llvm-project/clang/test/CodeGenCUDA
Justin Lebar d3a44f6885 [CUDA] Add -fcuda-flush-denormals-to-zero.
Summary:
Setting this flag causes all functions are annotated with the
"nvvm-f32ftz" = "true" attribute.

In addition, we annotate the module with "nvvm-reflect-ftz" set
to 0 or 1, depending on whether -cuda-flush-denormals-to-zero is set.
This is read by the NVVMReflect pass.

Reviewers: tra, rnk

Subscribers: cfe-commits

Differential Revision: http://reviews.llvm.org/D18671

llvm-svn: 265435
2016-04-05 18:26:20 +00:00
..
Inputs [CUDA] Make printf work. 2016-01-23 21:28:14 +00:00
address-spaces.cu [CUDA] Allow trivial constructors as initializer for __shared__ variables. 2015-09-10 17:26:58 +00:00
alias.cu [CUDA] Don't generate aliases for static extern "C" functions. 2016-01-25 22:36:37 +00:00
convergent.cu [CUDA] Mark all CUDA device-side function defs, decls, and calls as convergent. 2016-02-24 21:55:11 +00:00
cuda-builtin-vars.cu [cuda] Added support for CUDA built-in variables. 2015-04-21 22:14:13 +00:00
device-stub.cu Fixed test failure platforms with name mangling different from Linux. 2016-03-02 21:03:20 +00:00
device-var-init.cu [CUDA] Mark all CUDA device-side function defs, decls, and calls as convergent. 2016-02-24 21:55:11 +00:00
device-vtable.cu [CUDA] Make vtable construction aware of host/device side of CUDA compilation. 2015-12-17 18:12:36 +00:00
filter-decl.cu [CUDA] Emit host-side 'shadows' for device-side global variables 2016-03-02 18:28:50 +00:00
flush-denormals.cu [CUDA] Add -fcuda-flush-denormals-to-zero. 2016-04-05 18:26:20 +00:00
function-overload.cu [CUDA] Remove three obsolete CUDA cc1 flags. 2016-03-29 16:24:16 +00:00
host-device-calls-host.cu [CUDA] Remove three obsolete CUDA cc1 flags. 2016-03-29 16:24:16 +00:00
kernel-call.cu Move all CUDA testing inputs to Inputs/ subdirectory inside the tests. 2014-04-28 22:21:28 +00:00
launch-bounds.cu [cuda] Allow using integral non-type template parameters as launch_bounds attribute arguments. 2015-04-21 22:55:54 +00:00
link-device-bitcode.cu [CUDA] Add -disable-llvm-passes to CodeGenCUDA/link-device-bitcode.cu. NFC 2016-03-30 23:45:38 +00:00
llvm-used.cu When generating llvm.used, we may need an addrspacecast instead of a bitcast. 2015-02-02 21:05:49 +00:00
printf-aggregate.cu [CUDA] Don't crash when trying to printf a non-scalar object. 2016-02-11 02:00:52 +00:00
printf.cu [CUDA] Generate CUDA's printf alloca in its function's entry block. 2016-01-28 23:58:28 +00:00
ptx-kernels.cu [CUDA] __global__ functions should always be visible externally. 2015-09-23 17:44:53 +00:00