SI
SI
discoversearch

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor.  We ask that you disable ad blocking while on Silicon Investor in the best interests of our community.  If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.
Pastimes : Dream Machine ( Build your own PC ) -- Ignore unavailable to you. Want to Upgrade?


To: Zeuspaul who wrote (299)2/27/1998 11:35:00 AM
From: Michelino  Read Replies (1) | Respond to of 14778
 
Zeuspaul,

Thanks, this is very useful advice and some good links. Letting someone else burn-in a starting configuration and building up from there is attractive to me. I too, have thought of going with one processor to start. And multi-threading aside, it is still the OS that determines how efficient (or existent!) multi-proccessing is.

I've seen figures that say adding a second processor results in only a 60-70% increase in computing power. But that still adds enough additional independent data-crunching ability to suit me.

About how long should one expect to wait before the BTX motherboard bugs are addressed?

Regards,
Michael



To: Zeuspaul who wrote (299)2/27/1998 7:40:00 PM
From: Ken  Respond to of 14778
 
Zeus,

I agree completely with your comments on ground-up building. One advantage I've found is that building a machine from a table full of parts removes a lot of the mystery. I now know that I can fix (or at least diagnose) just about anything that goes wrong with my computer. After that first assembly project works, suddenly you realize that it is really simple. I am amazed at how many friends pay CompUSA to install memory or hard drives.

It really is just plugging things together. And when it doesn't work ---- well that is a learning opportunity. Some people just aren't interested in the nut and bolts aspects and think they avoid problems by just calling Dell or Micron and ordering a fully-configured machine. Maybe, but a friend just went through a four week delay involving numerous phone calls trying to get his dual-processor monster machine (used for video editing) running. Finally had to send it back to the factory. Turned out to be a loose cable and bad SCSI card. If he had built it himself, he probably could have diagnosed and fixed it. He may not have saved any money, but at least he would have learned something.

Ken



To: Zeuspaul who wrote (299)2/27/1998 9:20:00 PM
From: Zeuspaul  Respond to of 14778
 
Dual Processors....RAM is SHARED according to this poster

I asked this question on the Intel PII motherboard forum

zeuspaul wrote in message ...
>>If you have two CPU's and 128 MB RAM...How is the RAM divided? Does NT split the RAM between the two processors yielding an effective 64 MB RAM per processor? If your application is not multithreaded are you then working in a 64 MB environment even though you have 128 MB on the motherboard?<<

I received this response
newsgroups.intel.com

>>From
"Nima Ahrari"
Date Sent
Fri, 27 Feb 1998 13:56:37 -0700
Subject
Re: Two CPUs -how do they use RAM
Organization

The RAM is not divided, but shared between the two CPUs no matter
how many CPUs you have. Both CPUs have access to the RAM and
other motherboard resources and just share them. The CPUs just
work together (at least that's the idea!) to divide the jobs if
the software was written with multithreading in mind. If the
software was not specially written for a multithreading platform,
the only performance gained is usually in operating system related
tasks such as copying files, openning files, reading files, etc.
so you will still see some improvement (in NT only). If the software
was specifically written for multithreading platform, then you
can expect performance increase in the range of 50 to 100% or even
more in some cases depending on the type of application.

-- Nima<<