Speed now & then

Eric Korpela korpela at ssl.berkeley.edu
Wed Apr 11 11:31:34 CDT 2018

It might break the rules since it only goes back to 1999, but here's
Moore's law for integer speed, floating point speed, number of processors,
memory sizes and disk sizes for the machines connected to SETI at home.  Plots
are averages and medians, unfiltered for errors.   At least one of the
parameters (total credited ops) is no longer used.    expavg credit
includes GPU work even though it's not included in the processor speed
numbers.  If someone wants a distillation of the data for comparing to
earlier machines, let me know.


On Thu, Mar 29, 2018 at 10:53 AM, Paul Koning via cctalk <
cctalk at classiccmp.org> wrote:

> > On Mar 29, 2018, at 12:20 PM, Murray McCullough via cctalk <
> cctalk at classiccmp.org> wrote:
> >
> > I’m not trying to date myself but have things truly sped up? In 1970’s
> > Toronto I had a classic computer, sorry can’t recall what it was,
> connected
> > to a 300 baud modem; by early 80’s had ‘zoomed’ to 9600 baud. Oh, my! [ A
> > typical file size to download was probably 1 MB. ] Speed indeed! Yet now,
> > here in rural Ontario, Canada, I’m at 5MB/s. Yikes! (Friends in Toronto
> are
> > at 50MB/s.) We can do the math but content, particularly multimedia, has
> > swollen in size.[ 1 GB is not unheard of. ] Were classic computing days
> > that much slower? Happy computing. Murray  -:)
> I remember downloading the GCC release kit over a 56k dialup line, in
> 2000.  Took a while.
> The ARPAnet in its early days had "high speed backbone" links which were
> 56k bps.  Terminal links presumably 110 bps, that being the speed of ASCII
> teletypes.  And back in the late 1970s you could still find even slower
> links in some places, such as 6 bit links connecting teletype machines for
> newspaper "wire service" feeds.
> It would be fun to do a "generalized Moore's Law" chart, showing not just
> transistor count growth (Moore's subject) but also the many other scaling
> changes of computing: disk capacity, recording density, disk IOPS, disk
> bandwidth, ditto those for tape, CPU MIPS, memory size, memory bandwidth,
> network bandwidth...
> All these have grown dramatically, but very clearly not in the same
> proportion, for some of these the changes are smooth while others are
> jumps, and the rate of change sometimes varies dramatically over the
> decades.
>         paul

Eric Korpela
korpela at ssl.berkeley.edu

More information about the cctalk mailing list