🚀 go-pugleaf

In article <60570032@hpcuhe.cup.hp.com> ken@hpcuhe.cup.hp.com (Kenneth M. Sumrall) writes:
>What is the difference between "Static Column DRAM", "Nibble Mode DRAM", and
>"Fast Page Mode" DRAM?

Internally, a DRAM is a matrix -- usually square -- of cells, and a read
operation internally reads out an entire row.  Folks noticed that it was
kind of wasteful to read an entire row out and then supply only one bit
to the outside world.  Of course, if you're reading bits at random, there
isn't much that can be done about it.  But if you're reading them in
sequence, getting the next bit out of an already-read row is a lot
quicker than reading a new row.  Several ways of exploiting this were
invented.

In a static-column DRAM, once the row read has been done, you're looking
at something that behaves much like a static RAM:  change the address
lines, and after a specified access time, that bit shows on the output.

In a page-mode DRAM, you have to clock the new column address in with
one of the control signals to make things happen.  Fast page mode is
just page mode implemented well; early page-mode implementations often
were rather slow.

In a nibble-mode DRAM, you feed the first column address in by the usual
method, and then after that, just wiggle one control line and it increments
automatically.  Usually there's only a two-bit counter, so you can get a
maximum of four bits out this way, hence "nibble mode".

>Also, what is "Write Per Bit", and why do I want it?

DRAMs whose outside-world interface is wider than one bit have only one
set of control signals for all N (usually N=4) bits.  For many purposes,
this is fine because you read/write all N bits at once.  There are a
few applications that really want finer control.  Consider a 32-bit-wide
memory implemented with 4-bit-wide DRAMs.  Okay, that's 8 chips.  Now add
one parity bit per byte.  That's 4 bits, one more chip... except that if
the system does byte writes into the memory, you need to be able to change
one of those parity bits without changing the others.
-- 
"God willing... we shall return."       | Henry Spencer @ U of Toronto Zoology
       -Gene Cernan, the Moon, Dec 1972 |  henry@zoo.toronto.edu  utzoo!henry