GitLab

Handles all 4 Fp8 dtypes listed here : https://onnx.ai/onnx/technical/float8.html
Follows saturation/clipping logic from table there as well : https://onnx.ai/onnx/technical/float8.html#cast
Only adding fp8e4m3fnuz in MIGraphX IR for now.

The inaccuracy was caused by ONNX round requires nearest integer rounding for halway (0.5) cases.
std::round rounds away from zero, thus giving wrong results with halfway cases.
Replaced std::round with std::nearbyint which uses the correct rounding by default.

This will enable MLIR by default for these cases:

Any convolution fusion
Any int8 gemm fusion
All Navi3 standalone convolutions
With a flag(ie MIGRAPHX_ENABLE_MLIR) to enable MLIR for floating-point gemm fusions
Except:

3x3 winnograd convolutions fusions (except on Navi)
K > 2048 on gemm (as CK)
Also there is MIGRAPHX_DISABLE_MLIR to disable MLIR completely.

MIGraphX verification by default uses normalized RMS error as the basis for the verification.  This change adds some logic to allow migraphx to do "np.allclose" type of elementwise verification using atol and rtol.

Commit also includes changes to consistently pass "gold" or "expected" results as the second argument for "verify_range()" calls.  Default RMS tolerance inside driver is set to 0.001 which IMO is high for FP32 compared to what we had earlier. Need better defaults

This PR constitutes the MIGraphX-side changes needed to not break the build in the presence of
ROCmSoftwarePlatform/rocMLIR#1136 , and updates what data is sent in to MLIR during the kernel generation and tuning process.

* Add initial tuning support

* Format

* Add extra param

* Format

* Use exauhstive flag

* Format

* Set expected shapes

* Format

* Format

* Fix missing symbol

* Format

* Add missing license header

* Format

* Update src/targets/gpu/include/migraphx/gpu/mlir.hpp

Fixes the failing test case in #1815. Added a test that would otherwise fail without the change.

Bump MLIR commit to include latest supported pointwise ops.
Expand the MLIR approve list 
Ensure that operations such as tanh() that don't have integer implementations (at least in MLIR) aren't used within MLIR modules.
Add additional tests.

Add mlir quant_dot operator support

Co-authored-by: Paul Fultz II <pfultz2@yahoo.com>

Move CI to support the rocm5.5 release

Exposed the mlir_enabled() call the decide for lowering pipeline's enablement
Disabled the rewrite quantization pipeline in mlir compilation
Added quant convolution as anchor ops
Fixed the return type expectations
Added the fall back hip implementation for quantizelinear and dequantizelinear
Will need advises to improve the implementation for quantizelinear

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

* [MLIR] add dot offloads with manual tuning support
* This commit adds dot + pointwise fusion support
along with manual tuning using rocMLIR.

Fixes #1595

Added CMakeFlag for hipRTC. MIGRAPHX_USE_HIPRTC.
Added stages in Jenkins for hipRTC.
Fixes for some of the pending issues from hipRTC.

Need this for when we debug and use MIGRAPHX_TRACE_EVAL() to show tuples
Without this we break when reading our buffer due to the use of visit()
This came up as part of #1283 debugging.

Update dialect registration interface
Update 2nd build pipeline call and use full arch name

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

* 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.

Stream sync changes and associated API level changes

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

* 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

* 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.