195295fm.exe   Server 195/295 Field Maint. Test 2.0
295diag.exe       Server 195 and Server 295 diagnostic diskette
295fw170.exe   Server 295 FirmWare upgrade version 1.7
295ref.exe         Server 195 and Server 295 reference diskette version 1.7.1
295start.exe      Server 195 and Server 295 system startup diskette
mpext21f.exe     Server 295 HPFS fix for MP extensions
mass160.exe      Server 195 and Server 295 MASS/2 version 1.60
raidnw.exe         Server 195 and Server 295 RAID installation diskette for NetWare

ECC-P Memory with 4MB 30 Pin SIMMs
The memory card has (32) 30 pin SIMM sockets. Memory was sold in 32MB increments. That would be one block with eight 30 pin SIMMs.

32+4=36, not enough for ECC, BUT... 36+36=72... 64 bits for data, and 8 bits for ECC... Two 32 bit EDCs can be chained for a 64 bit word.

So... you need one bank of 8 SIMMs each for ECC-P, so the minimum is 32MB, the next is two banks (64MB), three banks (96MB), and four banks for a maximum of 128MB.

NOTE: I have seen a configuration screen, with a choice of "ECC: ON | OFF".

Previous IBM servers such as the IBM Server 85 were able to use standard memory to implement what is known as ECC-P.  ECC-P takes advantage of the fact that a 64-bit word needs 8 bits of parity in order to detect single-bit errors (one bit/byte of data). Since it is also possible to use an ECC algorithm on 64 bits of data with 8 check bits, IBM designed a memory controller which implements the ECC algorithm using the standard memory SIMMs.

The following shows the implementation of ECC-P. When ECC-P is enabled via the reference diskette, the controller reads/writes two 32-bit words and 8 bits of check information to standard parity memory. Since 8 check bits are available on a 64-bit word, the system is able to correct single-bit errors and detect double-bit errors just like ECC memory.



While ECC-P uses standard non-expensive memory, it needs a specific memory controller that is able to read/write the two memory blocks and check and generate the check bits.  Also, the additional logic necessary to implement the ECC circuitry make it slightly slower than true ECC memory. Since the price difference between a standard memory SIMM and an ECC SIMM has narrowed, IBM no longer implements ECC-P.

Memory Management and OS/2 version 1.3
Up to 128MB of 80ns ECC-P memory can be installed in the Server 295.  Although OS/2 1.3 can only address up to 16MB of this memory per processor, the rest can be accessed via the Transparent Reserved Memory Manager (RMM) utility.

When configured in multiprocessor mode, up to 16MB of memory can be assigned to each of the FP and the AP processors, for a total of 32MB.  This allocation is done using the utility MPSETUP.

The remaining memory can be accessed via RMM, which is installed as a device driver.  RMM enables this extra memory to be accessed in a number of ways:

o Disk buffers for the PDAs
o HPFS cache
o Reserved memory for some OS/2 applications - for example in response to DosAllocSeg and DosAllocHuge segment requests
o RAM-based virtual disk
o SWAPPER.DAT file used for segment swapping to disk by OS/2

RMM memory is allocated by the /TRMM= switch in the CONFIG.SYS and CONFIG.FP files.  Status information on RMM can be displayed using the RMMSTAT utility.

RMM is needed because of the 16-bit limitations of OS/2 1.3, and results in substantial performance improvements for some applications because massive amounts of data can be held in memory rather than on disk.

Bus Architecture:
64 bit 200 MB/sec Interprocessor Bus; hierarchical design to interconnect CPU(s), memory, SCSI controllers, and Remote Maintenance Processor (RMP); 6 slots.

Parity protection on data, address, and control buses (IP-Bus).

Inside the Parallan Super Server

"Superservers: Finding a Home in User Networks" Network World Jun 3, 1991 Page 23

Parallan Server 290 Architecture

 "Super Servers Waiting in the Wings",  InfoWorld Apr 6, 1992 Page 48