[LLVMdev] [Proposal] Parallelize post-IPO stage.

Shuxin Yang shuxin.llvm at gmail.com
Tue Jul 16 14:10:19 PDT 2013

On 7/16/13 2:04 PM, Xinliang David Li wrote:
> On Tue, Jul 16, 2013 at 1:49 PM, Shuxin Yang <shuxin.llvm at gmail.com> wrote:
>> I have actually came up the 3 approaches to build the post-ipo object
>> independently.
>> The "3rd approach" here is the 1st solution in my original proposal. Almost
>> all coworkers call it sucks:-)
>> Now I accept it because the it has no way to be adaptive.
>> Consider the scenario we compile the llvm compiler. We use "make -j16" for
>> computer with 8 processor, each make-thread invoke a compiler which may
>> blindly invoke 16 threads!
>> So, we end up to have 16*16 threads.
> Determining the right parallelism is not the job of the compiler
> (builtin) nor that of a developer -- the underlying build system
> should take care of the scheduling :)
> David
People told me we have some magic lib which is able to figure out how 
heavy the
system is.  Compiler just need to link against them.  I have not yet got 
to try these libs, neither do I know their names at this moment.

Since how to compiler post-LTO is small part of this project,  I'd like 
to pick up
whichever easier first. We may (very likely) end up all these possible 
three approaches.

More information about the llvm-dev mailing list