GitLab

Local Threads of multiples 32 were introduced in #1348
But LocalThreads that are not multiple of 64 are causing correctness issues.

Upgraded Dockerfiles and fixed tidy issues to make Ubuntu 20.04 and ROCm 5.3.0 the default 

updated GPU pad to now use JIT version.
added range functions for JIT kernels.

Fixes an observed regression error on certain Frozen Protobuf models due to PR 1280

Reiterate the assertion on the standard shape but relax it for the multibroadcast ops deliberately inserted to explicit the broadcast.

Refactor dynamic compute
- add a compute_output_shape object that implicitly converts to a new dyn_output or shape object
- dyn_output object can handle computing the static output shape of an operator given the input arguments shapes
  change an operator's compute function to argument compute(const dyn_output& dyn_out, std::vector<argument> args) to 
  use dyn_output object

Dynamic ref unary functions
-  Included these changes to have an example of the refactored dynamic compute being used
-  Changes to unary base class to handle dynamic shapes
-  Changed elu and leaky_relu to use unary base class and pointwise JIT

* use find2.0 for the convolution
Co-authored-by: Vasilii Filippov <DrizztDoUrden@users.noreply.github.com>
Co-authored-by: Chris Austen <causten@users.noreply.github.com>

* Enable non-standard shape
* Use perfdb for non xdlops
* Fix transpose+broadcast strides
Co-authored-by: jungpark-mlir <jungwook.park@amd.com>

Removes use_dynamic_same_auto_pad
Change padding_mode to be used for dynamic padding
Move compute_padded_shape to pad_calc.cpp as it will be used in other dynamic padding cases
Fix same_lower compute_padded_shape bug and add a test.

Rewrites the TF batch norm like operators to other MIGX operators
Removes the code related to batch_norm_inference

Stream sync changes and associated API level changes

optimize the softmax operator

Adds two methods for the custom_ops virtual class.

bool runs_on_offload_target(), if the custom op runs directly on the gpu then it should be set to true. in this case, custom op expects its parameters to reside in GPU memory and writes output to the GPU memory. If it is set to false then, custom op expects it's parameter to reside on the host and puts back the result into the host memory.

output_alias, if output of the custom op is aliasing the input buffer. i.e. interpreting the same input buffer with differnet shape and strides.

Update as_vector() in C++ API to handle non-standard shapes. It required exposing element_index to space_index conversion method for the shape class.

Improvements/Additions to be made:

changes for the quant_convolution,
changes for the deconvolution,
Macros for MIOpen status checks

test_gpu_pack_int8_args fails on gfx908 machine, because it doesn't set compute_fp32 flag correctly. This PR fixes the test such that it checks for the device-name, and rocblas-versions and sets this flag accordingly.

Implement operator for CPU and GPU implementations

* Remove device functions
* Update tests

This PR allows for other values of epsilon to be matched when finding layernorm. Similarly, the calculation now uses the variable for epsilon.

Compute mean and variance in same reduction
Set block size to numbers divisible by 32 instead powers of 2
Global is also set exactly instead of being divisible by block size
More exact matching of global/local can help get rid of branching/loops
Reduce vectors first before doing dpp_reduce
Explicitly vectorize array operators since the compiler doesnt always vectorize them
Still uses old for loop when its computing at compile-time since the reinterpret_cast nor the all the vector types is supported

* fix typo for add_sigmoid

* Replaced `find_library` with `find_package` to locate MLIR static library
* Unified the include dir for headers and remove backward compatibility
* Embedded the external/include dir into the exported library

* Implement concat using jit compilation

Improves performance for 4/6 GEMMs used by huggingface BERT models with batch_size>1 by using a non-batched rocBLAS call for GEMMs where the B input has a broadcasted batch dimension.
The four verify tests added reflect the actual configurations used by bert-base-cased, with varied batch sizes.

Also adds a matcher to simplify_reshapes to move multibroadcasts after concats.

* Remove unused headers

* Fix accuracy bug when vectorizing slices

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

Rewrite_gelu pass replaces the gelu formula of x * (1/2) * (1 + erf(x/sqrt(2))) with the sigmoid approximation of x * Sigmoid(x * 1.702)

* Track kernel time

This will rewrite dot operators like X(Y + b) to XY + Xb when b is constant as we can fold the add away.
This improves handling pointwise with broadcasted operators, this helps improves const propagation.
Improve gemm fusion with a mul_add
Improve support for broadcast shapes in gemm

* Update is_supported
* Return object from is_supported
* Return by reference in interator

remove print from source

Once
https://github.com/ROCmSoftwarePlatform/llvm-project-mlir/pull/690
lands, the ABI for MLIR-generated kernels will change. This commit
prepares MIGraphX for the change by conditionally selecting the new
ABI if MLIR reports a sufficiently high API version in its headers.

* Dynamic shape handling in shape object

* rewrite empty lens multibroadcast test

* Shape class changes to handle dynamic
* More throw errors for functions that don't make sense for dynamic shape
* Print output changes
* Serialization changes

* Fixing serialization errors

* Remove const on dyn_dim copy getters

* Dynamic shape tests

* Fix serialize errors

* Add dyn_data struct to avoid ambiguous constructor

* Tidy fix: emplace_back() over for loop

* Tidy fix: use move

* Use std::initializer_list in constructor
Reverts the dyn_data struct change
Should get around the ambiguous braced initialization list error

* avoid typedef

* element_space, min,max,opt _lens change

* formatting

* Comments fix

* dynamic bytes() test

* Seralize and reflect changes

* formatting

* Test the dynamic lens functions

* progress

* Formatting

* Dynamic conv draft progress

* Add operator<< tests for coverage

* Coverage update

* Add to conv dynamic batch test

* Dynamic image size test

* Dynamic weight handling

* Dyn image shape test change, fix dyn weight cond

* Comment update

* Dynamic weights shape test and fix

* Use ternary operator

* Tidy fixes

* Handle dynamic graph input shapes in ONNX parser

* Formatting

* Handle dynamic shape for convolution

* formatting

* cppcheck fixes

* Add onnx test files

* Fix typo

* Disable auto_pad for dynamic input shape

* check_shapes object checks for allowing dynamic shapes

* Fix any_of

* Change to maintain const objectness

* Formatting

* Check shapes allow dynamic

* Refactor compute_shape() call into op.compute()
Allows for per operator differences with handling dynamic shape
Fix operation.hpp change to use the generator

* Comment fix

* Refactor normalize_attributes() calls to use max_lens()

* Comment addition

* Update other normalize_attributes() calls

* Change to using constructor and add tests

* Use const member function

* Add more dynamic shape support

* Add tests for error code coverage

* Fix opt shape bug and add shape tests

* capture all by ref

* Fix typo with img shape calculation

* Add more tests

* dynamic auto pad attempt
Linker error with pad_calc.cpp

* Fix parse dyn auto_pad
Should only need to use dynamic auto pad when the image shape or kernel
shape are dynamic. For a dynamic batch size, the auto pad calculation is
the same.

* Fix linking error

* Fix auto_pad bug
Fixed input tensor with auto_pad setting on

* auto_pad onnx tests

* Fix auto_pad calculation, evaluate in ref_conv
add ref_ops tests

* Add shape tests, fix bugs

* Refactor first two output dynamic len calculation

* Conv MLIR test update

* i64 MLIR test fix

* Fix MLIR test typo
Co-authored-by: Chris Austen <causten@users.noreply.github.com>

Currently, while copying a host buffer to the device, it first registers/maps the host buffer pointer to address space of the device.

If the host buffer has been allocated by the hipHostMalloc then, it is implicitly registered to the device's address space, and no need to register again. This PR adds a check for the same.

* Add in changes for onnx Mod operator

Initial operator for mod implementation and test cases for integer and floating based types.

Need to use fmod from stdlib for floating point types. half_float::half thankfully is specced to the use the existing std::fmod() call when looking at the half.hpp implementation.

fmod_flag should mirror the onnx fmod attribute. Right now using a floating point type without setting that on the user side to true will result in an exception.

Ref ticket #1283 

* Add is_supported to the target
* Add get_target_assignments
* Rename assignment to target_assignments
* Add ref target header to test
* Add fpga target
* Make context const in compute