Re: Nocona [Intel 64-bit cpu timing]

tomstdenis_at_gmail.com
Date: 08/26/05 

Date: 25 Aug 2005 17:38:33 -0700

tomstdenis@gmail.com wrote:
> Andrew Swallow wrote:
> > tomstdenis@gmail.com wrote:
> > [snip]
> > >
> > > If I didn't buy this Intel box solely for benchmarking I'd be really
> > > pissed at it's craptacular performance. I mean when a single core
> > > AMD64 [cheaper than the EMT64 core] can beat things like AES and RSA
> > > often by huge margins ... you know there are problems..
> >
> > Multi-core machines assume that reading from and writing
> > to main memory is rare. Or put it another way, if you
> > need more than one mega byte of shared memory a single
> > core machine will normally be faster. (There are
> > special cases like where one of the processors is
> > the disk controller and the program is in private
> > memory.)
>
> That'd be an interesting fact except my AMD64 is a dual-core. Running
> single thread tasks it gets roughly the performance of a single-core
> AMD64. Obviously there is overhead from even something like an idle
> thread...

Whoops, context here... [sorry for the confusion]

My comment was that dual-core RSA [using two parallel threads] on the
nocona is STILL SLOWER than a single-core AMD64 running RSA. A
dual-core AMD64 just murders it [>2x faster].

The AMD64 scales nicely with multiple cores, though so does the P4. I
get nearly 2x the speed from my dual-core RSA plugin for LibTomCrypt on
the nocona. So if the ALU itself wasn't so pathetic this might have
been a contender :-)

Tom

Relevant Pages

  • Re: Nocona [Intel 64-bit cpu timing]
    ... memory at 4.5GiB/sec. ... As in the case of the Nocona vs. AMD64. ... my Prescott when on a Pentium M. ... The Prescott tops out at probably around ...
    (sci.crypt)
  • Re: vm.kmem_size settings doesnt affect loader?
    ... If your machine has some small amount of memory, ... then you probably shouldn't be using ZFS. ... The i386 vs. amd64 argument is bogus, ... kernel code and absolutely no experience with kernel memory management. ...
    (freebsd-stable)
  • Re: vm.kmem_size settings doesnt affect loader?
    ... If your machine has some small amount of memory, ... I know that i386 is highly unrecommended for ZFS, but it's just a home box for testing and learning purposes - I just want to know what I'm doing and what should I expect when I decide to put ZFS on server machines:) Currently, from posts on freebsd-fs, I conclude that even with a gigs of kmem and using AMD64, we still can experience panic from kmem_malloc. ... Of course I've followed tuning guide carefully. ...
    (freebsd-stable)
  • Re: swap
    ... I'm wondering how memory gets packed when ... I have both on this AMD64 system, but I find myself most of the time ... packages being compiled for a 64 bit system on a 64 bit system will work ... As for swap space, even in the AMD64 version of Etch that I have, it ...
    (Debian-User)