llvm-project/libclc
Jan Vesely e0edcaa4a9 r600/fmin: Flush denormals before calling builtin.
Same reason as amdgcn.
Fixes fmin, minmag CTS on turks.
Reviewer: Tom Stellard <tstellar@redhat.com>
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>

llvm-svn: 334228
2018-06-07 20:27:58 +00:00
..
amdgcn/lib Add initial support for half precision builtins 2018-05-17 22:55:30 +00:00
amdgcn-amdhsa/lib amdgcn: Fix build after GDS/const AS swap in r325030 2018-02-23 07:37:01 +00:00
amdgpu/lib Add initial support for half precision builtins 2018-05-17 22:55:30 +00:00
build configure.py: Make python3 friendly 2017-08-02 15:00:59 +00:00
generic math/fma: Add fp32 software implementation 2018-06-07 20:27:43 +00:00
ptx/lib ptx: Use __clc_nextafter to implement nextafter 2017-10-08 19:34:00 +00:00
ptx-nvidiacl/lib Implement mem_fence on ptx 2017-10-09 19:43:04 +00:00
r600/lib r600/fmin: Flush denormals before calling builtin. 2018-06-07 20:27:58 +00:00
test
utils utils: Adapt to llvm r325155 2018-02-23 07:37:03 +00:00
www Update page to list supported targets 2016-02-13 01:02:06 +00:00
.gitignore .gitignore: Ignore amdgcn-mesa object directory 2017-02-24 20:32:18 +00:00
.travis.yml travis: Add build using llvm-6 2018-03-12 19:46:48 +00:00
CREDITS.TXT
LICENSE.TXT Update copyright year to 2016. 2016-03-30 22:39:03 +00:00
README.TXT
check_external_calls.sh check_external_calls.sh: Print number of calls in tested file. 2017-10-08 20:07:56 +00:00
compile-test.sh
configure.py configure.py: Add gfx900 (Vega, Raven) 2017-11-27 11:14:06 +00:00

README.TXT

This file contains invisible Unicode characters

This file contains invisible Unicode characters that are indistinguishable to humans but may be processed differently by a computer. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

libclc
------

libclc is an open source, BSD licensed implementation of the library
requirements of the OpenCL C programming language, as specified by the
OpenCL 1.1 Specification. The following sections of the specification
impose library requirements:

  * 6.1: Supported Data Types
  * 6.2.3: Explicit Conversions
  * 6.2.4.2: Reinterpreting Types Using as_type() and as_typen()
  * 6.9: Preprocessor Directives and Macros
  * 6.11: Built-in Functions
  * 9.3: Double Precision Floating-Point
  * 9.4: 64-bit Atomics
  * 9.5: Writing to 3D image memory objects
  * 9.6: Half Precision Floating-Point

libclc is intended to be used with the Clang compiler's OpenCL frontend.

libclc is designed to be portable and extensible. To this end, it provides
generic implementations of most library requirements, allowing the target
to override the generic implementation at the granularity of individual
functions.

libclc currently only supports the PTX target, but support for more
targets is welcome.

Compiling and installing with Make
----------------------------------

$ ./configure.py --with-llvm-config=/path/to/llvm-config && make
$ make install

Note you can use the DESTDIR Makefile variable to do staged installs.

$ make install DESTDIR=/path/for/staged/install

Compiling and installing with Ninja
-----------------------------------

$ ./configure.py -g ninja --with-llvm-config=/path/to/llvm-config && ninja
$ ninja install

Note you can use the DESTDIR environment variable to do staged installs.

$ DESTDIR=/path/for/staged/install ninja install

Website
-------

http://www.pcc.me.uk/~peter/libclc/