GitLab

Reduce memory footprint by std::move of temporary (potentially very large) containers.
Minor cleanup for performance optimization: e.g. of Index() calculation -- which can get repeated millions of times in large tensors/vectors in a single Visit.

Throwing on these calls catches dynamic shape errors earlier rather than having to backpedal from a bad call

Python API with documentation updates

Solves #1311

Examples..

bin/driver verify /codes/onnx_models/resnet50-v1-7/resnet50-v1-7.onnx --split-single-dyn-dim --batch 3 --dyn-input-dim @data "[{min:1, max:4}, 3, 224, 224]"

bin/driver compile /codes/onnx_models/resnet50-v1-7/resnet50-v1-7.onnx --split-single-dyn-dim --default-dyn-dim "{min:1, max:10}" --output resnet50_batch1-10.mxr

bin/driver perf resnet50_batch1-10.mxr --batch 4

Makes the optimals into a std::set<std::size_t>
Changes shape object functions to handle the opts change
Changes to convolution, flatten, pooling, and convolution in that they no longer calculate the output optimal dimensions. Instead returns empty opts. Will need to change this in the future if we want to support dynamic shapes fully.
Many changes to tests and shape calls with respect to the new optimals

Creates the select_module operator that selects one of the submodules passed to it to run based on the submodule parameters.  The submodule is selected by having the exact same static shapes for the arguments to select_module as the parameters in the submodule

Extends pad operator to handle dynamic input shapes
Only handles computing the shape for adding constant padding to a dynamic shape
- adds the padding to the min, max, and opt values (unless opt is 0, where it keeps it 0)
- does not handle reflect padding with dynamic shapes

Implements the operator==(dynamic_dimension, size_t) functions

Updated Multibroadcast op to have a two input version for dynamic shapes
Current dynamic shape broadcasting logic
dynamic_dimensions must be the same or one of them is {1, 1, 0} or {1, 1, 1}
Works for dyn-dyn, dyn-static, and static-static shape combinations
Changed common.cpp for multibroadcasting for binary ops with dynamic shapes
Extended binary.hpp for dynamic shapes to test the new common.cpp stuff

Using not and or improves readability. The cppcheck rule will help ensure we are doing it consistently.

Initial sketch for changes to shape to handle dynamic dimensions

* Jit contiguous

Updated each source file in the repo with the existing license.

Add with_type to shape class

* Add definitions for all pointwise operators

* Formatting

* Add cpp generator class

* Formatting

* Move compilation to core

* Formatting

* Add clock to tmp name

* Add dynamic loader

* Formatting

* Add tests for code gen

* Formatting

* Add test for literals

* Formatting

* Use with_char

* Add missing header

* Fix mismerge

* Ignore tidy warning

* Fxx gcc 5 errors

* Apply fixits

* Skip signed bitwise of status

* Remove unused parameters

* Explicitly add c++14 flag

* Fix tidy warning

* Add tuple type to shape class

* Formatting

* Make data member private

* Formatting

* Add sub arguments

* Formatting

* Trun clang format off

* Disable clang-format

* Improve visiting tuples

* Formatting

* Add more argument tests

* Formatting

* Handle tuple in load

* Formatting

* Remove .o files

* Add tuple type to api

* Formatting

* Fix tidy warnings

* Fix tidy warnings

* Add a test for share method

* Formatting

* Add a test cpp_type

* Suppress tidy warning
Co-authored-by: Shucai Xiao <Shucai.Xiao@amd.com>

* Add eliminate_data_type pass

* Formatting

* Auto convert quant ops

* Formatting

* Flip the order of decompose

* Compute max size differently

* Formatting

* Clamp values in convert

* Formatting

* Fix loss of precision in reduce

* Formatting

* Fix bugs in reduction

* Fix accumulator type in reference softmax implementation

* Formatting

* Update convert test

* Remove unused variables

* Remove unnecessary quant_dot check

* Formatting

* Add tests

* Formatting

* Remove unused code

* Remove duplicate ops

* Remove blaze dependency

* Use set since shape::type_t is no hashable on gcc 5

* Formatting

* Add dnnl binary op

* Formatting

* Add binary and eltwise

* Formatting

* Add softmax

* Formatting

* Remove unused operators

* Add missing files

* Formatting

* Add lrn

* Formatting

* Add deconvolution

* Formatting

* Change allocate default

* Add reorder

* Formatting

* Add reductions

* Formatting

* Sort lines

* Change literals in another loop

* Add pow operator

* Formatting

* Add pow operator

* Formatting

* Make sure shapes are packed

* Allow broadcasted inputs

* Remove unused operators

* Simplify functions

* Remove softmax

* Add sub and erf functions

* Formatting

* Fix bug

* Formatting

* Improve parallism

* Formatting

* Allow multiple batch dimensions

* Formatting

* Move literal transforms out of lowering

* Formatting

* Add gather operator

* Sort lines

* Add early exit for carry

* Formatting

* Add missing concat

* Rename macro

* Fix deep nesting

* Formatting

* Fix cppcheck issues

* Remov else

* Move attribute to typedef

* Formatting

* Disable maybe-uninitialized warning since its broken on gcc

* Add constexpr default constructor

* Formatting

* Fix compiler warnings

* Fix adjust_allocation test

* Add layernorm matcher

* Add gelu_erf matcher

* Formatting

* Add gelu_tanh matcher

* Formatting

* Remove match namespace

* Formatting

* Use matcher instead of string

* Formatting

* Add fusions

* Formatting

* Add post op field

* Formatting

* Make post_ops serializable

* Formatting

* Add eltwise fusions

* Formatting

* Fix null conversions

* Formatting

* Add fuse_ops source files

* Formatting

* Set binary post op index correctly

* Formatting

* Fix serialization bugs

* Check if used once

* Formatting

* Fix error in get_primitive_attr

* Formatting

* Add compile function

* Formatting

* Limit fusions

* Formatting

* Disable with env variable instead of using compile arg

* Formatting

* Fix implicit conversion to bool

* Declar on seperate lines

* Formatting

* Fix cppcheck issues

* Fix ICE in pack_join

* Formatting

* Use const ref

* Make enum hashable

* Formatting

* Add explicit this

* Fix merge issues

* Fix dangling ref

* Formatting

* Add test for compile

* Formatting

* Add more value tests

* Formatting
Co-authored-by: Shucai Xiao <shucai@gmail.com>
Co-authored-by: mvermeulen <5479696+mvermeulen@users.noreply.github.com>

* Add definitions for all pointwise operators

* Formatting

* Add cpp generator class

* Formatting

* Move compilation to core

* Formatting

* Add clock to tmp name

* Add dynamic loader

* Formatting

* Add tests for code gen

* Formatting

* Add test for literals

* Formatting

* Use with_char

* Add missing header

* Fix mismerge

* Ignore tidy warning

* Fxx gcc 5 errors

* Apply fixits

* Skip signed bitwise of status

* Remove unused parameters

* Explicitly add c++14 flag

* Fix tidy warning

* Remove .o files

* Add method to compute shape with same layout

* Formatting

* Fix permutation with ambiguous layouts

* Formatting

* Propagate layout for pointwise operators

* Formatting

* Propagate layout for more operators

* Formatting

* Sort with lens

* Formatting

* Simplify permutation sorting

* Formatting

* Propagate layout for concat operator

* Formatting

* Use copy

* Formatting

* Remove header
Co-authored-by: mvermeulen <5479696+mvermeulen@users.noreply.github.com>

* Add eliminate_data_type pass

* Formatting

* Auto convert quant ops

* Formatting

* Flip the order of decompose

* Compute max size differently

* Formatting

* Clamp values in convert

* Formatting

* Fix loss of precision in reduce

* Formatting

* Fix bugs in reduction

* Fix accumulator type in reference softmax implementation

* Formatting

* Update convert test

* Remove unused variables

* Remove unnecessary quant_dot check

* Formatting

* Add tests

* Formatting

* Remove unused code

* Remove duplicate ops

* Remove blaze dependency

* Use set since shape::type_t is no hashable on gcc 5

* Formatting
Co-authored-by: Shucai Xiao <shucai@gmail.com>
Co-authored-by: mvermeulen <5479696+mvermeulen@users.noreply.github.com>

* Add flag to enable cpu backend

* Make buffers shared

* Enable optimizations

* Add onednn

* Formatting

* Formatting

* Add dnnl header

* Formatting

* Rewrite rnn first

* Formatting

* Call reference implementation

* Formatting

* Make literal data shared

* Formatting

* Add convolution

* Formatting

* Compensate for dilation

* Formatting

* Use name/make_op instead

* Formatting

* Rename gemm header

* Formatting

* Add dnnl convolution/gemm operators

* Formatting

* Add eliminate_contiguous

* Add faster pointwise operators

* Formatting

* Formatting

* Formatting

* Add dnnl op class

* Formatting

* Add add op

* Formatting

* Add concat operator

* Formatting

* Add more ops

* Create descriptor during finalization

* Formatting

* Dont rewrite pooling

* Enable memory coloring

* Formatting

* Add output aliases

* Formatting

* Fix errors

* Formatting

* Convert literals

* Add missing file

* Remove batch_norm

* Formatting

* Use strides

* Formatting

* Add some debug checks

* Formatting

* Fix big in adjusting shape for gemm

* Formatting

* Fix fallback dot operator

* Zero initialize buffers

* Add suport for group convolutions

* Formatting

* Make adjust allocation target independent

* Formatting

* Enable adjust_allocation for gpu/cpu

* Formatting

* Add copy to allocation model

* Formatting

* Add copy operator

* Formatting

* Better handling of output parameters in adjust_allocation

* Formatting

* Build with dnnl

* Make dnnl required

* Fix compile error

* Tidy fixes

* Formatting

* Tidy fixes

* Formatting

* Fix more tidy issues

* Formatting

* Add mul op

* Add mul op

* Set c compiler to clang as well

* Compensate for normalized compute shape

* Formatting

* Fix cppcheck errors

* Formatting

* Add onednn library to hcc

* Guard clang pragmas

* Disable cpu mode for gcc for now

* Leave it enabled it for gcc 7

* Fix cppcheck suppresion

* Fix compile error on gcc 5

* Remove unused code
Co-authored-by: Shucai Xiao <shucai.xiao@amd.com>
Co-authored-by: mvermeulen <5479696+mvermeulen@users.noreply.github.com>

* Make global variables const

* Tidy fixes

* Disable some lints

* Formatting

* Fix tidy const

* Formatting

* Add missing const keywords

* Formatting

* More fixes

* Fix remaining tidy issues

* Formatting

* Fix rocblas function call

* Formatting

* Fix nodiscard warnings

* Formatting

* Use named parameters

* Remove overload

* Add overload

* Remove noncps

* Use named param for node

* Add auto register header

* Use named parameters

* Refactor jenkinsfile

* Fix shadow

* Add missing body variable

* Add more const methods

* Add hip-clang docker builds

* Remove comments

* Add clang-format

* Add more const

* Formatting

* Rename stage

* Disable check

* Add another const

* Add python 2 dev packages

* Add sphinx to dockerfile

* Add initial serialization

* Formatting

* Add unit tests

* Formatting

* Add tests for serialization

* Formatting

* Use or not and

* Add value test

* Formatting

* Add more tests

* Add shape serialization

* Formatting

* Add serializtion for literal and argument

* Formatting

* Serialize empty types

* Formatting

* Tidy fixes

* Formatting

* Fix tidy issues

* Formatting

* Reformat value type macro

* Formatting

* Handle enum types

* Formatting

* Add float_equal

* Fix tidy issue

* Use declval for better sfinae

* Formatting

* Fix maro name

* Add more test cases to improve coverage

* Formatting

* Add more tests

* Formatting

* Fix assertion

* Fix bug with keyless assignment

* Formatting
Co-authored-by: mvermeulen <5479696+mvermeulen@users.noreply.github.com>

* Add decompose pass

* Add decompose test

* Formatting

* Add remap

* Formatting

* Add compute method for dot

* Formatting

* Add finder for horizontal fusion

* Formatting

* Formatting

* Reuse predicate

* Add gemm fusions

* Formatting

* Add some fixes for convolution

* Formatting

* Fix shape tests

* Formatting

* Reuse axis equal

* Add initial split fusion

* Formatting

* Update offset

* Workaround outputs that cant accept nonstandard shapes

* Formatting

* Add check for split concat

* Formatting

* Add missing headers

* Formatting

* Add tests

* Formatting

* Add more testing

* Formatting

* Fix when there is duplicate splits in inputs

* Formatting

* Fix mismatch iterators

* Add tests for dot fusions

* Formatting

* Add test for convolution

* Formatting

* Fix tidy issues

* Add more tests

* Formatting

* Ignore build directory for codecov

* Add test for groups

* Formatting

* Add more tests for groups

* Formatting

* Add test for missing end slice

* Add newline

* Remove unused function

* Add support for when beta is not 1

* Formatting

* Add test for scalar

* Add one more scalar test
Co-authored-by: mvermeulen <5479696+mvermeulen@users.noreply.github.com>

* Make contiguous preserve scalar shape

* Formatting

* Improve standard shape calulation

* Formatting

* Enable some optimizations on debug build

* Up optimization level

* Remove debug symbols
Co-authored-by: mvermeulen <5479696+mvermeulen@users.noreply.github.com>

* Fix bug in bert accuraccy

* Formatting

* add another test

* Fix add and overflow

* Formatting

* Fix bug in shape_for_each

* Use front instead of iterator

* Use result.front()

* Split add_unary files

* Formatting

* Fix incorrect last index

* Remove comment

* Inline function

* Fix carry check

* Fix metadata errors

* Formatting

* Reflow

* Reflow