home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: sci.electronics
- Path: sparky!uunet!wupost!spool.mu.edu!agate!usenet.ins.cwru.edu!neoucom.edu!wtm
- From: wtm@uhura.neoucom.edu (Bill Mayhew)
- Subject: Re: 486DX2 Crystals
- Message-ID: <1992Dec13.023819.9476@uhura.neoucom.edu>
- Organization: Northeastern Ohio Universities College of Medicine
- References: <1g4ul8INN2em@violet.csv.warwick.ac.uk> <1992Dec12.003507.1@research.ptt.nl>
- Date: Sun, 13 Dec 1992 02:38:19 GMT
- Lines: 29
-
- 486DX2 chips use what is essentially a phase locked loop circuit
- and 2* oscillator internal to the 486DX2 to synchronize internal
- operations to the external environment running at regular speed.
- As you might suspect, a clock doubled 25 -> 50 MHz 486DX2 performs
- many benchmarks faster than a 33 MHz 486DX. The clock doubling
- inside the chip is offset by the fact that the system clock is only
- 25 MHz.
-
- Several of the popular trade press publications have performed
- comparisons. There are some non-compute-bound scenarios where a
- standard 486DX-33 outperforms a 486DX2-50 (file servers, for
- instance).
-
- The moral of the story is to know what you are you are buying and
- to what the "MHz" rating refers. The system clock or the CPU
- doubled clock?
-
- There are many other factors such as local bus video, memory
- interleaving, secondary cache, EISA vs ISA vs MCA bus, disk system,
- etc. that effect performance. It is difficult to make a
- performance evaluation of a system without performing a task mix
- that models the actual working environment you intend for the
- computer.
-
-
- --
- Bill Mayhew NEOUCOM Computer Services Department
- Rootstown, OH 44272-9995 USA phone: 216-325-2511
- wtm@uhura.neoucom.edu (140.220.1.1) 146.580: N8WED
-