🚀 go-pugleaf

In article <1e824rINNlpu@iraul1.ira.uka.de> wolpers@i11s10.ira.uka.de (Andreas Wolpers) writes:
>Hello everybody,
>
>I'm having a minor problem on which I would welcome any comments:
>Since I'll have some money to spend real soon now, I've been pondering
>whether we should switch from Sun to HP. The usual benchmark results
>suggest that a change might result in faster execution of our pet
>program, a large "theorem prover" written in Lisp.

You should get Sun and HP to send you computers to test, then send back
the slowest.  From your comments below, I assume that you at least got
Sun to send you a machine...

>From the SpecInt92 results, one should expect a performance increase
>of about a factor of 2 when switching from a SS2 to either s SS10-30
>or HP720. Unfortunately, on both machines turned out to but just
>30% faster than a SS2 when running our system (for which ps NEVER
>shows a resident set under 3-4 MB, not even on an 8MB machine).

(a) I am sure the SPECInt's ops were all normal add's and sub's, and
I am guessing that Lisp's are all taged add's and sub's.  Can you
execute >1 tagged op per cycle?  Also did you re-compile the lisp
interpreter on the SS10 with a compiler that can schedule code for
the new CPU?

(b) If you only have 8M, you may be out of RAM even when ps says your
code is only using 3-4M.  First, ps doesn't (by default) report pages
used by shared libs, which can be quite a few pages (besudes if it did
report them, who should it assign them to?).  Second, if you are using
the generic kernel, it is over 1M.  Third, your buffer cache may be
using some RAM.  Forth, what else is running, just ps?

Try using vmstat to see how many pageins and pageouts are being done.
Get more RAM and see if it helps.

>Any explanations at hand? Did we encounter a bottleneck between
>CPU and memory, or what? What performace should I expect from
>a SS10-41 or SS10-52 (which have a larger cache, but still not 
>large enough to hold the Lisp's resident set. And if I'm not
>mistaken, the large cache results in a longer time spent for
>non-cache memory accesses (6 cycles instead of 3)). What 
>performace should I expect from a HP735?

Get Sun and HP to send you a machine, don't wriite a $50k check
on a guess.  Even a good one.

>I can buy enough memory so that disk speed is no criterion.
>Should I wait for machines with 8MB of cache? Should I shoot myself?

How long can you wait?  Wait long enough and lisp machines will come
back into style (half :-) ).

>Should I give YOU the money? :-)

Yes, in small unmarked bills.
-- 
           stripes@pix.com              "Security for Unix is like
      Josh_Osborne@Real_World,The          Multitasking for MS-DOS"
      "The dyslexic porgramer"                  - Kevin Lockwood
We all agree on the necessity of compromise.  We just can't agree on
when it's necessary to compromise.       - Larry Wall