On 08.03.22 02:53, Richard Henderson wrote:
> From: David Miller
>
> Prior to vector enhancements 2, the shift count was supposed to be equal
> for each byte lest the result be unpredictable, which allowed us to assume
> that the shift count was the same, and optimize accordingly.
>
> With vecto
From: David Miller
Prior to vector enhancements 2, the shift count was supposed to be equal
for each byte lest the result be unpredictable, which allowed us to assume
that the shift count was the same, and optimize accordingly.
With vector enhancements 2, the shift count is allowed to be differe