DIMMS and SIMMS: So what's the difference.
Edmond Underwood
faster than SIMMs, but by what margin. Example: if I bought a 64-bit 60ns
DIMM, how much faster would it be than a 60ns 32-bit SIMM. On the average.
---------------------------------------------------------------------------------
Edmond Underwood
Systems Management Group
Computing & Network Services (University of Colorado)
E-mail: unde...@Colorado.Edu
Robert C. Barris
Edmond Underwood <unde...@Colorado.Edu> wrote:
>Simple question if anybody has the real answer. DIMMs are said to be
>faster than SIMMs, but by what margin. Example: if I bought a 64-bit 60ns
>DIMM, how much faster would it be than a 60ns 32-bit SIMM. On the average.
>
I don't know who said they were faster? Chips is chips, DIMMs are just
another packing form factor for the same basic product.
My perception is that DIMM's will become more standard as PowerPC and
Pentium systems expand to take the lion's share of the personal computer
space (both using at least 64-bit data paths); just as it was getting a
little gross to have to install 8-bit wide SIMM's four at a time for a
32-bit data path, it would also be clunky to have to install 32-bit
(72-pin) SIMMs for a machine with a 128-bit backbone.
It boils down to convenience (as I see it) and reducing MB costs by
having fewer individual sockets etc... there is a speed issue involved
somewhat, since if it's easier and cost effective to design a machine
with such a wide bus, then your cache misses and what not can be
satisfied that much quicker (depending on memory controller design).
Rob Barris
Quicksilver Software Inc.
rba...@quicksilver.com
Jim Dresher
rba...@orion.oac.uci.edu (Robert C. Barris) writes:
> >Simple question if anybody has the real answer. DIMMs are said to be
> >faster than SIMMs, but by what margin. Example: if I bought a 64-bit 60ns
> >DIMM, how much faster would it be than a 60ns 32-bit SIMM. On the average.
> >
>
> I don't know who said they were faster? Chips is chips, DIMMs are just
> another packing form factor for the same basic product.
I don't know exact figures, but DIMMs support some sort of memory
interleaving which boosted speed 5-10% in the Quadras that supported
it. Could see similar figures for DIMM saavy PowerMacs. DIMMs are
64-bits and use lots more pins, so you'd think they were better, else
why would they be moving in that direction.
--
Jim Dresher
jdre...@xmission.com
OpenDoc is going to change where you want to go tomorrow.
Edmond Underwood
>another packing form factor for the same basic product.
>My perception is that DIMM's will become more standard as PowerPC and
>Pentium systems expand to take the lion's share of the personal computer
>space (both using at least 64-bit data paths); just as it was getting a
>little gross to have to install 8-bit wide SIMM's four at a time for a
>32-bit data path, it would also be clunky to have to install 32-bit
>(72-pin) SIMMs for a machine with a 128-bit backbone.
>It boils down to convenience (as I see it) and reducing MB costs by
>having fewer individual sockets etc... there is a speed issue involved
>somewhat, since if it's easier and cost effective to design a machine
>with such a wide bus, then your cache misses and what not can be
>satisfied that much quicker (depending on memory controller design).
That is what I thought, but I keep seeing on-line here and in some Mac
magazines that they are faster. I perosnally think that vendors will
move to EDRam. It is everything EDORam is supposed to be at 15ns and
faster. Sounds impressive to me.