home *** CD-ROM | disk | FTP | other *** search
- Path: news.mountain.net!usenet
- From: gene_heskett@wvlink.mpl.com (Gene Heskett)
- Newsgroups: comp.sys.amiga.hardware
- Subject: Re: Amiga vs. PC
- Date: 30 Mar 96 11:25:41 +0500
- Organization: MountainNet, Inc. Morgantown WV 800.444.1458
- Message-ID: <2862.6663T685T2067@wvlink.mpl.com>
- References: <sysop.02o5@sweetdreams.lahn.de>
- NNTP-Posting-Host: slip1.mpl.com
- X-Newsreader: THOR 2.22 (Amiga;TCP/IP)
-
-
- MB> On Mon 4-Mar-1996 19:57 Gene Heskett wrote:
-
- GH>> software. Hitachi picked up the ball and gave us the equ of the
- GH>> NEC V20-30-40 series of replacement cpu's for it, but contracts
- GH>> prevented them from admitting or advertizing that their chips
- GH>> were faster, better, and lower power by several orders than the
- GH>> Motorola version.
-
- MB> You mean the intel version, right? Motorola manufactures 486
- MB> clones for AMD right now in their british fabs but AFAIK never
- MB> made any x86 for intel.
-
- Not no, but HELL NO! I said, Matthias, and meant, Motorola. Hitachi
- had perms to build the chip, not from the Motorola masks, but from
- the emulation. They could do it any way they saw fit as long as it
- could drop into the Motorola socket and run without *any* software or
- hardware modifications. And thats *all* they could advertise
- legally, even at home in Japan, even now after nearly a decade.
-
- When I said "the equ" I was refering to the level of "improvements"
- NEC did to the Intel stuffs, but applied to the early Motorola chips,
- specifically the Hitachi equ of the Motorola 68x09EP, the Hitachi
- HD63x09EP. In a world of chips with an 8 bit data buss, the motorola
- was 16 bits internally, the Hitachi is 32 bits in some aspects, had a
- one cycle deep prefetch pipelined instruction cache, several more
- registers, and extremely optimized microcode for its instructions as
- opposed to the Motorola hardwired version. The results could at
- times be *very* surprising, like a 32/16 divide with 16 and 16
- results. In 37 clock cycles worst case! Using the Motorola 68x09,
- that can be done of course, but the clock cycle count is about 100x
- greater! This is in fact one 4 byte command... The NEC equ's to the
- Intels (V20 etc) never came close to doing such "improvements"
-
- Setting its "mode" register non-zero puts in in the "native" mode,
- where the prefetch is switched on, and that alone is an average of a
- 20% speed improvement. The operating system called os9, originally
- done for that chip, and which proved you *can* effectively multitask
- and multiuser a "game machine", is now running at least 2x the original
- speed thanks to taking advantage of the 63x09's new instructions.
- One of them is sitting on the next desk running right now, its called
- a TRS-80 Color Computer 3, has 2 megs of ram, a 130 meg hard drive, a
- modem, 3 floppies, 2 monitors, shares the printer with this Miggy,
- has a seriel mouse just like the big boys, and all the other
- appertanances of a *real* computer. Some of them are even online
- here on the net! 24 hours a day.
-
-
- /* Gene Heskett | I need help, please see my adv. */
- /* CE @ WDTV Weston/Clarksburg WV | in the newsgroup 'alt.tv.jobs' */
- /* <gene_heskett@wvlink.mpl.com> | We are an equal opportunity and */
- #include <std.disclaimer> | all that stuff employer.
-
-
-