[PATCH] SHA-1: Add SSSE3 implementation

[Jussi Kivilinna]

On 13.12.2013 23:13, Werner Koch wrote:
> On Fri, 13 Dec 2013 15:10, jussi.kivilinna at iki.fi said:
> 
>> cpu                Old        New        Diff
>> Intel i5-4570      9.02 c/B   5.22 c/B   1.72x
>> Intel i5-2450M     12.27 c/B  7.24 c/B   1.69x
>> Intel Core2 T8100  7.94 c/B   6.76 c/B   1.17x
> 
> I justed drafted an article which says
> 
>    For the important SHA-1 algorithm 1.6 improves over 1.5 but is not
>    yet up to what Nettle does.  The better improvement is likely due to
>    changes in helper functions (stack burning) and not due to changes in
>    the actual implementation.  There is definitely room for
>    improvements.
> 
> You are a really fast hacker :-).  Guess I need to re-run the
> benchmarks.  Yes, please add this.  If there is ever a problem with this
> code the admin can disable it using /etc/gcrypt/hwf.deny.

Heh.. and I now have hacked up BMI2/AVX optimized SHA-1 which does 3.85 c/B
on Intel Haswell. But lets leave these for 1.7, because of dependency on
largeish 'bulk hash transform' patch.

-Jussi

> 
> 
> Salam-Shalom,
> 
>    Werner
>