This is an archive of the discontinued LLVM Phabricator instance.

Differential D28326

[Docs] Update docs to indicate that CUDA compilation is supported on Windows.
ClosedPublic

Authored by jlebar on Jan 4 2017, 2:04 PM.

Details

Reviewers
None
Commits
rGe2cd288f57b6: [Docs] Update docs to indicate that CUDA compilation is supported on Windows.
rL291139: [Docs] Update docs to indicate that CUDA compilation is supported on Windows.
Summary

See "stack" section for the graph of commits needed to support CUDA on Windows.

Diff Detail

Event Timeline

jlebar updated this revision to Diff 83134.Jan 4 2017, 2:04 PM
jlebar retitled this revision from to [Docs] Update docs to indicate that CUDA compilation is supported on Windows..
jlebar updated this object.
jlebar added subscribers: llvm-commits, cfe-commits.
jlebar added parent revisions: D28325: [CUDA] Rename keywords used in macro so they don't conflict with MSVC., D28324: [CUDA] Don't define functions that the CUDA headers themselves define on Windows., D28323: [CUDA] Let NVPTX inherit the host's calling conventions., D28322: [CUDA] More correctly inherit primitive types from the host during device compilation., D28321: [CUDA] Add __declspec spellings for CUDA attributes., D28320: [Driver] Driver changes to support CUDA compilation on Windows., D28319: [CUDA] Make CUDAInstallationDetector take the host triple in its constructor., D28317: [Windows] Remove functions in intrin.h that are defined in Builtin.def., D28318: [TableGen] Only normalize the spelling of GNU-style attributes..Jan 4 2017, 2:05 PM
jlebar removed a parent revision: D28318: [TableGen] Only normalize the spelling of GNU-style attributes..
jlebar removed a parent revision: D28319: [CUDA] Make CUDAInstallationDetector take the host triple in its constructor..
jlebar updated this object.Jan 4 2017, 2:07 PM
jlebar added subscribers: tra, rnk, hfinkel.
jlebar added a subscriber: rryan.Jan 4 2017, 5:29 PM
Closed by commit rL291139: [Docs] Update docs to indicate that CUDA compilation is supported on Windows. (authored by jlebar). · Explain WhyJan 5 2017, 9:05 AM
This revision was automatically updated to reflect the committed changes.

Revision Contents

PathSize
llvm/
docs/
4 lines

Diff 83134

llvm/docs/CompileCudaWithLLVM.rst

Show All 29 Lines
driver for your nvidia GPU and the CUDA SDK. See `NVIDIA's CUDA installation driver for your nvidia GPU and the CUDA SDK. See `NVIDIA's CUDA installation
guide <https://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html>`_ guide <https://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html>`_
for details. Note that clang `does not support for details. Note that clang `does not support
<https://llvm.org/bugs/show_bug.cgi?id=26966>`_ the CUDA toolkit as installed <https://llvm.org/bugs/show_bug.cgi?id=26966>`_ the CUDA toolkit as installed
by many Linux package managers; you probably need to install nvidia's package. by many Linux package managers; you probably need to install nvidia's package.
You will need CUDA 7.0, 7.5, or 8.0 to compile with clang. You will need CUDA 7.0, 7.5, or 8.0 to compile with clang.
CUDA compilation is supported on Linux, and on MacOS as of XXXX-XX-XX. Windows CUDA compilation is supported on Linux, on MacOS as of 2016-11-18, and on
support is planned but not yet in place. Windows as of 2017-01-XX.
Invoking clang Invoking clang
-------------- --------------
Invoking clang for CUDA compilation works similarly to compiling regular C++. Invoking clang for CUDA compilation works similarly to compiling regular C++.
You just need to be aware of a few additional flags. You just need to be aware of a few additional flags.
You can use `this <https://gist.github.com/855e277884eb6b388cd2f00d956c2fd4>`_ You can use `this <https://gist.github.com/855e277884eb6b388cd2f00d956c2fd4>`_
▲ Show 20 Lines • Show All 512 Lines • Show Last 20 Lines