On 1999-03-04 [EMAIL PROTECTED] said:
>G'day Pete,
>> I had a chance to try windows 95 B on my computer whith 4 MB ram,
>>it worked, but was writting to the hard drive a LOT.
>I'm reporting you to the Prevention Of Cruelty To Computers Society;
>running Win95 in 4MB RAM is proof of your evil intentions towards
>defenceless old computers... ;-))))
Well, when windows 95 first came out minimum requirements wore a 386 CPU
or better and 4 MB ram or better.
>> I had to pull the new 4 meg simes I put in because I was having
>>all kinds of problems running windows, especially whith the
>>registry. The memory showed up in device maniger and mem in dos
>>knew about the new memory, but the new memory caused problems for
>me. [etc...]
>The "0" and "1" refer to memory _Banks_. A typical 386/486 machine
>using 30-pin SIMMS has two Banks, each with 4 SIMM slots. A 386sx
>has 2 SIMM slots per Bank. If you use a Bank, you have to fill it
>completely with identical capacity SIMMS. At the very least, fill
>Bank 0 or the computer won't start!
Yah, it beeped at me a bunch of times, then stoped and beeped some more when
bank 0 was empty. Thats how I knew for sure whitch was whitch. Bank 1 on
this computer is closest to the CPU and bank 0 is farthest from it.
>Although you can usually mix SIMMs from different manufacturers,
>and mix 3-chip with 9-chip SIMMs, in the one Bank; I have sometimes
>found that some SIMM mixtures will only work in a particular
>combination. For example, the four machines I have for my
>networking experiments have the following idiosyncrasies:
>- machine 1 is a 386-25 with four 256kB SIMMs in Bank 0 and four 1MB
>SIMMs in Bank 1, total 5MB. Swapping each lot of SIMMs to the other
>Bank results in a 4MB system (the 256kB "vanish" if in Bank 1). All
>SIMMs are from the same manufacturer and same speed.
>- machine 2 is a 386-33 with four 1MB SIMMs, two 9-chip from
>Toshiba and two 3-chip from Samsung, all the same speed and in Bank
>0. The Samsungs _must_ be in the first two slots for the computer
>to work at all.
>- machine 3 is a 486-25 with eight 1MB SIMMs from five different
>manufacturers, speeds of 60 and 70 nanoseconds. It doesn't matter
>where each SIMM is installed; this machine always finds 8MB even
>when it's running Linux.
>- machine 4 is a 386-40 with a strange SIMM slot layout which I
>found by trial and error: Bank 0 is slots 1,3,5,7 and Bank 1 is
>slots 2,4,6,8. So this machine has four 1MB SIMMs, each from a
>different manufacturer, in slots 1,3,5,7; and four 256kB SIMMs in
>the other slots for a total 5MB. If I swap each lot of SIMMs to
>the other Bank, *all* the memory "vanishes" and it won't start.
>It's no better with 72-pin SIMMs. The 586-100 I'm writing this on
>has two 8MB SIMMs, and they _must_ be identical or it won't start
>at all. This machine has 4 SIMM slots but only the first needs to
>be occupied (I started with one 8MB SIMM).
>You can also produce some amazing behaviour if you use SIMMs without
>parity in a machine that uses parity, or mixing parity/non-parity
>SIMMs in the same Bank. Not to mention EDO, SDRAM, Dual Pipeline...
>you might be interested in reading through the reviews at Tom's
>Hardware (http://www.tomshardware.com).
>I hope my examples illustrate how certain combinations of working
>SIMMs and working computers can fail to work when brought together;
>even when you and the computer shop know they should. So before
>you try any strange rituals involving burnt offerings etc to get
>the computer going; you might want to try experimenting with the
>possible permutations of SIMMs and slots. Don't forget anti-static
>precautions of course.
>cheers,
>Fraser Farrell
>http://www.dove.net.au/~fraserf/
Thanks!
Pete
To unsubscribe from SURVPC send a message to [EMAIL PROTECTED] with
unsubscribe SURVPC in the body of the message.
Also, trim this footer from any quoted replies.
