Re: What can old sparc be used for?

wrote:
>
> Dr. David Kirkby <> wrote:
> [...]
> > You won't get the performance from 4 CPUs in an SS20 that you will
> > with one in a Pentium III, but you will learn something that you could
> > never learn on most PCs, or the Ultra 1, Ultra 5, Ultra 10 or Ultra
> > 30. It's only when you go to the Ultra 60 (dual CPU) or Ultra 80 (quad
> > CPU) that you start getting multi-processor capability, but then the
> > machines cost a lot more.
>
> Hm - wouldn't the U2 then be an affordable alternative? Should be a
> good deal faster than the SS20 and quite a bit cheaper than U60/U80,
> right?
>
> Cheerio,
>
> Thomas

Sorry, you are right. The U2 is a dual processor box and would have a
price and performance somewhere between that of the SS20 and the Ultra
60, although it can't take the 4 CPUs the SS20 can.

Another use, which I forgot to mention for old SPARCs is what I use
them for. I have 5, each running a different operating system. (Two
versions of Linux, Solaris 2.5, NetBSD and OpenBSD). I then use them
to check the portability of code I write, ensuring it runs on older
systems, not just the Solaris 9 Ultra 80 I write it on. I've often
found bugs that appear on one system but not on another. With a bit of
effort, you can work around them, so ensuring makimum portabilitly.

Old SPARCs are cheap enough to do this with. Using even Ultra 1's
would add significantly to the cost of the hardware, while gaining me
very little, since in this case I don't care too much how long it
takes to compile/build/test the software. I just have a script that
copies a .tar.gz file to one of the SS20's by scp, unpacks it there,
configure, makes, tests etc. It's all done in the background on
multiple machines at once, and I don't care too much how long it
takes.


I've never had much success with multi-threaded code on old versions
of Linux. It could be a problem in my code, but I'm temted to belive
the thread libraries of older versions of Linux are broken.
--
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail