OpenVMS Alpha V7.2 for the MULTIA This kit is provided AS IS for Hobbyist Use Only, and is completely UNSUPPORTED by OpenVMS or Compaq Computer Corporation. Overview: The Multia (aka UDB or MTU) was not designed for general purpose use -- it was designed as a thin "NT Client", OpenVMS support was never planned for it - the Multia was supposed to be an X-Terminal-like appliance. Long after the Multia shipped as a NT-client, when it began to appear as a cheap UNIX box on the secondary market, OpenVMS investigated how much work it would take to support the Multia. The final conclusion was that it was feasible, but made little business sense (it was too slow, and was no longer being produced). Its design is based on an early LCA (Low Cost Alpha Chip) based workstation design that was scrapped because of low performance. The AlphaBook, which used a conventional EV45 also used a very similar design. Finally, the console firmware image available was designed for debug and diagnostics - but it was (and is) not product quality. The system image in this kit is essentially the proof-of-concept code that was done to show VMS would boot on it... plus some driver changes to make the network and graphics function. This kit will install on VMS V7.2. A kit for V7.1-2 will also be provided, as well as raw files for V6.3-1H3. Support for adding new platform support got much easier, cleaner, and standard starting in V7.1-2 (that is explicitly: no V7.0, V7.1, V7.1-1H1 or V7.1-1H2). Restrictions: 1 The PCMCIA support was completely removed, because the Intel chip on the Multia was not compatable with the Cirrus chip on the Alphabook. This means, of course, that you will not see and cannot use any PCMCIA cards. 2 The Multia uses shared interrupts, and as a result, a special ZLXp-E graphics driver is needed that does not use interrupts. 3 If you have a Multia with a PCI slot, you can't use any card that requires interrupts. 4 The SRM console on this system is very out of date, and fragile. It was designed only for diagnostic use. If things don't work... that's pretty much tough-luck - it will never get fixed. Requirements: A Multia: - at least 40mb of memory (DECnet/OSI requires 64mb or more) - a floppy drive - a hard drive - a CDROM drive (or the hard drive must already have VMS V7.2 or V7.1-2 loaded on it) - access to a PC system, or another means to create a DOS (FAT) formatted floppy and copy files to it. Instructions: The V7.2 and V7.1-2 kits are provided as a backup image of a floppy. This is called a "SHIP" kit, it is a PCSI kit that when used in conjunction with a VMS system disk (or installation CD) will allow a unknown system to boot (sometimes called a "foreign" boot). You can create a floppy from the backup image of this kit by issuing the command: $ BACKUP/IMAGE MULTIA_V71-2.FLOPPY/SAVE DVA0: INSTALL THE CONSOLE The first thing that you must do is to create a floppy that contains the SRM console image that will be written to flash memory. Create a FAT (PC format) floppy and copy the following files from the kit to the FAT floppy: DVA0:[VMS$COMMON.SRM]QSBYPASS.SCR DVA0:[VMS$COMMON.SRM]SRM.EXE Place the FAT floppy into a powered off system, power the system on, and the floppy will automatically load and flash the console. When it is done, remove the floppy and power cycle. You should then get a ">>>" prompt. INSTALL OPENVMS CLUSTER: If you will boot the Multia into a cluster, the kit can be installed by simply placing the floppy into the system and doing: $ PRODUCT INSTALL * /SOURCE=dva0:[vms$common] STANDALONE: If you boot standalone, its a bit more complicated... you must insert the floppy, and do a foreign boot. If you are installing VMS from CD, the PRODUCT/INSTALL will be done as part of the final installation process. If you are booting a OpenVMS Installation CD: Place the installation CD into the CDROM drive, and the floppy kit into the floppy drive and type the following: >>> boot dkaNNN,dva0 -fl 0,80000 ^^^ (Your CDROM unit here) This will invoke a "foreign" boot. The questions are pretty simple. It will ask you where the secondary media is located (that's the floppy in DVA0). And it will prompt to make sure it's in the drive. Do the installation as normal. When you boot the system disk for the initial time, you must perform a foreign boot: >>> boot dkaNNN,dva0 -fl 0,80000 ^^^ (Your just created VMS disk unit here) The installation will install the Multia support automatically onto the system disk before invoking AUTOGEN and rebooting. After this, you do not need the floppy, and can boot as normal. STANDALONE USING EXISTING OPENVMS V7.2 DISK: If you are booting an already installed OpenVMS V7.2 hard drive, then boot it as you would have done the second time above: >>> boot dkaNNN,dva0 -fl 0,80000 ^^^ (Your VMS disk unit here) When the system has booted, log in and install the kit by issuing the command: $ PRODUCT INSTALL * /SOURCE=dva0:[vms$common] Hints: The Multia has a TGA (ZLXp-E1) controller onboard for graphics. There are 4 jumper blocks (8-pins) grouped on the motherboard behind the floppy, if they are all removed, you should get 1280x1024 at 72Hz. Along the edge of the motherboard are some jumpers for speed. The system can be overclocked up to 300MHz. Of course, the chip may not work at higher clock speeds. All jumers in (X X X) is 300MHz; (X o X) is 233MHz; (o X X) is 166MHz; (X X 0) is 266 MHz. Kit Contents: [VMS$COMMON.SRM]QSBYPASS.SCR;1 [VMS$COMMON.SRM]SRM.EXE;1 [VMS$COMMON.SYS$LDR]SYS$CPU_ROUTINES_0B04.EXE;1 [VMS$COMMON.SYS$LDR]SYS$EWBTDRIVER.EXE;1 [VMS$COMMON.SYS$LDR]SYS$EWDRIVER.EXE;1 [VMS$COMMON.SYS$LDR]SYS$GYBDRIVER.EXE;1 [VMS$COMMON.SYSEXE]APB.EXE;1 [VMS$COMMON.SYSEXE]FIRM_REV_MATRIX.DAT;1 [VMS$COMMON.SYSHLP]MULTIA-V0100.RELEASE_NOTES;1 [VMS$COMMON.SYSUPD]MULTIA_FIRM_CHECK.COM;1 [VMS$COMMON.SYSUPD]PCSI_POSTINSTALL.COM;1 [VMS$COMMON]UNSUPPORTED-AXPVMS-MULTIA-V0100--4.PCSI$DESCRIPTION;1 [VMS$COMMON]UNSUPPORTED-AXPVMS-MULTIA-V0100--4.PCSI$TLB;1 Credits (in random order): Paul Jacobi - did the initial proof-of-concept for V6.2 (he was the AlphaBook Project Leader). Dick Stockdale - did the network boot support, and the initial port of the V6.2 platform code to V7.1. Fred Kleinsorge - hacked up the graphics driver, created the final platform support code, and created this kit, and these notes. Steve Hoffman - who distributed it, and will put it on the FREEWARE disk. John Wisenewski - who bugged OpenVMS until we did the hobbyist program Al Meier - bugged and bribed us until we did the Multia work Dave Maryanski - who did the product management "grunt work" that makes it all possible, and who supported it from the beginning. Ann McQuaid - Daves boss, and the one who ultimately said yes.