Re: UDB speed comparison

Harald Koenig (koenig@tat.physik.uni-tuebingen.de)
Wed, 25 Sep 1996 08:28:49 +0200 (MET DST)

>
> "Tracy R. Reed" <treed@straylight.connectnet.com> wrote:
>
> > host CPU Time
> >
> > redhat-udb 166Mhz 21066 0:10.01
> > straylight 66Mhz 486 1:24.43
> > sd 90Mhz Pentium 0:43.55
> > rohan SS1000E 1cpu 0:09.17
> > voyager 133Mhz Pentium 0:15.58
> > helios SS20 0:10.71
> >
> > So the UDB came up with 2,400 digits of pi in 10.01 seconds. Not bad. :)
> > It's the second fastest machine I have access to, but just barely. I wish
> > I still had that old account on the Cray II at Kirtland AFB. :)
>
> Here's the C90 result (Cray cc, not gcc; not vectorized/parallelized/etc.):
>
> time ./pihex > /dev/null
> 5.7767u 0.0361s 0:05 97%
>
> Presumably you're counting user+system time, so 5.81 seconds.
>
> (Do I win? :-) )

but only very barely, I'm quite close with my PC (ASUS SP3G with PODP-83)
and a DEC UDB233 and XL266...

PODP-83 0:20.71 real, 20.55 user, 0.04 sys, 99% cpu
21066-233 0:07.32 real, 7.19 user, 0.03 sys, 98% cpu
21064-266 0:06.12 real, 6.08 user, 0.00 sys, 99% cpu

so at least in price/performace ratio your result isn't very good ;-)

Harald

--
All SCSI disks will from now on                     ___       _____
be required to send an email notice                0--,|    /OOOOOOO\
24 hours prior to complete hardware failure!      <_/  /  /OOOOOOOOOOO\
                                                    \  \/OOOOOOOOOOOOOOO\
                                                      \ OOOOOOOOOOOOOOOOO|//
Harald Koenig,                                         \/\/\/\/\/\/\/\/\/
Inst.f.Theoret.Astrophysik                              //  /     \\  \
koenig@tat.physik.uni-tuebingen.de                     ^^^^^       ^^^^^