Conference hydra::amiga_v1

4100.0. "Update on the 68040" by KAHUNA::SUMNER () Tue Sep 11 1990 10:32

<><><><><><><><>  T h e   V O G O N   N e w s   S e r v i c e  <><><><><><><><>

 Edition : 2148               Monday 10-Sep-1990            Circulation :  8392 

==============                                  [Basingstoke, England          ]



 Motorola - 68040 chip delay sparks ire
	{The Wall Street Journal, 7-Sep-90, p. B1}
   Motorola wowed the industry in early 1989 with the 68040, only to hold up
 mass production of the chip repeatedly because of design flaws. Motorola now
 pledges to start shipping the chip in quantity in October, computer makers
 say, as much as nine months later than it first promised. The delay's biggest
 casualty is Hewlett-Packard, a loyal Motorola customer that planned to use the
 68040 in more than half its workstations by the end of the year. HP has
 watched its market share slide - blaming, in part, the elusive Motorola chip.
 "If we had the 040 today, we would probably be stronger in the workstation
 business," says Douglas Johnson, product-marketing manager for HP's
 Motorola-based computers. The tardy chip may also delay the shipment of a new
 machine by Next Inc. And smaller computer makers say they have delayed
 machines by as much as a year because of the belated chip. "The 040 fiasco
 certainly hurts Motorola's credibility," says analyst Robert Herwick at
 Hambrecht & Quist. "Most of the market has already thrown the 68000 [series]
 into the ground. This is the last dirt on the coffin." The main culprit in
 Motorola's lateness is its chip's complexity. A single 68040 contains 1.2
 million transistors and combines functions that require as many as 50 chips in
 most computers, says Jack Browne, director of marketing for Motorola, who
 calls the delay "pretty serious but not disastrous." Motorola was already
 late when it unveiled a prototype of the chip in January that it planned to
 ship in volume  by the middle of this year. But the early samples it sent out
 had scores of bugs. For instance, the chips wouldn't work without cooling
 devices. There were instructions that didn't work. Some software caused
 strange results or froze the chip up completely. By late spring, it was clear
 Motorola wasn't going to meet its schedule. This summer, Motorola was still
 sending out samples with a long "errata sheet," a list of known flaws in the
 chip, says Joseph Ramunni, president of Dallas-based Mizar Inc., which makes
 computer boards. "If, in fact, there are chips that do work, we sure aren't
 getting them," he says. Bugs are commonplace in early chip versions. Intel's
 best-selling 80386 chips were shipped in high volumes with many bugs, which
 were later corrected, and Intel's 80486 ran into well publicized bugs last
 year. But Motorola's bugs are showing up when the competition already has
 viable, "debugged" options on the market. Mr. Browne says Motorola is now "on
 track" and doesn't foresee more delays. "Now it's like the question of when
 Iraq's going to leave Kuwait," says Ron Stack, sales manager for General Micro
 Systems Inc., Montclair, Calif., which planned to sell a 68040 product last
 January. "Is it ever going to come out or what?"


<><><><><><><><>   VNS Edition : 2148      Monday 10-Sep-1990   <><><><><><><><>