[Webkit-unassigned] [Bug 55158] Use VFP for double to int truncate in ARMv7

Fri Apr 29 11:12:37 PDT 2011

https://bugs.webkit.org/show_bug.cgi?id=55158

--- Comment #10 from Chao-ying Fu <fu at mips.com>  2011-04-29 11:12:37 PST ---
(In reply to comment #9)
> Gah, I thought I'd checked all platforms before making my comment, I'd missed the extra clause on the ifdef in MIPS's supportsFloatingPoint.
> 
> Based on a little wikipedia research, the only chips toting MIPS I architecture appear to be over 20 years old, and run at under 100MHz.  If this understanding is correct, I doubt there are many users out there running the JSC JIT on these chips, and as such the performance benefit for having (some) FP enabled is unlikely to be significant!

I know only Maciej W. Rozycki uses MIPS1.  But I don't know if he runs WebKit.  Other people should be ok.

> 
> As such, it doesn't seem unreasonable to unify the MIPS floating point support on:
>     #if WTF_MIPS_DOUBLE_FLOAT && WTF_MIPS_ISA_AT_LEAST(2)
> so that we can unify the supportsFloatingPoint & supportsFloatingPointTruncate methods.
> 
> Would be great to get some input from the mips folk to confirm that my understanding of the WTF_MIPS_ISA_AT_LEAST is correct, and that we would not be degrading the performance for real deployment uses.

I am fine to unify the floating-point support by adding WTF_MIPS_ISA_AT_LEAST(2).  Thanks!

Regards,
Chao-ying

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.