wb@zeta.UUCP (Bill Beblo) (10/19/85)
I am currently in the process of looking rather closely at some of the "workstations" being offered by companies such as Sun, Apollo, Masscomp, Integrated Solutions, etc. Most of these vendors offer a system based on the 68020 chip. Some of the vendors claim their 68020 is clocked at 16.67MHz while others claim 12MHz. Although I realize that a faster clock rate in itself does not guarantee better throughput, an interesting dispute has developed regarding two of these vendors. Vendor A (16.67MHz clock) claims they have made special arrangements with Motorola and are paying a premium for 68020 chips certified by Motorola at 16.67MHz. Vendor B (12MHz clock) claims that Motorola is not currently certifying the 68020 at 16.67MHz and that Vendor A is in reality getting 12MHz chips from Motorola, but using only those chips which they can successfully push to run at 16.67MHz. I tend to believe that since the clock rate of 16.67MHz is printed on Vendor A's glossy brochure they are telling the truth, but Vendor B's claim does not seem absurd. I'm not an Electrical Engineer so I'm turning to the net for answers/opinions. A reply from a qualified party at Motorola would also be appreciated. Bill Beblo Bell Communications Research 290 W. Mt. Pleasant Ave. Rm 1B-112 Livingston, New Jersey 07039 (201) 740-4421
pavlov@hscfvax.UUCP (840033@G.Pavlov) (10/21/85)
Hewlett_Packard has released a workstation (HP9000 series 300)that utilizes
a 16.67 68020 chip. Since HP is very conservative re hardware, I think that
the question about "official" vs "unofficial" releases of the chip are moot.
g.pavlov, FSTRF, Amherst N.Y.john@hp-pcd.UUCP (john) (10/21/85)
<<<< < Vendor B (12MHz clock) claims that Motorola is not currently certifying < the 68020 at 16.67MHz and that Vendor A is in reality getting 12MHz chips < from Motorola, but using only those chips which they can successfully push < to run at 16.67MHz. < If Murphy is correct we can expect to see Motorola start sorting 16 MHZ parts for Vendor C and the number of fast parts shipped to vendor A will drop to near zero. That is a very dangerous practice. John Eaton !hplabs!hp-pcd!john
chip@cadsys.UUCP (Chip Krauskopf) (10/22/85)
The testing of microprocessors is a very interesting subject. To be
certified at a given speed the chip must perform at a given frequency and
meet all the stated specs. across a wide temperature range. It is one
thing for the part to work at room temp, but high temp is another. I would
find it very surprising for a chip company to sell parts that don't pass
there full test set at some speed as if they did. Likewise, if they
have 16MHz parts, I would think they would sell them as such! The only way
I know to "push" a non-full spec. chip is to cool it and/or run at higher
voltage.
--
Chip Krauskopf; 386 Design Team, Intel, Santa Clara, Ca.
{pur-ee,hplabs,amd,scquaxd,dual,qantel}!intelca!cadsys!chip
N.B. The views above are personal and may not represent those of Intel.