[llvm-bugs] [Bug 43796] New: [WebAssembly] Add a way to run binaryen's wasm-opt automatically

[via llvm-bugs]

https://bugs.llvm.org/show_bug.cgi?id=43796

            Bug ID: 43796
           Summary: [WebAssembly] Add a way to run binaryen's wasm-opt
                    automatically
           Product: clang
           Version: unspecified
          Hardware: All
                OS: All
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: -New Bugs
          Assignee: unassignedclangbugs at nondot.org
          Reporter: dan433584 at gmail.com
                CC: aardappel at gmail.com, aheejin at gmail.com,
                    htmldeveloper at gmail.com, kripkensteinr at gmail.com,
                    llvm-bugs at lists.llvm.org, neeilans at live.com,
                    richard-llvm at metafoo.co.uk, sbc at chromium.org

Alon asked me recently if it would make sense for the clang driver to run
Binaryen's wasm-opt automatically. My initial reaction was that LLVM/clang
tools don't typically run external optimization tools to post-process their
output, but thinking about it more, it seems like it might be ok if we take
care.

As background, while the in-tree LLVM WebAssembly target emits correct output,
and does some wasm-specific optimizations, there are a lot more such
optimizations possible that it doesn't do. Binaryen (which is not part of LLVM)
has a tool called wasm-opt which does do many more of these optimizations, so
it can be quite beneficial to run it on LLVM's output. LLVM could conceivably
be made to do those optimizations too, but right now, no one is actively
working on this, so unless this changes, it's desirable to use wasm-opt.

We'd still want to ensure that LLVM doesn't *depend* on having these tools
available. But if they are available, it would be convenient for users to be
able to have the clang driver run them automatically.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20191024/a2b98261/attachment.html>