Robert E. Watts wrote:
> HI Gang !
>
> Inserting comments as I go..... as usual....
>
>
> "metronid" <> wrote in message
> news:9cad7112-4fc0-4e21-b521-...
> On Jul 17, 2:20 pm, Ben Myers <ben_my...@charter.net> wrote:
>> Kenn Caesius wrote:
>>> Thank for the information, Ben Myers.
>>> In a addition to my original post, would you happen to know if I could
>>> expect a noticeable performance gain if I do install the additional
>>> cache memory?
>
> Ah, a person who thinks just like I do.
> Yep, L2 cache is important, especially if you have NONE like ( for reasons
> that escape everyone ) Packard Bell *usually* provided.
> I have a few Packard Bell's with L2, and surprisingly, (only 256K, as
> opposed to everyone else using 512k on most Pentium boards. ) the difference
> is not that great.
> I always suspected that PB did "something" in BIOS that made their machines
> run pretty good without it. Don't know what it would be though..........
> I have searched out PB's with L2 without much success, although like I
> stated, I do have a few with it.
>
>
>>> ---End of message---
>> The general rule of thumb is that the faster processors benefit more
>> from cache memory than the slower ones. For example, a 486-33 or 486-25
>> would not show much performance improvement at all, whereas a 66MHz
>> 486-DX2 or a 100MHz 486-DX4 (or 83MHz Pentium OverDrive) would.
>>
>> Also, going from no (zero K) cache at all (as shipped on some PB
>> systems) to ANY size of cache will result in the largest performance
>> gains... Ben Myers
>
>
> Good idea to read what Ben types carefully, it's always good information.
> :-)
>
>
>
> Hello I am here but have to limit my time.
> I am still recovering and will be doing so for a while.
>
>
> From what ?
> Just curious. Hope you're doing well!
>
>
>
> I had a PB 450mb system legend 10CD
>
>
>
> I did use the above memory as it was cheap and available last
> century.
> It worked some what better with the original 486 processor.
> It worked better and best with 486 100mhz as opposed to the 83mhz
> POD
>
> In fact I was not that impressed with the pentium 83 which was
> really the
> 63 that was clocked at 83 by me.
>
> One note it did run much better when I used the modified bios that
> I rewrote.
>
>
> Which more or less confirms what I said earlier. The way a BIOS is written
> can make a difference in performance.
>
> Also, the POD usually disables the L2 on PB motherboards. Dumb.
> PB did so many DUMB things.
>
>
>
> I did not actually write it I hijacked it and made some
> modifications.
> If my memory is still active it would allow for a HD of around
> 4gb.
> It may have just ran better because the added HD was not running
> in compressed mode.
>
>
> I know someone is going to say the heat with overdriving.
> I believe that Intel just labelled them each as such and they
> were maybe
> the same CPU.
> Maybe not the 1st run but after a few runs they knew it was
> cheaper to use one
> assembly line as opposed to runiing a 2nd that was not needed .
>
>
>
> Nothing so tricky. Intel merely labled them 83 or 63 MHz for the "masses".
> Most people can't figure out bus speed, so it was easier if people knew if
> their computer ran at (16), 20, 25, 33, 50, or 66MHz which Overdrive POD to
> buy.
> You and I didn't care, just cranked up the bus speed and let 'er rip.
> Ever tried the POD at 100MHz ? ( 2½X40MHz ) Runs pretty good !
>
>
>
> I am still able to ramble perhaps I may be getting back to para
> normal.
>
> Nothing wrong with rambling.
>
> :-)
>
>
The Socket 3 Pentium OverDrive was Intel's worst OverDrive disaster
because it was designed to run with a write-through external cache
rather than write-back, and probably 0.00005% faster than if it had been
designed for write-through. Prior to the annoucement and general
availability of the POD, almost all of the 486 boards ever built had
either write-back cache or no cache at all (like many PB models).
Installing a POD on one of these write-through boards made the system
incredibly slow. Intel also had an interposer board that would sit
between the POD and the Socket 3, effectively disabling the write-back
feature, enabling write-back instead. But interposers were always
extremely difficult to come by.
There was a class action suit against Gateway for advertising and
delivering "Pentium Ready" 486 systems. The suit was settled by paying
off all the lawyer suits and giving class action participants a couplon
good for $50 (???) off on the same stupid Pentium OverDrive that would
not work on their systems.
I sold hand-assembled kits using a 133MHz AMD 486 workalike, a 486
socket interposer, cooling fan, and a write back/write through jumper.
I never had a customer return one for lack of satisfaction. Evergreen
and Kingston also sold kits, far more than my 200 or so kits. I also
had a bundle with cache chips and nicely illustrated documentation. I
still have some of this stuff around for chump change, if anyone wants
to torque up their PB system. Not sure how much I have in the pile of
archeological remains from previous computer civilizations... Ben Myers
|
Similar Threads
Linear Mode
