OSX #610
Merged
OSX #610
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
Not every device allows for local work group size to be greater than [1;1;1].
Tests for inplace_reduce, radix_sort and radix_sort_by_key are disabled for CPU devices on Apple platform because for CPU devices on Apple platform when local memory is used in a kernel, local work group size must be [1;1;1]. Those algorithms are designed for GPU devices, therefore we do no lose any functionality.
This commit adds asserts that check if used template type is correct (is integral, is a floating point type).
From tests on Travis-CI it seems that on Apple OpenCL platform clCreatBuffer() does not return NULL and does not set error to CL_INVALID_BUFFER_SIZE when size of the buffer memory object is greater than CL_DEVICE_MAX_MEM_ALLOC_SIZE.
| : m_n(std::distance(first, last)), | ||
| m_probabilities(std::distance(first, last)) | ||
| : m_n((std::max)(size_t(1), static_cast<size_t>(std::distance(first, last)))), | ||
| m_probabilities((std::max)(size_t(1), static_cast<size_t>(std::distance(first, last)))) |
There was a problem hiding this comment.
I think we still need to handle the case where first == last better. Looking at the documentation for std::discrete_distribution, when the input range is empty (or calling the default constructor), we should just initialize with a single probability of 1. But this is currently not implemented, and we don't have to do it in this pull request. I'll open a separate bug for this.
|
Thanks for all the fixes! |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 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Outdated suggestions cannot be applied.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.
This PR resolves issues with Boost.Compute running on OSX.
Tests for
detail::inplace_reduce,radix_sort,radix_sort_by_keyand matrix transpose example are disabled for CPU devices on Apple platform because for CPU devices on Apple platform when local memory is used in a kernel local work group size must be [1;1;1]. Those algorithms and matrix example are designed for GPUs, so from practical perspective we do not lose any functionality.I've also fixed
boost::compute::discrete_distributionconstructor. There was a read from uninitialized memorym_probabilities[i-1](withi = 0).I've found out that there's a bug in
clCreateBuffer()on Apple platform.clCreatBuffer()does not returnNULLand does not set error toCL_INVALID_BUFFER_SIZEwhen size of the buffer memory object is greater thanCL_DEVICE_MAX_MEM_ALLOC_SIZE. You can see bug example here - 42949672960 is the size of allocated memory whereCL_DEVICE_MAX_MEM_ALLOC_SIZEis only 1073741824 bytes.After this PR, Travis-CI OSX build is not allowed to fail.