home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!cs.utexas.edu!geraldo.cc.utexas.edu!geraldo.cc.utexas.edu!usenet
- From: amigamat@ccwf.cc.utexas.edu (Mark Thomas)
- Newsgroups: comp.sys.amiga.hardware
- Subject: Re: Any News on the new Motorola Chip?
- Date: 25 Jan 1993 03:38:12 GMT
- Organization: The University of Texas at Austin, Austin TX
- Lines: 58
- Message-ID: <1jvnb4INNg64@geraldo.cc.utexas.edu>
- References: <hanscs.74.727431258@dhhalden.no> <C18o89.6MA@cs.columbia.edu> <1k1bn6INNau@newsman.csu.murdoch.edu.au>
- NNTP-Posting-Host: donald.cc.utexas.edu
-
- In article <1k1bn6INNau@newsman.csu.murdoch.edu.au> meek@csu.murdoch.edu.au (Lindsay Meek) writes:
- >won@cs.columbia.edu (Won Y. Kim) writes:
- >
- >>In article <hanscs.74.727431258@dhhalden.no> hanscs@dhhalden.no (HANS CHRISTIAN SANDER) writes:
- >>[stuff deletedp
- >>>>
- >>>The 68060 will be released(Amiga Computing) late 93. And will do aprox. 100
- >>>MIPS at 25 MHz.
- >>>
- >>Hmm, How can this be ?? Is the 060 a multi processor in one?
- >>At a clock speed of 25 Mhz, the most a "normal" cpu (that I am aware of)
- >>can execute at most 25 million instructions per second.
- >
- >>The # is usually smaller unless the code can utilize the pipleline 100 % (not
- >>bloody likely!!)
- >
- > They might have increased the width of the data bus. This would allow
- >/explain the 4x speed up at the same clock speed. However, the current state
- >of-the-art in VLIW (Very Large Instruction word) designs requires that you
- >have a customised code generator on your compiler to schedule the other three
- >instruction units correctly. I think some of the new workstations use VLIW (?)./
-
- 68060 key features:
-
- * Superscaler implementation of the M68000 architecture
- * IEEE compatible floating point
- * 8k instruction and 8k data caches
- * Branching target cache to minimize branch latency
- * Four entry write posting buffer
- * Bus snooping to maintain cache coherency
- * Full 32 nit non-multiplexed address and data busses
- * Low power mode for portable operation
-
- BTC (Branch Target Cache)
- The 68060 has a logically addressed 256 entry cache which associates branch
- instruction addresses with their corresponding branch target address.
- Correctly predicted taken branches execute in 0 clocks. Correclty predicted
- not taken branches execute in 1 clock.
-
- IEC (Instruction Execution Controller)
- Has dual integer execution pipelines which allow simultaneous instruction
- execution. Floatint point instructions are passed the the internal FPU.
- Most integer instructions are designed to execute in one clock.
-
- FPU (Floating Point Unit)
- The FPU is 100% compatible witht the 68040 programming model.
-
- At 50 MHz, the 68060 will offer more than three times the performance
- of a 25MHz 68040, but maintains code campatibilty.
-
- (Paraphrased from CTACS Monitor, the newsletter of the Central Texas
- Amiga Computer Society.)
-
- Mark
- --
- -------------------------------------------------------------------
- Mark A. Thomas / (amigamat@ccwf.cc.utexas.edu)
- CS Student : University of Texas at Austin
-