Artem Belevich f53524b4f6 [NVPTX] Added support for half-precision floating point. ...

Only scalar half-precision operations are supported at the moment.

- Adds general support for 'half' type in NVPTX.
- fp16 math operations are supported on sm_53+ GPUs only
  (can be disabled with --nvptx-no-f16-math).
- Type conversions to/from fp16 are supported on all GPU variants.
- On GPU variants that do not have full fp16 support (or if it's disabled),
  fp16 operations are promoted to fp32 and results are converted back
  to fp16 for storage.

Differential Revision: https://reviews.llvm.org/D28540

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@291956 91177308-0d34-0410-b5e6-96231b3b80d8

2017-01-13 20:56:17 +00:00

..

Analysis

[SCEV] Make howFarToZero max backedge-taken count check for precondition.

2017-01-11 21:07:15 +00:00

Assembler

[DebugInfo] Added support for Checksum debug info feature.

2016-12-25 10:12:09 +00:00

Bindings

[cmake] Canonicalize CMake booleans to 0/1 for lit interop

2017-01-06 21:33:48 +00:00

Bitcode

[ThinLTO] Subsume all importing checks into a single flag

2017-01-05 14:32:16 +00:00

BugPoint

…

CodeGen

[NVPTX] Added support for half-precision floating point.

2017-01-13 20:56:17 +00:00

DebugInfo

[DebugInfo] Handle same locations in DILocation::getMergedLocation

2017-01-12 20:34:35 +00:00

Examples

…

ExecutionEngine

RuntimeDyldELF: implement R_AARCH64_PREL64 reloc

2017-01-10 11:05:30 +00:00

Feature

…

FileCheck

Commit a test for match-full-lines.

2017-01-09 23:11:25 +00:00

Instrumentation

[asan] Don't overalign global metadata.

2017-01-12 23:26:20 +00:00

Integer

…

JitListener

[cmake] Canonicalize CMake booleans to 0/1 for lit interop

2017-01-06 21:33:48 +00:00

LibDriver

LibDriver: Allow resource files to be archive members.

2016-12-15 19:37:46 +00:00

Linker

Renumber testcase metadata nodes after r290153.

2016-12-22 00:45:21 +00:00

LTO

Renumber testcase metadata nodes after r290153.

2016-12-22 00:45:21 +00:00

MC

[AMDGPU] Assembler: SDWA/DPP should not accept scalar registers and immediate operands

2017-01-11 11:46:30 +00:00

Object

Object: Make IRObjectFile own multiple modules and enumerate symbols from all modules.

2016-12-13 20:20:17 +00:00

ObjectYAML

[ObjectYAML] Missed one mixup in the debug_line test

2017-01-10 06:24:24 +00:00

Other

Move test of lazy BFI with ORE to a generic directory

2017-01-13 00:16:23 +00:00

SymbolRewriter

…

TableGen

[tablegen] Delete duplicates from a vector without skipping elements

2016-12-01 19:38:50 +00:00

ThinLTO/X86

[ThinLTO] Import static functions from the same module as caller

2017-01-12 22:04:45 +00:00

tools

Revert r291903 and r291898. Reason: they break check-lld on the bots.

2017-01-13 16:45:15 +00:00

Transforms

[InstCombine] use m_APInt to allow shl folds for vectors with splat constants

2017-01-13 18:39:09 +00:00

Unit

…

Verifier

[TBAAVerifier] Be stricter around verifying scalar nodes

2016-12-29 15:47:05 +00:00

YAMLParser

…

.clang-format

…

CMakeLists.txt

[llvm-config] Print --system-libs only when static linking

2017-01-06 21:33:54 +00:00

lit.cfg

[llvm-config] Print --system-libs only when static linking

2017-01-06 21:33:54 +00:00

lit.site.cfg.in

[llvm-config] Print --system-libs only when static linking

2017-01-06 21:33:54 +00:00

TestRunner.sh

…