimension i
is computed as follows:
If
i
is present inbatch_dims
(ie is equal tobatch_dims[k]
for somek
) then we pick the corresponding dimension bounds out ofstart_indices.shape
, skippingindex_vector_dim
(ie pickstart_indices.shape.dims
[k
] ifk
<index_vector_dim
andstart_indices.shape.dims
[k
+1
] otherwise).If
i
is present inoffset_dims
(ie equal tooffset_dims
[k
] for somek
) then we pick the corresponding bound out ofslice_sizes
after accounting forcollapsed_slice_dims
(ie we pickadjusted_slice_sizes
[k
] whereadjusted_slice_sizes
isslice_sizes
with the bounds at indicescollapsed_slice_dims
removed).
Formally, the operand index In
corresponding to a given output index Out
is calculated as follows:
Let
G
= {Out
[k
] fork
inbatch_dims
}. UseG
to slice out a vectorS
such thatS
[i
] =start_indices
[Combine(G
,i
)] where Combine(A, b) inserts b at positionindex_vector_dim
into A. Note that this is well defined even ifG
is empty -- ifG
is empty thenS
=start_indices
.Create a starting index,
S
in
, intooperand
usingS
by scatteringS
usingstart_index_map
. More precisely:S
in
[start_index_map
[k
]] =S
[k
] ifk
<start_index_map.size
.S
in
[_
] =0
otherwise.
Create an index
O
in
intooperand
by scattering the indices at the offset dimensions inOut
according to thecollapsed_slice_dims
set. More precisely:O
in
[remapped_offset_dims
(k
)] =Out
[offset_dims
[k
]] ifk
<offset_dims.size
(remapped_offset_dims
is defined below).O
in
[_
] =0
otherwise.
In
isO
in
+S
in
where + is element-wise addition.
remapped_offset_dims
is a monotonic function with domain [ 0
, offset_dims.size
) and range [ 0
, operand.rank
) \ collapsed_slice_dims
. So if, eg, offset_dims.size
is 4
, operand.rank
is 6
and collapsed_slice_dims
is { 0
, 2
} then remapped_offset_dims
is { 0
→ 1
, 1
→ 3
, 2
→ 4
, 3
→ 5
}.
If indices_are_sorted
is set to true then XLA can assume that start_indices
are sorted (in ascending start_index_map
order) by the user. If they are not then the semantics is implementation defined.
If unique_indices
is set to true then XLA can assume that all element scattered to are unique. So XLA could use non-atomic operations. If unique_indices
is set to true and the indices being scattered to are not unique then the semantics is implementation defined.
Informal Description and Examples
Informally, every index Out
in the output array corresponds to an element E
in the operand array, computed as follows:
We use the batch dimensions in
Out
to look up a starting index fromstart_indices
.We use
start_index_map
to map the starting index (whose size may be less than operand.rank) to a "full" starting index into theoperand
.We dynamic-slice out a slice with size
slice_sizes
using the full starting index.We reshape the slice by collapsing the
collapsed_slice_dims
dimensions. Since all collapsed slice dimensions must have a bound of 1, this reshape is always legal.We use the offset dimensions in
Out
to index into this slice to get the input element,E
, corresponding to output indexOut
.
index_vector_dim
is set to start_indices.rank
- 1
in all of the examples that follow. More interesting values for index_vector_dim
do not change the operation fundamentally, but make the visual representation more cumbersome.
To get an intuition on how all of the above fits together, let's look at an example that gathers 5 slices of shape [8,6]
from a [16,11]
array. The position of a slice into the [16,11]
array can be represented as an index vector of shape S64[2]
, so the set of 5 positions can be represented as a S64[5,2]
array.
The behavior of the gather operation can then be depicted as an index transformation that takes [ G
, O
0
, O
1
], an index in the output shape, and maps it to an element in the input array in the following way:
We first select an ( X
, Y
) vector from the gather indices array using G
. The element in the output array at index [ G
, O
0
, O
1
] is then the element in the input array at index [ X
+ O
0
, Y
+ O
1
].
slice_sizes
is [8,6]
, which decides the range of O 0
and O 1
, and this in turn decides the bounds of the slice.
This gather operation acts as a batch dynamic slice with G
as the batch dimension.
The gather indices may be multidimensional. For instance, a more general version of the example above using a "gather indices" array of shape [4,5,2]
would translate indices like this:
Again, this acts as a batch dynamic slice G
0
and G
1
as the batch dimensions. The slice size is still [8,6]
.
The gather operation in XLA generalizes the informal semantics outlined above in the following ways:
We can configure which dimensions in the output shape are the offset dimensions (dimensions containing
O
0
,O
1
in the last example). The output batch dimensions (dimensions containingG
0
,G
1
in the last example) are defined to be the output dimensions that are not offset dimensions.The number of output offset dimensions explicitly present in the output shape may be smaller than the input rank. These "missing" dimensions, which are listed explicitly as
collapsed_slice_dims
, must have a slice size of1
. Since they have a slice size of1
the only valid index for them is0
and eliding them does not introduce ambiguity.The slice extracted from the "Gather Indices" array ((
X
,Y
) in the last example) may have fewer elements than the input array rank, and an explicit mapping dictates how the index should be expanded to have the same rank as the input.
As a final example, we use (2) and (3) to implement tf.gather_nd
:
G
0
and G
1
are used to slice out a starting index from the gather indices array as usual, except the starting index has only one element, X
. Similarly, there is only one output offset index with the value O
0
. However, before being used as indices into the input array, these are expanded in accordance to "Gather Index Mapping" ( start_index_map
in the formal description) and "Offset Mapping" ( remapped_offset_dims
in the formal description) into [ X
, 0
] and [ 0
, O
0
] respectively, adding up to [ X
, O
0
]. In other words, the output index [ G
0
, G
1
, O
0
] maps to the input index [ GatherIndices
[ G
0
, G
1
, 0
], X
] which gives us the semantics for tf.gather_nd
.
slice_sizes
for this case is [1,11]
. Intuitively this means that every index X
in the gather indices array picks an entire row and the result is the concatenation of all these rows.
GetDimensionSize
See also XlaBuilder::GetDimensionSize
.
Returns the size of the given dimension of the operand. The operand must be array shaped.
GetDimensionSize(operand, dimension)
Arguments | Type | Semantics |
---|---|---|
operand | XlaOp | n dimensional input array |
dimension | int64 | A value in the interval [0, n) that specifies the dimension |
SetDimensionSize
See also XlaBuilder::SetDimensionSize
.
Sets the dynamic size of XlaOp's given dimension. The operand must be array shaped.
SetDimensionSize(operand, size, dimension)
Arguments | Type | Semantics |
---|---|---|
operand | XlaOp | n dimensional input array. |
size | XlaOp | int32 representing the runtime dynamic size. |
dimension | int64 | A value in the interval [0, n) that specifies the dimension. |
Pass through the operand as result, with dynamic dimension tracked by the compiler.
Padded values will be ignored by downstream reduction ops.
let v: f32[10] = f32[10]{1, 2, 3, 4, 5, 6, 7, 8, 9, 10};
let five: s32 = 5;
let six: s32 = 6;
// Setting dynamic dimension size doesn't change the upper bound of the static
// shape.
let padded_v_five: f32[10] = set_dimension_size(v, five, /*dimension=*/0);
let padded_v_six: f32[10] = set_dimension_size(v, six, /*dimension=*/0);
// sum == 1 + 2 + 3 + 4 + 5
let sum:f32[] = reduce_sum(padded_v_five);
// product == 1 * 2 * 3 * 4 * 5
let product:f32[] = reduce_product(padded_v_five);
// Changing padding size will yield different result.
// sum == 1 + 2 + 3 + 4 + 5 + 6
let sum':f32[] = reduce_sum(padded_v_six);
GetTupleElement
See also XlaBuilder::GetTupleElement
.
Indexes into a tuple with a compile-time-constant value.
The value must be a compile-time-constant so that shape inference can determine the type of the resulting value.
This is analogous to std::get<int N>(t)
in C++. Conceptually:
let v: f32[10] = f32[10]{0, 1, 2, 3, 4, 5, 6, 7, 8, 9};
let s: s32 = 5;
let t: (f32[10], s32) = tuple(v, s);
let element_1: s32 = gettupleelement(t, 1); // Inferred shape matches s32.
See also tf.tuple
.
Infeed
See also XlaBuilder::Infeed
.
Infeed(shape)
Argument | Type | Semantics |
---|---|---|
shape | Shape | Shape of the data read from the Infeed interface. The layout field of the shape must be set to match the layout of the data sent to the device; otherwise its behavior is undefined. |
Reads a single data item from the implicit Infeed streaming interface of the device, interpreting the data as the given shape and its layout, and returns a XlaOp
of the data. Multiple Infeed operations are allowed in a computation, but there must be a total order among the Infeed operations. For example, two Infeeds in the code below have a total order since there is a dependency between the while loops.
result1 = while (condition, init = init_value) {
Infeed(shape)
}
result2 = while (condition, init = result1) {
Infeed(shape)
}
Nested tuple shapes are not supported. For an empty tuple shape, the Infeed operation is effectively a no-op and proceeds without reading any data from the Infeed of the device.
Iota
See also XlaBuilder::Iota
.
Iota(shape, iota_dimension)
Builds a constant literal on device rather than a potentially large host transfer. Creates an array that has specified shape and holds values starting at zero and incrementing by one along the specified dimension. For floating-point types, the produced array is equivalent to ConvertElementType(Iota(...))
where the Iota
is of integral type and the conversion is to the floating-point type.
Arguments | Type | Semantics |
---|---|---|
shape | Shape | Shape of the array created by Iota() |
iota_dimension | int64 | The dimension to increment along. |
For example, Iota(s32[4, 8], 0)
returns
[[0, 0, 0, 0, 0, 0, 0, 0 ],
[1, 1, 1, 1, 1, 1, 1, 1 ],
[2, 2, 2, 2, 2, 2, 2, 2 ],
[3, 3, 3, 3, 3, 3, 3, 3 ]]
Iota(s32[4, 8], 1)
returns
[[0, 1, 2, 3, 4, 5, 6, 7 ],
[0, 1, 2, 3, 4, 5, 6, 7 ],
[0, 1, 2, 3, 4, 5, 6, 7 ],
[0, 1, 2, 3, 4, 5, 6, 7 ]]
Map
See also XlaBuilder::Map
.
Map(operands..., computation)
Arguments | Type | Semantics |
---|---|---|
operands | sequence of N XlaOp s | N arrays of types T 0..T {N-1} |
computation | XlaComputation | computation of type T_0, T_1, ..., T_{N + M -1} -> S with N parameters of type T and M of arbitrary type |
dimensions | int64 array | array of map dimensions |
Applies a scalar function over the given operands
arrays, producing an array of the same dimensions where each element is the result of the mapped function applied to the corresponding elements in the input arrays.
The mapped function is an arbitrary computation with the restriction that it has N inputs of scalar type T
and a single output with type S
. The output has the same dimensions as the operands except that the element type T is replaced with S.
For example: Map(op1, op2, op3, computation, par1)
maps elem_out <- computation(elem1, elem2, elem3, par1)
at each (multi-dimensional) index in the input arrays to produce the output array.
Pad
See also XlaBuilder::Pad
.
Pad(operand, padding_value, padding_config)
Arguments | Type | Semantics |
---|---|---|
operand | XlaOp | array of type T |
padding_value | XlaOp | scalar of type T to fill in the added padding |
padding_config | PaddingConfig | padding amount on both edges (low, high) and between the elements of each dimension |
Expands the given operand
array by padding around the array as well as between the elements of the array with the given padding_value
. padding_config
specifies the amount of edge padding and the interior padding for each dimension.
PaddingConfig
is a repeated field of PaddingConfigDimension
, which contains three fields for each dimension: edge_padding_low
, edge_padding_high
, and interior_padding
.
edge_padding_low
and edge_padding_high
specify the amount of padding added at the low-end (next to index 0) and the high-end (next to the highest index) of each dimension respectively. The amount of edge padding can be negative -- the absolute value of negative padding indicates the number of elements to remove from the specified dimension.
interior_padding
specifies the amount of padding added between any two elements in each dimension; it may not be negative. Interior padding occurs logically before edge padding, so in the case of negative edge padding, elements are removed from the interior-padded operand.
This operation is a no-op if the edge padding pairs are all (0, 0) and the interior padding values are all 0. The figure below shows examples of different edge_padding
and interior_padding
values for a two-dimensional array.

Recv
See also XlaBuilder::Recv
.
Recv(shape, channel_handle)
Arguments | Type | Semantics |
---|---|---|
shape | Shape | shape of the data to receive |
channel_handle | ChannelHandle | unique identifier for each send/recv pair |
Receives data of the given shape from a Send
instruction in another computation that shares the same channel handle. Returns a XlaOp for the received data.
The client API of Recv
operation represents synchronous communication. However, the instruction is internally decomposed into 2 HLO instructions ( Recv
and RecvDone
) to enable asynchronous data transfers. See also HloInstruction::CreateRecv
and HloInstruction::CreateRecvDone
.
Recv(const Shape& shape, int64 channel_id)
Allocates resources required to receive data from a Send
instruction with the same channel_id. Returns a context for the allocated resources, which is used by a following RecvDone
instruction to wait for the completion of the data transfer. The context is a tuple of {receive buffer (shape), request identifier (U32)} and it can only be used by a RecvDone
instruction.
RecvDone(HloInstruction context)
Given a context created by a Recv
instruction, waits for the data transfer to complete and returns the received data.
Reduce
See also XlaBuilder::Reduce
.
Applies a reduction function to one or more arrays in parallel.
Reduce(operands..., init_values..., computation, dimensions)
Arguments | Type | Semantics |
---|---|---|
operands | Sequence of N XlaOp | N arrays of types T_0, ..., T_{N-1} . |
init_values | Sequence of N XlaOp | N scalars of types T_0, ..., T_{N-1} . |
computation | XlaComputation | computation of type T_0, ..., T_{N-1}, T_0, ..., T_{N-1} -> Collate(T_0, ..., T_{N-1}) . |
dimensions | int64 array | unordered array of dimensions to reduce. |
Where:
- N is required to be greater or equal to 1.
- The computation has to be "roughly" associative (see below).
- All input arrays must have the same dimensions.
- All initial values have to form an identity under
computation
. - If
N = 1
,Collate(T)
isT
. - If
N > 1
,Collate(T_0, ..., T_{N-1})
is a tuple ofN
elements of typeT
.
This operation reduces one or more dimensions of each input array into scalars. The rank of each returned array is rank(operand) - len(dimensions)
. The output of the op is Collate(Q_0, ..., Q_N)
where Q_i
is an array of type T_i
, the dimensions of which are described below.
Different backends are allowed to reassociate the reduction computation. This can lead to numerical differences, as some reduction functions like addition are not associative for floats. However, if the range of the data is limited, floating-point addition is close enough to being associative for most practical uses.
Examples
When reducing across one dimension in a single 1D array with values [10, 11, 12, 13]
, with reduction function f
(this is computation
) then that could be computed as
f(10, f(11, f(12, f(init_value, 13)))
but there are also many other possibilities, eg
f(init_value, f(f(10, f(init_value, 11)), f(f(init_value, 12), f(init_value, 13))))
The following is a rough pseudo-code example of how reduction could be implemented, using summation as the reduction computation with an initial value of 0.
result_shape <- remove all dims in dimensions from operand_shape
# Iterate over all elements in result_shape. The number of r's here is equal
# to the rank of the result
for r0 in range(result_shape[0]), r1 in range(result_shape[1]), ...:
# Initialize this result element
result[r0, r1...] <- 0
# Iterate over all the reduction dimensions
for d0 in range(dimensions[0]), d1 in range(dimensions[1]), ...:
# Increment the result element with the value of the operand's element.
# The index of the operand's element is constructed from all ri's and di's
# in the right order (by construction ri's and di's together index over the
# whole operand shape).
result[r0, r1...] += operand[ri... di]
Here's an example of reducing a 2D array (matrix). The shape has rank 2, dimension 0 of size 2 and dimension 1 of size 3:

Results of reducing dimensions 0 or 1 with an "add" function:

Note that both reduction results are 1D arrays. The diagram shows one as column and another as row just for visual convenience.
For a more complex example, here is a 3D array. Its rank is 3, dimension 0 of size 4, dimension 1 of size 2 and dimension 2 of size 3. For simplicity, the values 1 to 6 are replicated across dimension 0.

Similarly to the 2D example, we can reduce just one dimension. If we reduce dimension 0, for example, we get a rank-2 array where all values across dimension 0 were folded into a scalar:
| 4 8 12 |
| 16 20 24 |
If we reduce dimension 2, we also get a rank-2 array where all values across dimension 2 were folded into a scalar:
| 6 15 |
| 6 15 |
| 6 15 |
| 6 15 |
Note that the relative order between the remaining dimensions in the input is preserved in the output, but some dimensions may get assigned new numbers (since the rank changes).
We can also reduce multiple dimensions. Add-reducing dimensions 0 and 1 produces the 1D array [20, 28, 36]
.
Reducing the 3D array over all its dimensions produces the scalar 84
.
Variadic Reduce
When N > 1
, reduce function application is slightly more complex, as it is applied simultaneously to all inputs. The operands are supplied to the computation in the following order:
- Running reduced value for the first operand
- ...
- Running reduced value for the N'th operand
- Input value for the first operand
- ...
- Input value for the N'th operand
For example, consider the following reduction function, which can be used to compute the max and the argmax of a 1-D array in parallel:
f: (Float, Int, Float, Int) -> Float, Int
f(max, argmax, value, index):
if value >= max:
return (value, index)
else:
return (max, argmax)
For 1-D Input arrays V = Float[N], K = Int[N]
, and init values I_V = Float, I_K = Int
, the result f_(N-1)
of reducing across the only input dimension is equivalent to the following recursive application:
f_0 = f(I_V, I_K, V_0, K_0)
f_1 = f(f_0.first, f_0.second, V_1, K_1)
...
f_(N-1) = f(f_(N-2).first, f_(N-2).second, V_(N-1), K_(N-1))
Applying this reduction to an array of values, and an array of sequential indices (ie iota), will co-iterate over the arrays, and return a tuple containing the maximal value and the matching index.
ReducePrecision
See also XlaBuilder::ReducePrecision
.
Models the effect of converting floating-point values to a lower-precision format (such as IEEE-FP16) and back to the original format. The number of exponent and mantissa bits in the lower-precision format can be specified arbitrarily, although all bit sizes may not be supported on all hardware implementations.
ReducePrecision(operand, mantissa_bits, exponent_bits)
Arguments | Type | Semantics |
---|---|---|
operand | XlaOp | array of floating-point type T . |
exponent_bits | int32 | number of exponent bits in lower-precision format |
mantissa_bits | int32 | number of mantissa bits in lower-precision format |
The result is an array of type T
. The input values are rounded to the nearest value representable with the given number of mantissa bits (using "ties to even" semantics), and any values that exceed the range specified by the number of exponent bits are clamped to positive or negative infinity. NaN
values are retained, although they may be converted to canonical NaN
values.
The lower-precision format must have at least one exponent bit (in order to distinguish a zero value from an infinity, since both have a zero mantissa), and must have a non-negative number of mantissa bits. The number of exponent or mantissa bits may exceed the corresponding value for type T
; the corresponding portion of the conversion is then simply a no-op.
ReduceScatter
See also XlaBuilder::ReduceScatter
.
ReduceScatter is a collective operation that effectively does an AllReduce and then scatters the result by splitting it into shard_count
blocks along the scatter_dimension
and replica i
in the replica group receives the ith
shard.
ReduceScatter(operand, computation, scatter_dim, shard_count, replica_group_ids, channel_id)
Arguments | Type | Semantics |
---|---|---|
operand | XlaOp | Array or a non-empty tuple of arrays to reduce across replicas. |
computation | XlaComputation | Reduction computation |
scatter_dimension | int64 | Dimension to scatter. |
shard_count | int64 | Number of blocks to split scatter_dimension |
replica_groups | vector of vectors of int64 | Groups between which the reductions are performed |
channel_id | optional int64 | Optional channel ID for cross-module communication |
- When
operand
is a tuple of arrays, the reduce-scatter is performed on each element of the tuple. -
replica_groups
is a list of replica groups between which the reduction is performed (replica id for the current replica can be retrieved usingReplicaId
). The order of replicas in each group determines the order in which the all-reduce result will be scattered.replica_groups
must either be empty (in which case all replicas belong to a single group), or contain the same number of elements as the number of replicas. When there are more than one replica groups, they all must be of the same size. For example,replica_groups = {0, 2}, {1, 3}
performs reduction between the replicas0
and2
, and1
and3
and then scatters the result. -
shard_count
is the size of each replica group. We need this in cases wherereplica_groups
are empty. Ifreplica_groups
is not empty,shard_count
must be equal to the size of each replica group. -
channel_id
is used for cross-module communication: onlyreduce-scatter
operations with the samechannel_id
can communicate with each other.
The output shape is the input shape with the scatter_dimension
made shard_count
times smaller. For example, if there are two replicas and the operand has the value [1.0, 2.25]
and [3.0, 5.25]
respectively on the two replicas, then the output value from this op where scatter_dim
is 0
will be [4.0]
for the first replica and [7.5]
for the second replica.
ReduceWindow
See also XlaBuilder::ReduceWindow
.
Applies a reduction function to all elements in each window of a sequence of N multi-dimensional arrays, producing a single or a tuple of N multi-dimensional arrays as output. Each output array has the same number of elements as the number of valid positions of the window. A pooling layer can be expressed as a ReduceWindow
. Similar to Reduce
, the applied computation
is always passed the init_values
on the left-hand side.
ReduceWindow(operands..., init_values..., computation, window_dimensions, window_strides, padding)
Arguments | Type | Semantics |
---|---|---|
operands | N XlaOps | A sequence of N multi-dimensional arrays of types T_0,..., T_{N-1} , each representing the base area on which the window is placed. |
init_values | N XlaOps | The N starting values for the reduction, one for each of the N operands. See Reduce for details. |
computation | XlaComputation | Reduction function of type T_0, ..., T_{N-1}, T_0, ..., T_{N-1} -> Collate(T_0, ..., T_{N-1}) , to apply to elements in each window of all the input operands. |
window_dimensions | ArraySlice<int64> | array of integers for window dimension values |
window_strides | ArraySlice<int64> | array of integers for window stride values |
base_dilations | ArraySlice<int64> | array of integers for base dilation values |
window_dilations | ArraySlice<int64> | array of integers for window dilation values |
padding | Padding | padding type for window (Padding::kSame, which pads so as to have the same output shape as input if the stride is 1, or Padding::kValid, which uses no padding and "stops" the window once it no longer fits) |
Where:
- N is required to be greater or equal to 1.
- All input arrays must have the same dimensions.
- If
N = 1
,Collate(T)
isT
. - If
N > 1
,Collate(T_0, ..., T_{N-1})
is a tuple ofN
elements of type(T0,...T{N-1})
.
Below code and figure shows an example of using ReduceWindow
. Input is a matrix of size [4x6] and both window_dimensions and window_stride_dimensions are [2x3].
// Create a computation for the reduction (maximum).
XlaComputation max;
{
XlaBuilder builder(client_, "max");
auto y = builder.Parameter(0, ShapeUtil::MakeShape(F32, {}), "y");
auto x = builder.Parameter(1, ShapeUtil::MakeShape(F32, {}), "x");
builder.Max(y, x);
max = builder.Build().ConsumeValueOrDie();
}
// Create a ReduceWindow computation with the max reduction computation.
XlaBuilder builder(client_, "reduce_window_2x3");
auto shape = ShapeUtil::MakeShape(F32, {4, 6});
auto input = builder.Parameter(0, shape, "input");
builder.ReduceWindow(
input,
/*init_val=*/builder.ConstantLiteral(LiteralUtil::MinValue(F32)),
*max,
/*window_dimensions=*/{2, 3},
/*window_stride_dimensions=*/{2, 3},
Padding::kValid);

Stride of 1 in a dimension specifies that the position of a window in the dimension is 1 element away from its adjacent window. In order to specify that no windows overlap with each other, window_stride_dimensions should be equal to window_dimensions. The figure below illustrates the use of two different stride values. Padding is applied to each dimension of the input and the calculations are the same as though the input came in with the dimensions it has after padding.

For a non-trivial padding example, consider computing reduce-window minimum (initial value is MAX_FLOAT
) with dimension 3
and stride 2
over the input array [10000, 1000, 100, 10, 1]
. Padding kValid
computes minimums over two valid windows: [10000, 1000, 100]
and [100, 10, 1]
, resulting in the output [100, 1]
. Padding kSame
first pads the array so that the shape after the reduce-window would be the same as input for stride one by adding initial elements on both sides, getting [MAX_VALUE, 10000, 1000, 100, 10, 1, MAX_VALUE]
. Running reduce-window over the padded array operates on three windows [MAX_VALUE, 10000, 1000]
, [1000, 100, 10]
, [10, 1, MAX_VALUE]
, and yields [1000, 10, 1]
.
The evaluation order of the reduction function is arbitrary and may be non-deterministic. Therefore, the reduction function should not be overly sensitive to reassociation. See the discussion about associativity in the context of Reduce
for more details.
ReplicaId
See also XlaBuilder::ReplicaId
.
Returns the unique ID (U32 scalar) of the replica.
ReplicaId()
The unique ID of each replica is an unsigned integer in the interval [0, N)
, where N
is the number of replicas. Since all the replicas are running the same program, a ReplicaId()
call in the program will return a different value on each replica.
Reshape
See also XlaBuilder::Reshape
and the Collapse
operation.
Reshapes the dimensions of an array into a new configuration.
Reshape(operand, new_sizes)
Reshape(operand, dimensions, new_sizes)
Arguments | Type | Semantics |
---|---|---|
operand | XlaOp | array of type T |
dimensions | int64 vector | order in which dimensions are collapsed |
new_sizes | int64 vector | vector of sizes of new dimensions |
Conceptually, reshape first flattens an array into a one-dimensional vector of data values, and then refines this vector into a new shape. The input arguments are an arbitrary array of type T, a compile-time-constant vector of dimension indices, and a compile-time-constant vector of dimension sizes for the result. The values in the dimension
vector, if given, must be a permutation of all of T's dimensions; the default if not given is {0, ..., rank - 1}
. The order of the dimensions in dimensions
is from slowest-varying dimension (most major) to fastest-varying dimension (most minor) in the loop nest which collapses the input array into a single dimension. The new_sizes
vector determines the size of the output array. The value at index 0 in new_sizes
is the size of dimension 0, the value at index 1 is the size of dimension 1, and so on. The product of the new_size
dimensions must equal the product of the operand's dimension sizes. When refining the collapsed array into the multidimensional array defined by new_sizes
, the dimensions in new_sizes
are ordered from slowest varying (most major) and to fastest varying (most minor).
For example, let v be an array of 24 elements:
let v = f32[4x2x3] { { {10, 11, 12}, {15, 16, 17} },
{ {20, 21, 22}, {25, 26, 27} },
{ {30, 31, 32}, {35, 36, 37} },
{ {40, 41, 42}, {45, 46, 47} } };
In-order collapse:
let v012_24 = Reshape(v, {0,1,2}, {24});
then v012_24 == f32[24] {10, 11, 12, 15, 16, 17, 20, 21, 22, 25, 26, 27,
30, 31, 32, 35, 36, 37, 40, 41, 42, 45, 46, 47};
let v012_83 = Reshape(v, {0,1,2}, {8,3});
then v012_83 == f32[8x3] { {10, 11, 12}, {15, 16, 17},
{20, 21, 22}, {25, 26, 27},
{30, 31, 32}, {35, 36, 37},
{40, 41, 42}, {45, 46, 47} };
Out-of-order collapse:
let v021_24 = Reshape(v, {1,2,0}, {24});
then v012_24 == f32[24] {10, 20, 30, 40, 11, 21, 31, 41, 12, 22, 32, 42,
15, 25, 35, 45, 16, 26, 36, 46, 17, 27, 37, 47};
let v021_83 = Reshape(v, {1,2,0}, {8,3});
then v021_83 == f32[8x3] { {10, 20, 30}, {40, 11, 21},
{31, 41, 12}, {22, 32, 42},
{15, 25, 35}, {45, 16, 26},
{36, 46, 17}, {27, 37, 47} };
let v021_262 = Reshape(v, {1,2,0}, {2,6,2});
then v021_262 == f32[2x6x2] { { {10, 20}, {30, 40},
{11, 21}, {31, 41},
{12, 22}, {32, 42} },
{ {15, 25}, {35, 45},
{16, 26}, {36, 46},
{17, 27}, {37, 47} } };
As a special case, reshape can transform a single-element array to a scalar and vice versa. For example,
Reshape(f32[1x1] { {5} }, {0,1}, {}) == 5;
Reshape(5, {}, {1,1}) == f32[1x1] { {5} };
Rev (reverse)
See also XlaBuilder::Rev
.
Rev(operand, dimensions)
Arguments | Type | Semantics |
---|---|---|
operand | XlaOp | array of type T |
dimensions | ArraySlice<int64> | dimensions to reverse |
Reverses the order of elements in the operand
array along the specified dimensions
, generating an output array of the same shape. Each element of the operand array at a multidimensional index is stored into the output array at a transformed index. The multidimensional index is transformed by reversing the index in each dimension to be reversed (ie, if a dimension of size N is one of the reversing dimensions, its index i is transformed into N - 1 - i).
One use for the Rev
operation is to reverse the convolution weight array along the two window dimensions during the gradient computation in neural networks.
RngNormal
See also XlaBuilder::RngNormal
.
Constructs an output of a given shape with random numbers generated following the \(N(\mu, \sigma)\) normal distribution. The parameters \(\mu\) and\(\sigma\), and output shape have to have a floating point elemental type. The parameters furthermore have to be scalar valued.
RngNormal(mu, sigma, shape)
Arguments | Type | Semantics |
---|---|---|
mu | XlaOp | Scalar of type T specifying mean of generated numbers |
sigma | XlaOp | Scalar of type T specifying standard deviation of generated numbers |
shape | Shape | Output shape of type T |
RngUniform
See also XlaBuilder::RngUniform
.
Constructs an output of a given shape with random numbers generated following the uniform distribution over the interval \([a,b)\). The parameters and output element type have to be a boolean type, an integral type or a floating point types, and the types have to be consistent. The CPU and GPU backends currently only support F64, F32, F16, BF16, S64, U64, S32 and U32. Furthermore, the parameters need to be scalar valued. If \(b <= a\) the result is implementation-defined.
RngUniform(a, b, shape)
Arguments | Type | Semantics |
---|---|---|
a | XlaOp | Scalar of type T specifying lower limit of interval |
b | XlaOp | Scalar of type T specifying upper limit of interval |
shape | Shape | Output shape of type T |
RngBitGenerator
Generates an output with a given shape filled with uniform random bits using the specified algorithm (or backend default) and returns an updated state (with the same shape as initial state) and the generated random data.
Initial state is the initial state of the current random number generation. It and the required shape and valid values are dependent on the algorithm used.
The output is guaranteed to be a deterministic function of the initial state but it is not guaranteed to be deterministic between backends and different compiler versions.
RngBitGenerator(algorithm, key, shape)
Arguments | Type | Semantics |
---|---|---|
algorithm | RandomAlgorithm | PRNG algorithm to be used. |
initial_state | XlaOp | Initial state for the PRNG algorithm. |
shape | Shape | Output shape for generated data. |
Available values for algorithm
:
rng_default
: Backend specific algorithm with backend specific shape requirements.rng_three_fry
: ThreeFry counter-based PRNG algorithm. Theinitial_state
shape isu64[2]
with arbitrary values. Salmon et al. SC 2011. Parallel random numbers: as easy as 1, 2, 3.rng_philox
: Philox algorithm to generate random numbers in parallel. Theinitial_state
shape isu64[3]
with arbitrary values. Salmon et al. SC 2011. Parallel random numbers: as easy as 1, 2, 3.
Scatter
The XLA scatter operation generates a sequence of results which are the values of the input array operands
, with several slices (at indices specified by scatter_indices
) updated with the sequence of values in updates
using update_computation
.
See also XlaBuilder::Scatter
.
scatter(operands..., scatter_indices, updates..., update_computation, index_vector_dim, update_window_dims, inserted_window_dims, scatter_dims_to_operand_dims)
Arguments | Type | Semantics |
---|---|---|
operands | Sequence of N XlaOp | N arrays of types T_0, ..., T_N to be scattered into. |
scatter_indices | XlaOp | Array containing the starting indices of the slices that must be scattered to. |
updates | Sequence of N XlaOp | N arrays of types T_0, ..., T_N . updates[i] contains the values that must be used for scattering operands[i] . |
update_computation | XlaComputation | Computation to be used for combining the existing values in the input array and the updates during scatter. This computation should be of type T_0, ..., T_N, T_0, ..., T_N -> Collate(T_0, ..., T_N) . |
index_vector_dim | int64 | The dimension in scatter_indices that contains the starting indices. |
update_window_dims | ArraySlice<int64> | The set of dimensions in updates shape that are window dimensions . |
inserted_window_dims | ArraySlice<int64> | The set of window dimensions that must be inserted into updates shape. |
scatter_dims_to_operand_dims | ArraySlice<int64> | A dimensions map from the scatter indices to the operand index space. This array is interpreted as mapping i to scatter_dims_to_operand_dims[i] . It has to be one-to-one and total. |
indices_are_sorted | bool | Whether the indices are guaranteed to be sorted by the caller. |
Where:
- N is required to be greater or equal to 1.
-
operands
[0
], ...,operands
[N-1
] must all have the same dimensions. -
updates
[0
], ...,updates
[N-1
] must all have the same dimensions. - If
N = 1
,Collate(T)
isT
. - If
N > 1
,Collate(T_0, ..., T_N)
is a tuple ofN
elements of typeT
.
If index_vector_dim
is equal to scatter_indices.rank
we implicitly consider scatter_indices
to have a trailing 1
dimension.
We define update_scatter_dims
of type ArraySlice<int64>
as the set of dimensions in updates
shape that are not in update_window_dims
, in ascending order.
The arguments of scatter should follow these constraints:
Each
updates
array must be of rankupdate_window_dims.size + scatter_indices.rank - 1
.Bounds of dimension
i
in eachupdates
array must conform to the following:- If
i
is present inupdate_window_dims
(ie equal toupdate_window_dims
[k
] for somek
), then the bound of dimensioni
inupdates
must not exceed the corresponding bound ofoperand
after accounting for theinserted_window_dims
(ieadjusted_window_bounds
[k
], whereadjusted_window_bounds
contains the bounds ofoperand
with the bounds at indicesinserted_window_dims
removed). - If
i
is present inupdate_scatter_dims
(ie equal toupdate_scatter_dims
[k
] for somek
), then the bound of dimensioni
inupdates
must be equal to the corresponding bound ofscatter_indices
, skippingindex_vector_dim
(iescatter_indices.shape.dims
[k
], ifk
<index_vector_dim
andscatter_indices.shape.dims
[k+1
] otherwise).
- If
update_window_dims
must be in ascending order, not have any repeating dimension numbers, and be in the range[0, updates.rank)
.inserted_window_dims
must be in ascending order, not have any repeating dimension numbers, and be in the range[0, operand.rank)
.operand.rank
must equal the sum ofupdate_window_dims.size
andinserted_window_dims.size
.scatter_dims_to_operand_dims.size
must be equal toscatter_indices
[index_vector_dim
], and its values must be in the range[0, operand.rank)
.
For a given index U
in each updates
array, the corresponding index I
in the corresponding operands
array into which this update has to be applied is computed as follows:
- Let
G
= {U
[k
] fork
inupdate_scatter_dims
}. UseG
to look up an index vectorS
in thescatter_indices
array such thatS
[i
] =scatter_indices
[Combine(G
,i
)] where Combine(A, b) inserts b at positionsindex_vector_dim
into A. - Create an index
S
in
intooperand
usingS
by scatteringS
using thescatter_dims_to_operand_dims
map. More formally:-
S
in
[scatter_dims_to_operand_dims
[k
]] =S
[k
] ifk
<scatter_dims_to_operand_dims.size
. -
S
in
[_
] =0
otherwise.
-
- Create an index
W
in
into eachoperands
array by scattering the indices atupdate_window_dims
inU
according toinserted_window_dims
. More formally:-
W
in
[window_dims_to_operand_dims
(k
)] =U
[k
] ifk
is inupdate_window_dims
, wherewindow_dims_to_operand_dims
is the monotonic function with domain [0
,update_window_dims.size
) and range [0
,operand.rank
) \inserted_window_dims
. (For example, ifupdate_window_dims.size
is4
,operand.rank
is6
, andinserted_window_dims
is {0
,2
} thenwindow_dims_to_operand_dims
is {0
→1
,1
→3
,2
→4
,3
→5
}). -
W
in
[_
] =0
otherwise.
-
-
I
isW
in
+S
in
where + is element-wise addition.
In summary, the scatter operation can be defined as follows.
- Initialize
output
withoperands
, ie for all indicesJ
, for all indicesO
in theoperands
[J
] array:
output
[J
][O
] =operands
[J
][O
] - For every index
U
in theupdates
[J
] array and the corresponding indexO
in theoperand
[J
] array, ifO
is a valid index foroutput
:
(output
[0
][O
], ..., output[
N-1][
O]) =
update_computation(
output[
0][
O], ..., ,
output[
N-1][
O],
updates[
0][
U], ...,
updates[
N-1][
U`])
The order in which updates are applied is non-deterministic. So, when multiple indices in updates
refer to the same index in operands
, the corresponding value in output
will be non-deterministic.
Note that the first parameter that is passed into the update_computation
will always be the current value from the output
array and the second parameter will always be the value from the updates
array. This is important specifically for cases when the update_computation
is not commutative .
If indices_are_sorted
is set to true then XLA can assume that start_indices
are sorted (in ascending start_index_map
order) by the user. If they are not then the semantics is implementation defined.
Informally, the scatter op can be viewed as an inverse of the gather op, ie the scatter op updates the elements in the input that are extracted by the corresponding gather op.
For a detailed informal description and examples, refer to the "Informal Description" section under Gather
.
Select
See also XlaBuilder::Select
.
Constructs an output array from elements of two input arrays, based on the values of a predicate array.
Select(pred, on_true, on_false)
Arguments | Type | Semantics |
---|---|---|
pred | XlaOp | array of type PRED |
on_true | XlaOp | array of type T |
on_false | XlaOp | array of type T |
The arrays on_true
and on_false
must have the same shape. This is also the shape of the output array. The array pred
must have the same dimensionality as on_true
and on_false
, with the PRED
element type.
For each element P
of pred
, the corresponding element of the output array is taken from on_true
if the value of P
is true
, and from on_false
if the value of P
is false
. As a restricted form of broadcasting , pred
can be a scalar of type PRED
. In this case, the output array is taken wholly from on_true
if pred
is true
, and from on_false
if pred
is false
.
Example with non-scalar pred
:
let pred: PRED[4] = {true, false, false, true};
let v1: s32[4] = {1, 2, 3, 4};
let v2: s32[4] = {100, 200, 300, 400};
==>
Select(pred, v1, v2) = s32[4]{1, 200, 300, 4};
Example with scalar pred
:
let pred: PRED = true;
let v1: s32[4] = {1, 2, 3, 4};
let v2: s32[4] = {100, 200, 300, 400};
==>
Select(pred, v1, v2) = s32[4]{1, 2, 3, 4};
Selections between tuples are supported. Tuples are considered to be scalar types for this purpose. If on_true
and on_false
are tuples (which must have the same shape!) then pred
has to be a scalar of type PRED
.
SelectAndScatter
See also XlaBuilder::SelectAndScatter
.
This operation can be considered as a composite operation that first computes ReduceWindow
on the operand
array to select an element from each window, and then scatters the source
array to the indices of the selected elements to construct an output array with the same shape as the operand array. The binary select
function is used to select an element from each window by applying it across each window, and it is called with the property that the first parameter's index vector is lexicographically less than the second parameter's index vector. The select
function returns true
if the first parameter is selected and returns false
if the second parameter is selected, and the function must hold transitivity (ie, if select(a, b)
and select(b, c)
are true
, then select(a, c)
is also true
) so that the selected element does not depend on the order of the elements traversed for a given window.
The function scatter
is applied at each selected index in the output array. It takes two scalar parameters:
- Current value at the selected index in the output array
- The scatter value from
source
that applies to the selected index
It combines the two parameters and returns a scalar value that's used to update the value at the selected index in the output array. Initially, all indices of the output array are set to init_value
.
The output array has the same shape as the operand
array and the source
array must have the same shape as the result of applying a ReduceWindow
operation on the operand
array. SelectAndScatter
can be used to backpropagate the gradient values for a pooling layer in a neural network.
SelectAndScatter(operand, select, window_dimensions, window_strides, padding, source, init_value, scatter)
Arguments | Type | Semantics |
---|---|---|
operand | XlaOp | array of type T over which the windows slide |
select | XlaComputation | binary computation of type T, T -> PRED , to apply to all elements in each window; returns true if the first parameter is selected and returns false if the second parameter is selected |
window_dimensions | ArraySlice<int64> | array of integers for window dimension values |
window_strides | ArraySlice<int64> | array of integers for window stride values |
padding | Padding | padding type for window (Padding::kSame or Padding::kValid) |
source | XlaOp | array of type T with the values to scatter |
init_value | XlaOp | scalar value of type T for the initial value of the output array |
scatter | XlaComputation | binary computation of type T, T -> T , to apply each scatter source element with its destination element |
The figure below shows examples of using SelectAndScatter
, with the select
function computing the maximal value among its parameters. Note that when the windows overlap, as in the figure (2) below, an index of the operand
array may be selected multiple times by different windows. In the figure, the element of value 9 is selected by both of the top windows (blue and red) and the binary addition scatter
function produces the output element of value 8 (2 + 6).

The evaluation order of the scatter
function is arbitrary and may be non-deterministic. Therefore, the scatter
function should not be overly sensitive to reassociation. See the discussion about associativity in the context of Reduce
for more details.
Send
See also XlaBuilder::Send
.
Send(operand, channel_handle)
Arguments | Type | Semantics |
---|---|---|
operand | XlaOp | data to send (array of type T) |
channel_handle | ChannelHandle | unique identifier for each send/recv pair |
Sends the given operand data to a Recv
instruction in another computation that shares the same channel handle. Does not return any data.
Similar to the Recv
operation, the client API of Send
operation represents synchronous communication, and is internally decomposed into 2 HLO instructions ( Send
and SendDone
) to enable asynchronous data transfers. See also HloInstruction::CreateSend
and HloInstruction::CreateSendDone
.
Send(HloInstruction operand, int64 channel_id)
Initiates an asynchronous transfer of the operand to the resources allocated by the Recv
instruction with the same channel id. Returns a context, which is used by a following SendDone
instruction to wait for the completion of the data transfer. The context is a tuple of {operand (shape), request identifier (U32)} and it can only be used by a SendDone
instruction.
SendDone(HloInstruction context)
Given a context created by a Send
instruction, waits for the data transfer to complete. The instruction does not return any data.
Scheduling of channel instructions
The execution order of the 4 instructions for each channel ( Recv
, RecvDone
, Send
, SendDone
) is as below.

-
Recv
happens beforeSend
-
Send
happens beforeRecvDone
-
Recv
happens beforeRecvDone
-
Send
happens beforeSendDone
When the backend compilers generate a linear schedule for each computation that communicates via channel instructions, there must not be cycles across the computations. For example, below schedules lead to deadlocks.

Slice
See also XlaBuilder::Slice
.
Slicing extracts a sub-array from the input array. The sub-array is of the same rank as the input and contains the values inside a bounding box within the input array where the dimensions and indices of the bounding box are given as arguments to the slice operation.
Slice(operand, start_indices, limit_indices, strides)
Arguments | Type | Semantics |
---|---|---|
operand | XlaOp | N dimensional array of type T |
start_indices | ArraySlice<int64> | List of N integers containing the starting indices of the slice for each dimension. Values must be greater than or equal to zero. |
limit_indices | ArraySlice<int64> | List of N integers containing the ending indices (exclusive) for the slice for each dimension. Each value must be greater than or equal to the respective start_indices value for the dimension and less than or equal to the size of the dimension. |
strides | ArraySlice<int64> | List of N integers that decides the input stride of the slice. The slice picks every strides[d] element in dimension d . |
1-dimensional example:
let a = {0.0, 1.0, 2.0, 3.0, 4.0}
Slice(a, {2}, {4}) produces:
{2.0, 3.0}
2-dimensional example:
let b =
{ {0.0, 1.0, 2.0},
{3.0, 4.0, 5.0},
{6.0, 7.0, 8.0},
{9.0, 10.0, 11.0} }
Slice(b, {2, 1}, {4, 3}) produces:
{ { 7.0, 8.0},
{10.0, 11.0} }
Sort
See also XlaBuilder::Sort
.
Sort(operands, comparator, dimension, is_stable)
Arguments | Type | Semantics |
---|---|---|
operands | ArraySlice<XlaOp> | The operands to sort. |
comparator | XlaComputation | The comparator computation to use. |
dimension | int64 | The dimension along which to sort. |
is_stable | bool | Whether stable sorting should be used. |
If only one operand is provided:
If the operand is a rank-1 tensor (an array), the result is a sorted array. If you want to sort the array into ascending order, the comparator should perform a less-than comparison. Formally, after the array is sorted, it holds for all index positions
i, j
withi < j
that eithercomparator(value[i], value[j]) = comparator(value[j], value[i]) = false
orcomparator(value[i], value[j]) = true
.If the operand has higher rank, the operand is sorted along the provided dimension. For example, for a rank-2 tensor (a matrix), a dimension value of
0
will independently sort every column, and a dimension value of1
will independently sort each row. If no dimension number is provided, then the last dimension is chosen by default. For the dimension which is sorted, the same sorting order applies as in the rank-1 case.
If n > 1
operands are provided:
All
n
operands must be tensors with the same dimensions. The element types of the tensors may be different.All operands are sorted together, not individually. Conceptually the operands are treated as a tuple. When checking whether the elements of each operand at index positions
i
andj
need to be swapped, the comparator is called with2 * n
scalar parameters, where parameter2 * k
corresponds to the value at positioni
from thek-th
operand, and parameter2 * k + 1
corresponds to the value at positionj
from thek-th
operand. Usually, the comparator would thus compare parameters2 * k
and2 * k + 1
with each other and possibly use other parameter pairs as tie breakers.The result is a tuple that consists of the operands in sorted order (along the provided dimension, as above). The
i-th
operand of the tuple corresponds to thei-th
operand of Sort.
For example, if there are three operands operand0 = [3, 1]
, operand1 = [42, 50]
, operand2 = [-3.0, 1.1]
, and the comparator compares only the values of operand0
with less-than, then the output of the sort is the tuple ([1, 3], [50, 42], [1.1, -3.0])
.
If is_stable
is set to true, the sort is guaranteed to be stable, that is, if there are elements which are considered to be equal by the comparator, the relative order of the equal values is preserved. By default, is_stable
is set to false.
Transpose
See also the tf.reshape
operation.
Transpose(operand)
Arguments | Type | Semantics |
---|---|---|
operand | XlaOp | The operand to transpose. |
permutation | ArraySlice<int64> | How to permute the dimensions. |
Permutes the operand dimensions with the given permutation, so ∀ i . 0 ≤ i < rank ⇒ input_dimensions[permutation[i]] = output_dimensions[i]
.
This is the same as Reshape(operand, permutation, Permute(permutation, operand.shape.dimensions)).
TriangularSolve
See also XlaBuilder::TriangularSolve
.
Solves systems of linear equations with lower or upper triangular coefficient matrices by forward- or back-substitution. Broadcasting along leading dimensions, this routine solves one of the matrix systems op(a) * x = b
, or x * op(a) = b
, for the variable x
, given a
and b
, where op(a)
is either op(a) = a
, or op(a) = Transpose(a)
, or op(a) = Conj(Transpose(a))
.
TriangularSolve(a, b, left_side, lower, unit_diagonal, transpose_a)
Arguments | Type | Semantics |
---|---|---|
a | XlaOp | a rank > 2 array of a complex or floating-point type with shape [..., M, M] . |
b | XlaOp | a rank > 2 array of the same type with shape [..., M, K] if left_side is true, [..., K, M] otherwise. |
left_side | bool | indicates whether to solve a system of the form op(a) * x = b ( true ) or x * op(a) = b ( false ). |
lower | bool | whether to use the upper or lower triangle of a . |
unit_diagonal | bool | if true , the diagonal elements of a are assumed to be 1 and not accessed. |
transpose_a | Transpose | whether to use a as is, transpose it or take its conjugate transpose. |
Input data is read only from the lower/upper triangle of a
, depending on the value of lower
. Values from the other triangle are ignored. Output data is returned in the same triangle; the values in the other triangle are implementation-defined and may be anything.
If the rank of a
and b
are greater than 2, they are treated as batches of matrices, where all except the minor 2 dimensions are batch dimensions. a
and b
must have equal batch dimensions.
Tuple
See also XlaBuilder::Tuple
.
A tuple containing a variable number of data handles, each of which has its own shape.
This is analogous to std::tuple
in C++. Conceptually:
let v: f32[10] = f32[10]{0, 1, 2, 3, 4, 5, 6, 7, 8, 9};
let s: s32 = 5;
let t: (f32[10], s32) = tuple(v, s);
Tuples can be deconstructed (accessed) via the GetTupleElement
operation.
While
See also XlaBuilder::While
.
While(condition, body, init)
Arguments | Type | Semantics |
---|---|---|
condition | XlaComputation | XlaComputation of type T -> PRED which defines the termination condition of the loop. |
body | XlaComputation | XlaComputation of type T -> T which defines the body of the loop. |
init | T | Initial value for the parameter of condition and body . |
Sequentially executes the body
until the condition
fails. This is similar to a typical while loop in many other languages except for the differences and restrictions listed below.
- A
While
node returns a value of typeT
, which is the result from the last execution of thebody
. - The shape of the type
T
is statically determined and must be the same across all iterations.
The T parameters of the computations are initialized with the init
value in the first iteration and are automatically updated to the new result from body
in each subsequent iteration.
One main use case of the While
node is to implement the repeated execution of training in neural networks. Simplified pseudocode is shown below with a graph that represents the computation. The code can be found in while_test.cc
. The type T
in this example is a Tuple
consisting of an int32
for the iteration count and a vector[10]
for the accumulator. For 1000 iterations, the loop keeps adding a constant vector to the accumulator.
// Pseudocode for the computation.
init = {0, zero_vector[10]} // Tuple of int32 and float[10].
result = init;
while (result(0) < 1000) {
iteration = result(0) + 1;
new_vector = result(1) + constant_vector[10];
result = {iteration, new_vector};
}
