Partial math function support in in the hemi:: namespace and some tests #10
Conversation
This file contains 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
… the appropriate single or double-precision function. Switching is done based on HEMI_DEV_CODE preprocessor statement.
…m brev. Unsure if fma actually works or is just a infinite loop trap on host.
compilers. Changed test makefile to look like examples' Started test harness for math.h functions in hemi. Double-precision round() on the device is not passing an assert.
…++11 in GCC-4.8 it can come back.
|
Please make some changes.
|
|
|
||
| // templated POW function. For a CUDA device, it casts the arguments to double. | ||
| // For a host, it defaults to std::pow in <cmath> | ||
| template <class T> |
Moved { } to their own lines.
Removed hemi::fma for now, no c++11 support in nvcc (at least experimentally the compiler is choking on STL).
|
Submitting a new pull request. |
|
Github is confusing. Sorry for the extra noise. |
…orted. Uses OpenMP locks as an example.
…I didn't catch this the first time, this particular issue isn't something I have managed to come across before.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
Added a tests folder for the math function support. There's a failing assert on the device for double-precision round() (internal function), at least for the case of 0^1 = 0.
Everything's defined as a inline function and should be usable just as their std:: counterparts. It also selects the single-precision versions when the input arguments are floats. We could probably improve on this behavior, but it'd probably need a global flipped in the hemi namespace or something similar.
I've tested it with CUDA 6 and g++ 4.8.