[tor-bugs] #9663 [Tor]: Table-based basepoint multiply optimizations for ntor handshake

[Tor Bug Tracker & Wiki]

#9663: Table-based basepoint multiply optimizations for ntor handshake
-------------------------+-------------------------------------------------
     Reporter:  nickm    |      Owner:
         Type:           |     Status:  needs_revision
  enhancement            |  Milestone:  Tor: 0.2.7.x-final
     Priority:  normal   |    Version:
    Component:  Tor      |   Keywords:  tor-relay, performance, ntor,
   Resolution:           |  026-triaged-0, 027-triaged-1-in
Actual Points:           |  Parent ID:  #9662
       Points:           |
-------------------------+-------------------------------------------------

Comment (by nickm):

 Replying to [comment:18 yawning]:
 > Replying to [comment:17 nickm]:
 > > eg (untested)
 > >
 > > {{{
 > > #if defined(__SSE2__) && !(defined(__x86_64__) || defined(__amd64__))
 > > ....
 > > #endif
 > > }}}
 > > ?
 >
 > This should work (might need `x86intrin.h`?).

 (Probably won't need that; those are supposed to be predefined by the
 compiler.)

 >  I'd like to see a comparison between ref10, 32 bit donna, and SSE2
 donna before we go down this path, and for the purposes of the feature
 branch as is, I will ignore the issue for now since doing the integration
 work in a safe manner is more important.

 Makes sense.

 > Will we ever consider removing the ref10 code/fallback stuff?  I don't
 think it particularly makes a difference to how I'm going to do this, but
 I think the portable versions of the donna code is well tested at this
 point.

 We can consider removing it, sure.  I think I'd like to leave it in for a
 while as a fallback though.

--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/9663#comment:19>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online