[comp.sys.ibm.pc.hardware] Question: How fast must memory for 12Mhz At be ?

craig@ernie.Rosemount.COM (Craig Taylor) (05/21/91)

I have some old 150ns 256x1 DRAMs and was considering picking up
a 12Mhz AT motherboard and using them.  What speed must the DRAMS
be for a 12Mhz AT?  Are wait states usually jumperable on the mother-
board?  Is there a simple formula to use to determine the required
speed?  Can slow memory chips be used as extended memory on
a RAM card with a fast AT?  Any help appreciated.

Craig Taylor
craig@ernie.rosemount.com

phil@brahms.amd.com (Phil Ngai) (05/23/91)

craig@ernie.Rosemount.COM (Craig Taylor) writes:
>I have some old 150ns 256x1 DRAMs and was considering picking up
>a 12Mhz AT motherboard and using them.  What speed must the DRAMS
>be for a 12Mhz AT?  Are wait states usually jumperable on the mother-
>board?  Is there a simple formula to use to determine the required
>speed?  Can slow memory chips be used as extended memory on
>a RAM card with a fast AT?  Any help appreciated.

I have some gasoline and was considering buying a four door car
to put it in. For some reason, I don't want to ask the person
selling me the car about technical and highly variable information
so I'm going to ask the world how much gasoline I should put
into the car. Oh, and I won't tell you which car it is either.
It might be a 4 cylinder car or an 8 cylinder car. It might
have a radio, it might not.

If the gasoline isn't right for the car, can I use it in a motorcycle?

--
For the Welfare system to flourish, its clients must not.
Conflict of interest?

whitney@reed.cs.unlv.edu (Lee Whitney) (05/29/91)

To answer your question, you would be better off using your RAM for an
XT (8086 based) computer.  150ns is generally too slow for a 286
computer.

To Phil who responded to you:

If you are saying this was a stupid question, count how many times you
have asked a question someone already knew the answer to, and
reconsider your judgement.  Just because you may have already learned
something does not mean that it is then stupid for anyone else to
learn it.  If everyone on the net tries to help a little rather than
mock someone trying to learn, we can have most of our questions
answered.

If you did not have intentions of ridicule, please disregard the above
note and enlighten us as to the meaning of your wit.

phil@brahms.amd.com (Phil Ngai) (05/29/91)

whitney@reed.cs.unlv.edu (Lee Whitney) writes:
>To answer your question, you would be better off using your RAM for an
>XT (8086 based) computer.  150ns is generally too slow for a 286
>computer.

Gosh, I guess C&T must be lying when they say on the front of their
CS8221 NEW ENHANCED AT DATA BOOK that they support 12 MHz operation
with 150 ns DRAMs. But then, you probably don't know what page mode
or interleave operation are. In fact, I'd be willing to bet a dollar
that you've never even read the C&T NEAT data book.

>If everyone on the net tries to help a little rather than
>mock someone trying to learn, we can have most of our questions
>answered.

Yes, and with answers like yours, whose relationship with reality is
only accidental. Any answer is better than no answer, whether or not it
is correct?

--
The media is in the business of distorting people's perception of
reality, by emphasising the out of the ordinary.

whitney@reed.cs.unlv.edu (Lee Whitney) (05/30/91)

whitney@reed.cs.unlv.edu (Lee Whitney) writes:
>>To answer your question, you would be better off using your RAM for an
>>XT (8086 based) computer.  150ns is generally too slow for a 286
>>computer.

Phil writes:
>Gosh, I guess C&T must be lying when they say on the front of their
>CS8221 NEW ENHANCED AT DATA BOOK that they support 12 MHz operation
>with 150 ns DRAMs. But then, you probably don't know what page mode
>or interleave operation are. In fact, I'd be willing to bet a dollar
>that you've never even read the C&T NEAT data book.

Did I say 150ns RAM would not operate in any 286?  No.  I said 150ns
RAM is generally too slow for a 286, which I still stand behind.
Consider this:

     1.)  Not everone has a 286 board with the features you speak of,
          and on many 286-12 boards, 150ns memory CAN and HAS caused
          problems.

     2.)  Even if the board does have interleaved memory (which, by
	  the way, your on for any monetary bet you would like to
	  wager on my knowledge of this subject) this DOES NOT
	  GUARANTEE that you will get the best performance from your
	  system.  If simple memory tricks like that were the end all
	  answer, we would not need more advanced schemes to reduce
	  memory access such as caching and RISC's register based
	  architecture. 

Under these circumstances I feel the original answer had a little more
than "an accidental relationship with reality".



>>If everyone on the net tries to help a little rather than
>>mock someone trying to learn, we can have most of our questions
>>answered.

>Yes, and with answers like yours, whose relationship with reality is
>only accidental. Any answer is better than no answer, whether or not it
>is correct?

My point was not argue answer validity with you, although I will if
your goal is to contribute to everyone's knowledge.

My point was that even though you are someone who has apparently
attained some degree of technical knowledge, instead of helping to answer
a simple question, or ignoring it, you chose to try to belittle and
make fun of a less experienced user.  I cannot see how your initial
response was helpful, or even funny.

phil@brahms.amd.com (Phil Ngai) (05/30/91)

whitney@reed.cs.unlv.edu (Lee Whitney) writes:
>     1.)  Not everone has a 286 board with the features you speak of,
>          and on many 286-12 boards, 150ns memory CAN and HAS caused
>          problems.

Maybe we should have a show of hands. How many people with a 286 have
a C&T NEAT chipset?

Oh, I see. NEAT is a really rare and unusual chipset and hardly anyone
has it.

These people who had trouble, they weren't the kind of people who don't
bother to consult with their vendor, are they?

>     2.)  Even if the board does have interleaved memory (which, by
>	  the way, your on for any monetary bet you would like to
>	  wager on my knowledge of this subject) this DOES NOT
>	  GUARANTEE that you will get the best performance from your
>	  system.  If simple memory tricks like that were the end all
>	  answer, we would not need more advanced schemes to reduce
>	  memory access such as caching and RISC's register based
>	  architecture. 

The way I see it, two answers have been provided here, mine, which is
to the effect that "it's complicated and depends on a lot of stuff,
your vendor is the best source of info", and yours, which is "no, it
won't work. well, ok, after Mr. Ngai complained about my answer, I'll
change it to be sometimes it works and sometimes it doesn't.  actually,
it's complicated, but I think you shouldn't do it"

>I cannot see how your initial
>response was helpful, or even funny.

What can I say, other people did.

--
The media is in the business of distorting people's perception of
reality, by emphasising the out of the ordinary.