cliffhanger@cup.portal.com (Cliff C Heyer) (02/07/90)
Re: Digital Review 1/29/90 P. 1 "DEC Set For Move into Board Level Market" I can't believe my eyes. I don't see how DEC can hope to compete in this market! With 8+ MIPS 68030 boards going for $2500. And the new Heurikon 68040 25MHz board at 13.5 MIPS going for $3400. Lets note DEC's KA650 MicroVAX 3000 board goes for $10,000 USED - for a paltry 4 MIPS. I guess the new price is about $12000... This uses the 78034 uP. Oh - but wait! DEC's new SBC "board level product" is not the KA650, it's the "RTVAX 300" offshoot of the KA620 priced in the $2,000 range. At 3.5 MVUPS (4.8 MIPS) Note that this is $416/MIPS compared to $253/MIPS with the 68040. Also note the new SBC uses the MicroVAX II 78032 chip (not the MV 3000 78034) which on the MicroVAX II only had 1 MIPS at 20MHz(Digital Review 9/85 p. 60), now suddenly it has 4.8 MIPS at 20MHz in the new SBC product. What goes? If the chips are the same I must conclude VMS does not give a single job full uP MIPS, which is what I always suspected. A $10,000 price difference for 2 DEC boards: The MicroVAX 3000 and the "Real Time VAX" board that seem to have the same MIPS (?) Why should the VMS board cost any more than the SBC if the hardware is the same? (Please, only respond to my assertions if you know FACTS such as having a board in front of you to look at. I don't need a lecture on the features of VMS, etc, that "take a lot of special chips" or other such useless generalities.) How would MicroVAX owners feel if it were true that their CPU board sold for $10,000 less in the SBC market so DEC could compete with Intel and Motorola SBCs? This is the real issue. DEC alleges that VAX MIPS cost more than RISC MIPS. If this is true, why would the SBC cost less? Does anyone on the net have a DEC SBC as well as a MicroVAX 3000 or II handy to compare the boards?
grr@cbmvax.commodore.com (George Robbins) (02/07/90)
In article <26668@cup.portal.com> cliffhanger@cup.portal.com (Cliff C Heyer) writes: > Re: Digital Review 1/29/90 P. 1 > "DEC Set For Move into Board Level Market" > > I can't believe my eyes. I don't see how DEC can > hope to compete in this market! With 8+ MIPS 68030 > boards going for $2500. And the new Heurikon 68040 > 25MHz board at 13.5 MIPS going for $3400. ... > A $10,000 price difference for 2 DEC boards.. > > (Please, only respond to my assertions if you know > FACTS such as having a board in front of you to > look at. I don't need a lecture on the features of > VMS, etc, that "take a lot of special chips" or > other such useless generalities.) What facts do you want? It's obvious the DEC pricing for CPU boards used in DEC systems bears no relation to the cost of manufacturing the board or it's marketability. This is done intentionally to prevent 3-rd parties from using DEC cpu boards to make "compatible" systems and to preserve profit margins on "upgrades" and "service spares". If you are not convinced, look at the prices for things like 6300 cpu vs 6200 cpu card (essentially the same) or for a 785 or 8650 boardset vs. the original. The system COST of a 6300 is probably identical to a 6200 and the 6400 might cost maybe 5-10% more to build. The $50K price difference between base systems reflects what DEC wants to charge to create a "product line" and what customers are willing to pay in the absence of effective competition (VMS captive base). Now, after a period of years, DEC is remembering that they used to make money selling Q-bus PDP-11 SBC's and maybe they feel it's time to hawk u-VAX cards. You can only charge n times what the competition is charging and hope to have any success even with the "value added" of VMS, ELN or whatever, so they have to more reaonable pricing on these products. -- George Robbins - now working for, uucp: {uunet|pyramid|rutgers}!cbmvax!grr but no way officially representing: domain: grr@cbmvax.commodore.com Commodore, Engineering Department phone: 215-431-9349 (only by moonlite)
tihor@acf4.NYU.EDU (Stephen Tihor) (02/08/90)
First off make sure you are comparing apples and apples. DEC has produced and market four revisions of the 70832 family chipset if I have been keeping count properly. There may be a couple more I missed and I know that there are more in developement. Second the rtVAX has a slight braindamage, one level of page table translation is bypassed. This provides either (1) cheaper simpler hardware or (2) the ability to use regular vax chips with defects in this area. It also prevents you from running a standard Digital full function operating system (VMS, Ultrix, BSD UNIX) on it. YOu could probably do a real UNIX port for one but no one has. [after all its about the same level of page system support as a 386 chip] THus its selling into a wider market without impact on their captive integrated systems market. [See the recent discussions on comp.arch about the $50,000 card cage most vendors sell you when you buy midrange systems. Or the DEC pricing on the CI star coupler, a completely inert electrical bus that costs a small fortune to buy and maintain.]
carl@grc.UUCP (Carl A. Pick) (02/08/90)
>I can't believe my eyes. I don't see how DEC can >hope to compete in this market! With 8+ MIPS 68030 >boards going for $2500. And the new Heurikon 68040 >25MHz board at 13.5 MIPS going for $3400. The product isn't priced based on anything but what customers are willing to pay for it. I'll bet Heurikon's board costs them more to make than DEC's costs them. If companies have a hardware investment in the Qbus, software investment in VMS, require the DEC cachet for their customers, need worldwide service, or whatever, nobody elses SBC will do. Small wonder they sell more of theirs than all other venders combined.
kaiser@cheese.enet.dec.com (02/08/90)
All the speculation about Digital's pricing policies is invigorating and good for the blood pressure, but let's not mistake it for substantive comment on the topic of DR's article. ---Pete kaiser@cheese.enet.dec.com DEC, 2 Mt. Royal Ave. (UPO1-3), Marlboro MA 01752-9108 508-480-4345 (machine: 617-641-3450)
sritacco@hpdml93.HP.COM (Steve Ritacco) (02/09/90)
That article is probably also refering to the rumored VME based mips boards.
bcw@rti.UUCP (Bruce Wright) (02/11/90)
In article <26668@cup.portal.com>, cliffhanger@cup.portal.com (Cliff C Heyer) writes: > Re: Digital Review 1/29/90 P. 1 > "DEC Set For Move into Board Level Market" > > [...] > > Also note the new SBC uses the MicroVAX II 78032 > chip (not the MV 3000 78034) which on the MicroVAX > II only had 1 MIPS at 20MHz(Digital Review 9/85 p. > 60), now suddenly it has 4.8 MIPS at 20MHz in the > new SBC product. What goes? If the chips are the > same I must conclude VMS does not give a single job > full uP MIPS, which is what I always suspected. I don't know what the exact configuration of the new board-level machine is, so I can't really say whether the 4.8 MIPS at 20MHz is a realistic figure. But you should be aware that the standard MicroVAX II configuration does not have ANY cache memory. That's one of the major enhancements of the MicroVAX III line (not only does the chip have some built-in cache on the III, but some of the models also have additional cache). It would be quite possible for DEC to boost performance fairly cheaply by adding a cache to the board-level product. Anybody know if this is what they in fact did? The comment about VMS is rather silly - yes, machines that run it are much more expensive than lots of other machines, but its speed on a MicroVAX isn't really that different from, for example, Ultrix. The bottom line is that a MicroVAX II just isn't that fast, at least as it's usually configured. Bruce C. Wright