An IBM/360 model 50 was about the same performance as an 8088, is my
recollection. It was incredibly fast for its day, but its day was 1963.
When I first started to use one, in 1967, it was already considered
obsolete.
My next IBM mainframe was the 360 model 67, which ran TSS/360, one of the
early timesharing systems. TSS was massive, slow, and clunky, and if you
ran, as we did, 40-60 users on it, an 8088 would have looked blisteringly
fast by comparison. The 67 was a 65 with a memory map attached (the TLB
was of size 8, because the maximum number of pages one instruction could
cover was 6, and 8 was the next higher power). It had 750ns main memory
and 8us “bulk memory”, which we used instead of a paging drum. We had a
DMA access between the bulk memory and the main memory (of which we had
768K, one of the largest mainframe memories around, and a special order
since the 65/67 normally had a max of 512K) but discovered, ultimately,
that it was faster to execute code in the dead-slow 8us memory than to
page it it, because the paging overhead was so high. In the mid-1970s, we
replaced the 8us IBM memory with 2us Ampex core, but the bus still ran at
8us.
A 360/67 was about the same instruction performance as a PDP-11, but the
difference between mainframes and minis at the time was largely in I/O
throughput. A 360 equipped with a multiplexor channel could push more
bits per second through the CPU than a PDP-11 could support as an external
peripheral. Even today, the “mainframes” hold the blue ribbon in I/O
throughput, although it is now often measured in gigabytes/second.
We did not see solid state memory on computers until the late 1970s; prior
to that, everything was core. Our PDP-11s all used core memory. Our DEC
KL-10, a huge, fast ECL machine (1MIPS, screamingly fast for 1973 or so),
used core.
In the last week of the semester (as every student was trying to get their
programs running by end-of-semester), one of our engineers (we had
purchased the /360 by that time) decided to change the memory timing so
the bulk memory bus cycled at 2us instead of 8us, so we’d get 4x
performance from the bulk memory. The result was the whole machine went
down for a week, prompting this song (to “I’ve got rhythm”)
I’ve got DECtapes,
I’ve got magtapes,
I’ve got disk packs,
But has anybody seen my core?
Someone fiddled,
someone diddled,
Someone piddled,
Now no core!
(There are several more verses, but you get the idea)
joe
> Memory cost real money in those days. It was made of little magnetic
> donuts.
On IBM/360? really? (I think Soviet clones used small chips…)
> And the operating system was IBM DOS
And there was also “OS” (without “D”) which was more capable, and later
some VM hypervisor (MVS?).
BTW - what does “mainframe” mean? can you compare the PDP-11 which was
called “mini” and IBM/360 (which was called “mainframe”) of the same time?
Were /360s really much more powerful?
–
Maxim S. Shatskih
Windows DDK MVP
xxxxx@storagecraft.com
http://www.storagecraft.com
NTDEV is sponsored by OSR
For our schedule of WDF, WDM, debugging and other seminars visit:
http://www.osr.com/seminars
To unsubscribe, visit the List Server section of OSR Online at
http://www.osronline.com/page.cfm?name=ListServer