[LLVMdev] [lldb-dev] Has Fast register allocation been finished in 3.3 version?

[Andrew Trick]

On Oct 22, 2013, at 5:03 AM, 罗勇刚(Yonggang Luo) <luoyonggang at gmail.com> wrote:

> Forward to llvmdev
> 
> ---------- 转发的邮件 ----------
> 发件人："Chris Lattner" <clattner at apple.com>
> 日期：2013年10月22日 下午7:33
> 主题：Re: [lldb-dev] Has Fast register allocation been finished in 3.3	version?
> 收件人： <wansheg at gmail.com>
> 抄送： <lldb-dev at cs.uiuc.edu>
> 
> 
> On Oct 21, 2013, at 7:43 PM, wansheg at gmail.com wrote:
> 
>> 
>> When I use llc command with --regalloc=fast ,  get error 

And I presume the error is:
LLVM ERROR: regalloc=... not currently supported with -O0

-regalloc=fast is not a very useful flag because it requires “-O0” and is also implied by -O0.

However, if you still want all the other llc optimizations you can use an internal option instead of -O0:

-optimize-regalloc=false -regalloc=fast

Register allocation is not just the allocation algorithm. It’s a set of passes that cooperate. The ‘fast’ allocator uses a different pipeline than the optimizing global allocators.

>> In addition, is lli use fast register allocation?

AFAIK, the -O levels are the same, with -O0 using fast regalloc.

-Andy

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20131023/42ab1394/attachment.html>