[llvm-dev] LLVM Weekly - #231, June 4th 2018

Previous Topic Next Topic
classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view

[llvm-dev] LLVM Weekly - #231, June 4th 2018

Tim Northover via llvm-dev
LLVM Weekly - #231, June 4th 2018

If you prefer, you can read a HTML version of this email at

Welcome to the two hundred and thirty-first issue of LLVM Weekly, a weekly
newsletter (published every Monday) covering developments in LLVM, Clang, and
related projects. LLVM Weekly is brought to you by [Alex
Bradbury](https://www.linkedin.com/in/alex-bradbury/). Subscribe to future
issues at <http://llvmweekly.org> and pass it on to anyone else you think may
be interested. Please send any tips or feedback to <[hidden email]>, or
@llvmweekly or @asbradbury on Twitter.

## News and articles from around the web

Nathan Froyd
on the possibility of Firefox building using LLVM and Clang across all major
platforms, and the advantages/disadvantages/risks of standardising on a single
compiler toolchain.

## On the mailing lists

* Michael Kruse has written RFCs on [extending #pragma clang
loop](http://lists.llvm.org/pipermail/cfe-dev/2018-May/058141.html) and
[extending loop metadata in
LLVM](http://lists.llvm.org/pipermail/llvm-dev/2018-May/123690.html). As
explained in the RFC, the intent is to propose a form of these loop
transformation pragmas to the OpenMP standard.

* Sohail Somani
[proposes](http://lists.llvm.org/pipermail/llvm-dev/2018-May/123632.html) a
new API to simplify the interface for creating new forms of debug info

* Anast Gramm reports on an [investigation of the impact of the Scalar
Replacement of Aggregates pass on
Although the DExTer tool showed SROA had a negative impact on debug
experience, Anast's experiment indicated that SROA does a good job of
preserving debug information. Greg Bedwell offers
on some of the subtleties of interpreting DExTer results.

## LLVM commits

* A new WebAssembly exception handling prepare pass was added.

* ADDE/ADDC/SUBE/SUBC have been set to expand by default. Out-of-tree backends
that use these deprecated opcodes may need to be updated.

* The RISC-V backend gained peepholes to improve codegen for some common
global address lowering cases. The intent is to replace these peepholes with a
more complete MachineFunctionPass.

* The core logic of LoopInstSimplify has been rewritten to be both simpler and
more efficient. [r333461](https://reviews.llvm.org/rL333461).

* The MCSchedModel API was extended with new methods for querying latency and
reciprocal throughput information for an MCInst.

## Clang commits

* Initial infrastructure for the HIP language has started to land. As
explained in the [original
RFC](http://lists.llvm.org/pipermail/cfe-dev/2018-March/057426.html), HIP is a
language similar to CUDA but with a vendor neutral host API.

* The `-compiler-option-dump` flag will dump features/extensions lists to a
JSON file. [r333653](https://reviews.llvm.org/rL333653).

## Other project commits

* The XRay Profile Collector Service has been committed to compiler-rt.

* LLD's WebAssembly linker gained initial support for LTO.
LLVM Developers mailing list
[hidden email]