hmmm.... it's been years since I've played with programming under MS-Win~1
for multi-proc systems.... I think the OS will obviously take advantage of
the multi-proc (and the OS scheduler may schedule separate processes of apps
on different processors), but I think (if I recall) the applications have to
sense and use (and set ProcessorAffinity) for its respective *threads*, with
those types of system calls to the kernel in order for the threads to take
advantage of the multi-proc (Winnt and Win2k of course).

I do not know if Protaltium implements that or not....  And there is
probably some extra shared memory management and perhaps marshalling of
resources that would need to be written (which is fodder for bugs and
crashing! so hey, maybe they did ;-)  just kidding)...

Does anyone else know off the top of their heads about the
Processor_Affinity stuff? I think this is correct... (It's been a while)

I know for example, my multitrack recording software does purposefully
implement this stuff by hook or by crook for the threads and plug-in effects
running on different processors.

The multitrack recording software boasts about only a ** 35% ** increase in
performance with a dual processor system versus the single processor system,
pound for pound (or dyne for dyne  or slug for slug......   ;-).  But a dual
processor system is better for multitrack type software to keep those
separate threads on different processors in pseudo/apparent in-process dll
attachment type configuration (without having to enter/exit from the host
CPU to another machine to process things at high data rates)...

I think 2 computers might be better (but more costly???) as a second
computer at the same speed would be 100%...  but then you have file
integrity, revision control, possible merge-ification issues and multiple
copies of things.... but you may gain some crash safe redundancy...

I wonder how this would behave on the new Dell's dual 1.7GHz processor
400MHz memory bussed RDRAM machines? (which I suppose in theory brings the
bus speed to 200 MHz for each processor in theory... which is not much more
than 133MHz....).... I'd wait for 800MHz busses....

shoe shine boy
assistant bottle washer

-----Original Message-----
From: Andrew J Jenkins [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, August 07, 2001 7:49 PM
To: Protel EDA Forum
Subject: Re: [PEDA] Signal Integrity dialog locks up.

On 01:20 PM 8/7/2001 -0700, Brad Velander wrote:
>         welcome to the Protel(ooops, Altium) club. We all meet here
>regularly to lick our wounds, share war stories and generally commiserate
>about our problems. You will make a suitably good initiate candidate, you
>seem to have passed all the tests. 8^>
>         Seriously, I would warn you against looking for wonderful
>performance increases from going to dual processors. There are things to be
>said for dual processors but it doesn't necessarily make Protel that much

I will not respond...I will not...I will not....Arghhhhhh.....

The point about dual-proc systems is that they do not experience the
latencies single-proc computers do experience, routinely, btw, NOT that they
do the same job quicker (at least for programs that aren't written to take
advantage of the second proc.)

What they DO offer is a way out of sitting around while the computer is
locked up tight, at least allowing the user to fire up Eudora and write a
quick rant against Protel, do a net search for component data, write a
proposal (or that damned monthly report), open a second instance of Protel,
(not generally recommended, but I've done it to my advantage in the past.)
or otherwise utilize the PC whilst the resource hog blindly grinds away at
its task, happy in not knowing that it hasn't succeeded in stopping all
other work.

I say it again. Brad, you have to try personally running under a dual-proc
before judging this subject. Admittedly, IFf you can afford dual computers,
that's one way to go. But it won't solve the latency problems. 100% is 100%,

Ever auto-routed two boards simultaneously on a single computer? It CAN be
done. Two instances. Isolate 1st to Proc 1, second to Proc 2.



* * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
* To post a message: mailto:[EMAIL PROTECTED]
* To leave this list visit:
*                      - or email -
* mailto:[EMAIL PROTECTED]?body=leave%20proteledaforum
* Contact the list manager:
* Browse or Search previous postings:
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * *

Reply via email to