[PATCH] D52193: RFC: [clang] Multithreaded compilation support

[Zachary Turner via cfe-commits]

I can try to get some timings from my machine.  How do we handle crash
recovery in the case where we don't spawn a child process?  I thought the
whole reason for spawning the cc1 driver as a separate process was so that
we could collect and report crash information in a nice way.  Not having
that seems like a heavy price to pay.

On Wed, Oct 10, 2018 at 11:34 AM Alexandre Ganea via Phabricator <
reviews at reviews.llvm.org> wrote:

> aganea added a comment.
>
> Updated tests again. This time I've closed all applications, all unused
> services, and tried to have a "clean" machine.
>
> New timings without the child `clang-cl.exe` being invoked (hacked from
> https://reviews.llvm.org/D52411). The test consists in building
> Clang+LLVM+LLD at r343846 using the compiler specified on each line of the
> table.
>
> **Config 1:** Intel Xeon Haswell 6 cores / 12 HW threads, 3.5 GHz, 15 MB
> cache, 128 GB RAM, SSD 550 MB/s (Windows 10 Fall Creators update 1709)
>
> __Ninja:__
>
> | MSVC 15.8.6 (cl + link)
>            | 29 min 4 sec      |
> | clang-cl + lld-link 7.0 official release
>             | 25 min 45 sec     |
> | clang-cl + lld-link 8.0 r343846 //built with clang-cl 7.0 official//
> **(no child)** | **23 min 27 sec** |
> |
>
> **Config 2:** Intel Xeon Skylake 18 cores / 36 HW threads, x2 (Dual CPU),
> 72 HW threads total, 2.3 GHz, 24.75 MB cache each, 128 GB RAM, NVMe SSD 4.6
> GB/s (Windows 10 Fall Creators update 1709)
>
> __Ninja:__
>
> | MSVC 15.8.6 (cl + link)
>            | 6 min 40 sec     |
> | clang-cl + lld-link 7.0 official release
>             | 6 min 24 sec     |
> | clang-cl + lld-link 8.0 r343846 //built with clang-cl 7.0 official//
> **(no child)** | **5 min 10 sec** |
> |
>
> I'm wondering if the improvement is of the same order on Linux. I'd be
> interested to see your build times, out of curiosity? Can any of you try it
> with and without https://reviews.llvm.org/D52411?
>
>
> Repository:
>   rC Clang
>
> https://reviews.llvm.org/D52193
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20181010/cfce7909/attachment.html>