<feed xmlns='http://www.w3.org/2005/Atom'>
<title>emscripten-fastcomp/test, branch 1.13.0</title>
<subtitle>LLVM with the emscripten fastcomp javascript backend</subtitle>
<id>https://git.amat.us/emscripten-fastcomp/atom/test?h=1.13.0</id>
<link rel='self' href='https://git.amat.us/emscripten-fastcomp/atom/test?h=1.13.0'/>
<link rel='alternate' type='text/html' href='https://git.amat.us/emscripten-fastcomp/'/>
<updated>2014-02-28T20:39:49Z</updated>
<entry>
<title>Use range metadata instead of introducing selects.</title>
<updated>2014-02-28T20:39:49Z</updated>
<author>
<name>Dan Gohman</name>
<email>sunfish@mozilla.com</email>
</author>
<published>2014-02-28T20:39:49Z</published>
<link rel='alternate' type='text/html' href='https://git.amat.us/emscripten-fastcomp/commit/?id=dc59054db06864099c1b1051b8ef28a30f13e6c7'/>
<id>urn:sha1:dc59054db06864099c1b1051b8ef28a30f13e6c7</id>
<content type='text'>
When GlobalOpt has determined that a GlobalVariable only ever has two values,
it would convert the GlobalVariable to a boolean, and introduce SelectInsts
at every load, to choose between the two possible values. These SelectInsts
introduce overhead and other unpleasantness.

This patch makes GlobalOpt just add range metadata to loads from such
GlobalVariables instead. This enables the same main optimization (as seen in
test/Transforms/GlobalOpt/integer-bool.ll), without introducing selects.

The main downside is that it doesn't get the memory savings of shrinking such
GlobalVariables, but this is expected to be negligible.
</content>
</entry>
<entry>
<title>Update tests for recent changes which result in more parens being used in some places.</title>
<updated>2014-02-28T19:22:37Z</updated>
<author>
<name>Dan Gohman</name>
<email>sunfish@mozilla.com</email>
</author>
<published>2014-02-27T02:16:15Z</published>
<link rel='alternate' type='text/html' href='https://git.amat.us/emscripten-fastcomp/commit/?id=3ef11173bf12b5fdc0a21bfd883b057718676e8f'/>
<id>urn:sha1:3ef11173bf12b5fdc0a21bfd883b057718676e8f</id>
<content type='text'>
</content>
</entry>
<entry>
<title>Move ExpandI64 into lib/Target/JSBackend.</title>
<updated>2014-02-28T19:22:37Z</updated>
<author>
<name>Dan Gohman</name>
<email>sunfish@mozilla.com</email>
</author>
<published>2014-02-26T16:47:51Z</published>
<link rel='alternate' type='text/html' href='https://git.amat.us/emscripten-fastcomp/commit/?id=0dac79df771f7708a03e724126dded4f64531bc5'/>
<id>urn:sha1:0dac79df771f7708a03e724126dded4f64531bc5</id>
<content type='text'>
</content>
</entry>
<entry>
<title>Implement llvm.flt.rounds and disable the createRewriteLLVMIntrinsicsPass pass.</title>
<updated>2014-02-28T19:22:37Z</updated>
<author>
<name>Dan Gohman</name>
<email>sunfish@mozilla.com</email>
</author>
<published>2014-02-26T03:22:12Z</published>
<link rel='alternate' type='text/html' href='https://git.amat.us/emscripten-fastcomp/commit/?id=eaddc5d650f87a861ebdc6d3c49fb155ec791306'/>
<id>urn:sha1:eaddc5d650f87a861ebdc6d3c49fb155ec791306</id>
<content type='text'>
</content>
</entry>
<entry>
<title>Preserve alignment information when promoting integer loads and stores.</title>
<updated>2014-02-14T19:38:16Z</updated>
<author>
<name>Dan Gohman</name>
<email>sunfish@google.com</email>
</author>
<published>2014-02-14T03:24:56Z</published>
<link rel='alternate' type='text/html' href='https://git.amat.us/emscripten-fastcomp/commit/?id=71fb6dcb87d2abd302080149d4a583a5e35e3697'/>
<id>urn:sha1:71fb6dcb87d2abd302080149d4a583a5e35e3697</id>
<content type='text'>
</content>
</entry>
<entry>
<title>Preserve alignment information when splitting loads and stores.</title>
<updated>2014-02-14T19:38:16Z</updated>
<author>
<name>Dan Gohman</name>
<email>sunfish@google.com</email>
</author>
<published>2014-02-14T03:02:20Z</published>
<link rel='alternate' type='text/html' href='https://git.amat.us/emscripten-fastcomp/commit/?id=f3359807242e95f9e94bb35c0d90a73c794f1af0'/>
<id>urn:sha1:f3359807242e95f9e94bb35c0d90a73c794f1af0</id>
<content type='text'>
</content>
</entry>
<entry>
<title>Emscripten doesn't need to legalize pointers.</title>
<updated>2014-02-13T23:04:31Z</updated>
<author>
<name>Dan Gohman</name>
<email>sunfish@google.com</email>
</author>
<published>2014-02-13T22:51:59Z</published>
<link rel='alternate' type='text/html' href='https://git.amat.us/emscripten-fastcomp/commit/?id=7f2cf3ceadbcbb41227b9c0008c33d7e911e59e3'/>
<id>urn:sha1:7f2cf3ceadbcbb41227b9c0008c33d7e911e59e3</id>
<content type='text'>
</content>
</entry>
<entry>
<title>Don't leave behind unreachable blocks with illegal instructions.</title>
<updated>2014-02-13T19:47:44Z</updated>
<author>
<name>Dan Gohman</name>
<email>sunfish@google.com</email>
</author>
<published>2014-02-13T19:47:44Z</published>
<link rel='alternate' type='text/html' href='https://git.amat.us/emscripten-fastcomp/commit/?id=6f629a513c37587247dc83b961e86c177cef5d6d'/>
<id>urn:sha1:6f629a513c37587247dc83b961e86c177cef5d6d</id>
<content type='text'>
</content>
</entry>
<entry>
<title>Revamp ExpandI64.</title>
<updated>2014-02-13T06:30:21Z</updated>
<author>
<name>Dan Gohman</name>
<email>sunfish@google.com</email>
</author>
<published>2014-02-13T02:09:52Z</published>
<link rel='alternate' type='text/html' href='https://git.amat.us/emscripten-fastcomp/commit/?id=718e02f4cea34a6f34d7a6f1f82411960bff1ba4'/>
<id>urn:sha1:718e02f4cea34a6f34d7a6f1f82411960bff1ba4</id>
<content type='text'>
By using a reverse-postorder traversal of the basic blocks, we can perform
this transform in a single pass. This eliminates the need for tricky
coordination between the passes.
</content>
</entry>
<entry>
<title>Generalize PromoteIntegers to handle arbitrary bit widths.</title>
<updated>2014-02-13T06:30:21Z</updated>
<author>
<name>Dan Gohman</name>
<email>sunfish@google.com</email>
</author>
<published>2014-02-13T02:09:29Z</published>
<link rel='alternate' type='text/html' href='https://git.amat.us/emscripten-fastcomp/commit/?id=2b36da6f5426d0ea67c85fdee8016a5bbc828a23'/>
<id>urn:sha1:2b36da6f5426d0ea67c85fdee8016a5bbc828a23</id>
<content type='text'>
</content>
</entry>
</feed>
