frank@contex.UUCP (Frank Perdicaro) (08/31/90)
A Modest Proposal The 4D 25S is a fine machine; it could be made into a better machine. Imagine this. As the place where the GR1 board goes is vacant in the server, it should be possible to place a second IP10 in there. There should be no problem with power, because the majority of power that is consumed by a 4D 25G is consumed on the GR1 and related boards. If there is no problem with power use, there is unlikely to be a problem with power dissipation. The sheet metal modifications are not too bad -- the tooling already exists for the holes that need to be punched. As this is stricly a server ( lets call it the 4D 25SS ), there is no need of a graphical terminal. Send out a dual-session, dual-port terminal ( like the Wyse-60 ) to act as the console of both cpus. Proposed Configuration 1. Make the second cpu be a diskless workstation. Put 64Mbytes in the 16 SIMM slots. Use the new NQS software to move jobs back and forth, and place both machines as peers on an Ethernet trunk. [NOTE: this is the rack-mounted 4D 25 I want] Proposed Configuration 2. Make the second cpu be a diskless workstation. Put 64Mbytes in the 16 SIMM slots. Put a network bridge card in the VME slot of the primary cpu and have the seconday cpu be on its own sub-net. This provides for zero additional Ethernet traffic and good security possibilities. Proposed Configuration 3. Very minor sheet metal modification make it possible to run a full height drive in the top slot of a 4D 25. Make these modifications and place a drive there. Cut the existing SCSI cable in half, and have the top-slot drive serve the secondary cpu. This still provides for SCSI expansion in one cpu, but provides two independent cpu in one box. Proposed Configuration 4. Further sheet metal modifications would make it possible to have each cpu have a SCSI port to the outside world. Imagine each cpu with its own Iris File loaded with a Tahiti and an Exabyte or other SCSI peripherals. Proposed Configuration 5. Take configuration 4 and... Put a FDDI card in the VME slot of each cpu ( first right-shift the cost of a FDDI card 3 bits! ) and you have a very fast communications pathway between the cpus. I suggest that the 20 Mhz R3000s be replaced with 33 Mhz R3000s to take advantage of this high-speed communications ability. Proposed Configuration 6 Use the new Wren Elite drives in place of the Wren 7 drives, replace the 16 4*9 SIMMS with 16 8*9 SIMMS and say 'this is a fast computer'. Proposed Configuration 7 Use the 4D 25 that Ratheon sells as a base for a 4 cpu system similar to the ones described above. Please sell these for $9995 configured. Please note that I am making pleas and guesses. If I have just announced a product, it is certainly be accident. -- Frank E Perdicaro, Systems Admin, etc. Xyvision Design Systems LEGALIZE Guns, drugs and cash...today. 101 Edgewater Drive inhouse: frank@bugs Wakefield MA outhouse: contex!frank@uunet.uu.net 018801285
buck@drax.gsfc.nasa.gov (Loren (Buck) Buchanan) (09/05/90)
In article <1411@contex.UUCP> frank@contex.UUCP (Frank Perdicaro) writes: >A Modest Proposal In some respects similar to the "Personal Power IRIS" I have been asking about for more than a year. The answer I got at SIGGRAPH, is that it would be nice, but our engineering dollars are going else where. The most convincing argument they gave me (and the one that has finally convinced me to give up) is for a Personal Power IRIS to really work reasonably well is for it to get a private memory bus a la the Power Series. By the time you add up engineering and manufacturing costs it just doesn't make sense. Besides, I can hardly wait for what they are spending R&D dollars on (supposedly a machine that is getting into the price range I would consider for a home machine, about 7-8K for what I would guess would be a diskless 8 bit system). > > The 4D 25S is a fine machine; it could be made into a better machine. >Imagine this. As the place where the GR1 board goes is vacant in the server, >it should be possible to place a second IP10 in there. What you are proposing sounds reasonable to me, but there may be a problem of the two boards having a conflict on the bus (easy to fix, just cut the bus in two). > >Proposed Configuration 3. > Very minor sheet metal modification make it possible to run a full height >drive in the top slot of a 4D 25. Make these modifications and place a drive >there. Cut the existing SCSI cable in half, and have the top-slot drive >serve the secondary cpu. This still provides for SCSI expansion in one cpu, >but provides two independent cpu in one box. Or leave the metal alone, and just give the second CPU the largest half-height drive available for local swap space. This still gives room left over for a tape drive, CD ROM, floppy drive, etc. Of course all of the things you wanted in terms of faster CPUs, more memory, and related stuff to improve performance would of course be welcome for us ordinary users of PIs. B Cing U Buck Loren Buchanan | buck@drax.gsfc.nasa.gov | #include <std_disclaimer.h> CSC, 1100 West St. | ...!ames!dftsrv!drax!buck | typedef int by Laurel, MD 20707 | (301) 497-2531 | void where_prohibited(by law){} CD International lists over 40,000 pop music CDs, collect the whole set.