Changes to the tiled-convolution branch #2
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This option has existed for some time, but was never explicitly written in the main CMake file. While adding this, some old unused lines were removed from the CMake (the directory has long not been set by FindComputeCpp.cmake, so the lines were totally useless).
The current behaviour of `get_node` in the virtual pointer mapper will use `lower_bound` to get the buffer which is not less than the input pointer. If the pointer mapper has only allocated one buffer, and we are looking up a pointer offset into that buffer then the call to `lower_bound` will return an iterator pointing past the end of the map. This means we cannot dereference this iterator, as it does not point into the map, but we still want to decrement this iterator to get the buffer that the virtual pointer is referencing. By adding a check for whether the iterator is the end we can short circuit the iterator dereference, and so avoid this problem without otherwise changing the behaviour of the pointer mapper.
…_vptr This commit removes the unnecessary usage of the buffer allocator in …
The type buffer_t is defined in the class PointerMapper and redefined to a different type in get_buffer which results in shadowing warnings. SYCLmalloc doesn't have this issue but the alias should probably be changed for consistency.
Property lists are useful, and we still want to pass them to buffers through the SYCL malloc interface. In addition, a related test was failing because of an interface change, and has been removed.
Contains improvements designed to improve the reliablity of FindComputeCpp.cmake, brought on by the MSVC policy of supporting only two Clang versions at any given time. * Added way to check MSVC STL/Clang compatibility * Fix and simplify force-include for MSVC * Ensure device compiler arguments are always stored as CMake list
Small sample showing how to use built-in kernels in ComputeCpp.
Fixes issue codeplaysoftware#126 but allows for optional ComputeCpp use.
Specifically this shows a device selector that fails to find a device. Similar effect could be achieved with the fallback queue. This should fix codeplaysoftware#105.
Moving these compiler checks to a separate module keeps the main FindComputeCpp.cmake clean and makes the checks "opt-in", if you know that you won't need the check for your project.
* Use `handler::copy` instead of copying in kernel * Initialize output data to zero * Only create single temporary buffer object * Use on-chip memory for temporary buffer objects * Made `compute_index` a single function * Added existing sample from master * As `tiled-convolution-standard` * `input_data_info` * Changed OpenCL parameters * Addressed many PR comments * `constexpr matrix_size_t` * Typo fixes and formatting * Validate returns `bool` * No need to specifiy global buffer target * CMake cleanup * Display profiling
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
9 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
handler::copyinstead of copying in kernelcompute_indexa single functiontiled-convolution-standardinput_data_infoconstexpr matrix_size_tbool