🚀 go-pugleaf

In article <exuptr.123.712797525@exu.ericsson.se> exuptr@exu.ericsson.se (Patrick Taylor) writes:
>>In article <1992Jul22.194413.7274@ncsu.edu> reward@eos.ncsu.edu (Robert E. 
>Ward) writes:>>
>>Cut in CPU power?  From My understanding... a 486 is a 386 with an onboard
>>387 and cache.  A 486sx is a 486 that has the 387 disabled, leaving you
>>with what is basically a 386 with a cache.
>>Now, this means he's comparing a 386 40Mhz with a 386 25Mhz with a cache.
>>I leave the decision as to which is better up to y'all...
>
>A 486 is not a 386 with anything.  The chip has been redesigned and 
>instructions execute in fewer clock cycles.  A 486 DX 33, for example
>runs non-floating point instructions about 2.2 times faster than a 386 DX 33.
>
>So a 486 25 with or without co-processor is going to outrun a 386-40
>almost all the time, unless you write code specifically against the 486's
>optimization.
>
>I have one each (386 33 and 486 33) side by side and the difference is
>about the same as between the 386 33 and my 286 25.  And soon with the
>DX2 chips you can pop a 50 MHz chip into the 25 MHz board and add 70% to
>your performance again.  Now that's a machine.
>                                                                                        "This must be Thursday.  I never could get the hang of Thursdays"                                                                     - D Adams
>  - Patrick Taylor
>    Ericsson Network Systems                                                        exuptr@exu.ericsson.se             "Don't let the .se fool you"                 alternately, exuptr@ZGNews.Lonestar.Org


This is all fine and dandy in theory. Yes a 486 performes operations in 1
clock cycle and a 386 performs operations in 2 clock cycles. It would
seem that a 486/25 would be like a 386/50 (if there was such a thing) but 
this is not the case. there is much more going on than just how fast the
processes does the operation like memory speed, etc. It actual
use, as documented in many magazines, the 386/40 is faster than the 486SX.
Test were done with Norton SI, powermeter, etc. which really mean nothing.
Also, calculations were performed using Lotus 1-2-3, documents formatted
using word processors, etc. The 386/40 was faster than the 486SX in all
cases.

-- 
----------------------------------------------------------------
|  kskapin@ic.sunysb.edu  (Keith Skapinski)  |    place        |
|                                            |  witty quote    |
|  BEEPER / VOICE MAIL SYSTEM: 516-341-3641  |     here        |
----------------------------------------------------------------