[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: ZIP Question....
In article <Pine.3.07.9312280843.A1527-a100000@cap.gwu.edu>
dpfarrer@CAP.GWU.EDU (Daniel Pfarrer) writes:
> What is the fastest ZIP Mhz rate, I called them a few months ago, I
> remember it being something like 16Mhz for $659, correct me if i'm wrong.
I'd be surprised if ZIP Technology were selling ZIP IIgs's rated as
fast as 16 MHz. They might be offering 14 MHz, but the last I heard
(several months ago) is they have stopped offering cards faster than
10 MHz. If you want to go faster than that, you'd have to do it
yourself (buy the faster microprocessor, crystal and cache RAM).
If you want to go faster than 12 MHz, you have to modify the ZIP card.
Going faster than 14 MHz is tricky. There is a message posted somewhat
irregularly to this newsgroup which discusses the TransWarp GS and ZIP
GS.
> One more thing, could someone explain what "cache" means?
According to my dictionary: "A hiding-place for treasure, provisions,
ammunition, etc.; or stores so hidden." :-)
In computer terms, a "cache" is an area of memory used to hold
frequently accessed data.
There are two main types of caches: disk caches and memory caches.
A disk cache is just an area of computer memory used to hold data read
from disks. GS/OS provides a disk cache, which is mainly used to
store directories (since they are the most frequently accessed parts
of disks). The RamFAST SCSI card contains up to a megabyte of RAM
used to cache data read from the disk.
A memory cache is used in systems where the memory (RAM and ROM) is
not able to provide data as fast as the processor requests it. The
cache is a special type of fast memory which can be accessed by the
processor at full speed.
The cache holds the most recently accessed locations, the assumption
being that the computer is likely to access them again (which is
especially true of loops in programs).
When the processor requests a certain memory location, the cache
checks to see if that location has been stored. If so, it is provided
immediately to the processor. If not, the location is fetched from
main memory (a slower operation), stored in the cache for future
reference, and passed to the processor.
There is a limited amount of space in the cache, so data which has not
been accessed recently is discarded in preference to new data as it is
read into the cache.
In the case of the ZIP GS, you can have 8k, 16k, 32k or 64k of cache
RAM. The cache RAM runs at the same speed as the processor (e.g. in
an 8 MHz ZIP, the cache is able to keep up with accesses at 8 MHz).
The memory in the IIgs is not able to be accessed faster than the
original speed of the IIgs, 2.5 MHz or thereabouts (1 MHz for the
"slow" areas such as the I/O locations and video buffers).
The ZIP's cache works as I described above - if a requested location
is in the cache, it is returned to the processor immediately.
Otherwise, the ZIP has to slow down to 2.5 MHz or 1 MHz and read the
desired location from main memory.
In ideal circumstances, where the code being executed is fully in the
cache (e.g. a tight loop), the ZIP will be able to run at or near its
rated speed (e.g. 8 MHz). Every non-cached memory read (and all
memory writes) require the ZIP to slow down to the speed of the IIgs,
resulting in a slower average speed of operation.
This is why a ZIP rated at 10 MHz doesn't necessarily run four times
faster than an unaccelerated IIgs. It could theoretically run four
times faster, but only in ideal circumstances. Depending on the
operation being performed, it would run anywhere between 100% and 400%
of the standard IIgs speed (probably averaging to 300% for
calculations, or 200% for graphics operations).
The amount of cache RAM on the ZIP makes a difference. With more
cache RAM, more data can fit into the cache, and the ZIP will be able
to run closer to its maximum speed.
I hope that answered your question.
--
David Empson
dempson@swell.actrix.gen.nz
Snail mail: P.O. Box 27-103, Wellington, New Zealand