On 11/20/2015 05:50 PM, Dimitry Sibiryakov wrote:
> Hello, All.
>
> If we revive subj, should it have higher priority than application
> callback or lower?
>
May be treat providing both as a bug?
--
Go from
Hello, All.
Can I be sure that "from" and "to" pointers got by plugins's encrypt/decrypt
routines
are always aligned to 16 bytes boundary?
--
WBR, SD.
--
Go from Idea to Many App Stores Faster with Intel(R)
On 11/26/2015 03:38 PM, Dimitry Sibiryakov wrote:
> Hello, All.
>
> Can I be sure that "from" and "to" pointers got by plugins's
> encrypt/decrypt routines
> are always aligned to 16 bytes boundary?
>
I think we can add such requirement to caller.
26.11.2015 13:45, Alex Peshkoff wrote:
> I think we can add such requirement to caller.
Google say that Intel claim zero-impact for unaligned load in i7 core, so we
can forget
that.
--
WBR, SD.
--
Go from Idea
On 11/26/2015 7:38 AM, Dimitry Sibiryakov wrote:
> Hello, All.
>
> Can I be sure that "from" and "to" pointers got by plugins's
> encrypt/decrypt routines
> are always aligned to 16 bytes boundary?
>
I'm curious. Why does that matter?
26.11.2015 16:36, Leyne, Sean wrote:
> What about other CPU platforms (ie. ARM)???
I have no idea how they support _mm_loadu_si128() intrinsic (if at all).
--
WBR, SD.
--
Go from Idea to Many App Stores Faster