[llvm-dev] Why is Partial Unrolling off by default?

Friedman, Eli via llvm-dev llvm-dev at lists.llvm.org
Mon Jul 30 11:34:31 PDT 2018


On 7/30/2018 9:02 AM, Bhatu via llvm-dev wrote:
> Hi all,
>
> Why is partial unrolling off by default? Did enabling partial 
> unrolling result in any significant runtime or compile time 
> regressions in any benchmarks?

It's not off by default on all targets; it's controlled by the 
getUnrollingPreferences() target hook.  The exact threshold which makes 
sense tends to be target-specific (especially for CPUs which have 
special optimizations for small loops).

-Eli

-- 
Employee of Qualcomm Innovation Center, Inc.
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project



More information about the llvm-dev mailing list