A new view on "slow": 9401-P0x

Ladies and Gentlemen,

after tinkering around for a while with my acquired P02 and P03 models I take the freedom to offer you a retro-flashback. :-)

I lately managed to reinstall the P02 with two external 6104 disks attached, in a disk enclosure I had spare from a lot I got from my former employer. They've had a *lot* of external hard drives for their Macs, decades ago. Just looking at the processor LED when there's activity, I feel the CPU is less thwarted by paging. Here is a display when the machine is idle, as it's most of the time. ;-)

Size % I/O Request Read Write Read Write %
Unit Type (M) Used Rqs Size (K) Rqs Rqs (K) (K) Busy
1 6601 1031 18,8 0,1 1,5 0,0 0,0 1,4 1,6 0
2 6601 1031 16,4 0,0 4,9 0,0 0,0 2,8 8,6 0
3 6104 988 16,4 0,0 3,2 0,0 0,0 3,5 3,0 0
4 6104 988 16,7 0,0 2,4 0,0 0,0 2,4 2,4 0

Nevertheless, I'm tempted to move 8 Megs of memory from the P03 (now: 24) to the P02 (8). Even doing things on a locally connected Twinax terminal never lets you feel stressed by a too fast computer. ;-)

% CPU used . . . . . . . : 26,0 System ASP . . . . . . . : 4041 M
Elapsed time . . . . . . : 00:00:01 % system ASP used . . . : 17,1674
Jobs in system . . . . . : 55 Total aux stg . . . . . : 4041 M
% perm addresses . . . . : 0,917 Current unprotect used . : 1074 M
% temp addresses . . . . : 0,208 Maximum unprotect . . . : 1074 M

Sys Pool Rsrv Max -----DB----- ---Non-DB--- Act- Wait- Act-
Pool Size K Size K Act Fault Pages Fault Pages Wait Inel Inel
1 3568 1838 +++ 0,0 0,0 1,5 1,5 0,0 0,0 0,0
2 3583 0 4 0,0 0,0 0,0 0,0 0,0 0,0 0,0
3 80 0 1 0,0 0,0 0,0 0,0 0,0 0,0 0,0
4 961 0 2 0,0 0,0 13,7 78,9 31,7 0,0 0,0


Main storage size (K) . : 8192

Type changes (if allowed), press Enter.

Defined Max Allocated Pool -Paging Option-
Pool Size (K) Active Size (K) ID Defined Current
*MACHINE 3360 +++ 3360 1 *FIXED *FIXED
*BASE 3299 4 3299 2 *CALC *CALC
*INTERACT 1453 2 1453 4 *CALC *CALC
*SPOOL 80 1 80 3 *CALC *CALC

Both machines are connected via the only communication adapter for "networking" to the rest of my network through a Cisco router, and only talk SNA — which is less straining on the CPUs. Also, I don't use HPR, which also needs to run in the main CPU, while I've read that pure APPN is handled by the IOP for the comms adapter.

Connection speed is at 19200 bps. The largest value allowed by both machines. Both have an APPN adjacency to my model 150 over the Cisco Router connected to Token Ring, which in turn has another adjacency to my old MS SNA Server running on the IPCS in my 150. This enables me to just use Mocha on my Mac to "directly" open a tn5250 connection to the P0x over the SNA Server. Neat! The 150 handles Intermediate Session Routing.

The following examples are taken by such a remote session.

Simple tasks such as displaying another menu, or dspmsg take about 2 seconds until the display is drawn. Wrkmbrpdm qclsrc takes 6 seconds. Compilation of QSTRUP as batch is done in 20 seconds (from pressing return in PDM to the message waiting indicator popping up). Those two seconds easily go up to over ten when SNADS is transferring bulk data over the same link.

Next, I've adjusted the network attributes to allow data compression on my 150, and to request data compression on my P0x. I checked the mode descriptions on all machines (including the SNA Server) to at least allow data compression. Screen displays seem to not benefit from data compression. Data transfers might, though.

Especially straining is switching sessions with the ATTN key menu.

I have not yet done any "porting" of my stuff coded on my model 150 yet. But I think, what I've measured so far gives you a good idea. I'm curious to see if my attempt to write "old style" RPGLE in V4 pays off by being instantly compilable on the P0x. :-)

Have a nice weekend!

:wq! PoC