Re: [Qemu-devel] [PATCH v15 33/33] target-tilegx: Handle v1shl, v1shru, v1shrs

[Richard Henderson]

On 09/03/2015 08:02 AM, Eric Blake wrote:
> Is it any more efficient to use multiplies instead of looping, as in:
>
> uint64_t m;
>
> b &= 7;
> m = 0x0101010101010101ULL * ((1 << (8 - b)) - 1);
> return (a & m) << b;
>
> Or if multiplies are bad, what about straight-line expansion of the
> mask, as in:
>
> uint64_t m;
>
> b &= 7;
> m = (1 << (8 - b)) - 1;
> m |= m << 32;
> m |= m << 16;
> m |= m << 8;
> return (a & m) << b;

The multiply will (likely) be expanded by the compiler to the shift, but yes, I 
should have played the mask game after Peter pointed out that this wasn't a 
per-element variable shift in v14.


r~