Re: 14MHz 65816??

[lesherjt@vuse.vanderbilt.edu (Tilghman Lesher)]

Nathan Mates (nathan@visi.com) wrote:
: In article <4v3442$5qm@news.vanderbilt.edu>,
: Tilghman Lesher <lesherjt@vuse.vanderbilt.edu> wrote:
: >Tony Cianfaglione (tony@ccn.cs.dal.ca) wrote:
: >:    I have been in contact with Western Digital regarding their 14 MHz 
: >: 65816 chip which they tell me fits onto the accelerator cards such as the 
: >: TransWarp GS or Zip GS.  They were not sure if it would replace the stock 
: >: GS chip that plugs into the motherboard, which is what I want to do.  I 
: >: find the various accelerator cards too buggy to rely on.

: >Nope; can't do this.  Remember that the regular 2.8 MHz chip slows down
: >to 2.6 MHz because of the 1 MHz bus speed.

:    Argh. No. Wrong. The processor internally runs at 2.8Mhz, but due
: to ram refresh cycles, the effective speed is about 2.6Mhz. Things are

Whatever.  It's called an effective abbreviation where an explanation
works for simple users, even if the explanation is not technically
correct.  Such is used by many people (try the Perl instruction booklets)
to try to explain complex ideas in simple language without confusing
learners.

: dropped to 1Mhz under the following conditions: 1) User requests it
: (Control panel or via software). 2) "Slow" ram is accessed (banks E0,
: E1, which can happen transparently due to shadowing. 3) Softswitches
: (//e speaker, 5.25" disk access) are hit, which are just a special
: case of 2.

: >The accelerators are able to work that fast because they use a type of
: >memory called "cache", which quite simply is, fast memory.  The cache
: >is "mirrored" back to regular RAM ever so often, so that memory stays
: >intact.

:    I've already explained this on this thread. Cache is a small (8-32K)

Your messages seem to disappear.  Does your newsreader cancel articles
on the same thread when you post followups?  If so, maybe you need another
newsreader which doesn't do this.

: chunk of memory holding recently accessed memory, which can be read at
: high speeds. Writes are always done asap, not "every so often."

My point was that the refresh is done on a regular cycle (IOW, at a 
specific, predictable rate, which happens to be "as soon as possible")

: >If you wanted to use an accelerated chip on the motherboard, you'd have
: >to modify the motherboard to speed up the bus, get faster RAM, and hope
: >to God that it works (probably won't).

:    As I said earlier in this thread, you'd have to redesign the
: motherboard from scratch to be able to plug in a faster 65816 and
: hope it ran. And the odds of that happening are slim and none.

You don't think a motherboard could be modified to do it?  You have a
very low opinion of the abilities of EE's.  YES, IT WOULD BE DIFFICULT.
YES, IT PROBABLY WOULDN'T WORK ON THE FIRST TRY.  But that's a far
cry from saying that it cannot be done.

: Nathan Mates

Tilghman