On 8/10/2014 11:21 AM, Henk wrote:
Just by luck, I did some timing measurements for a
friend on an 11/94.
As it is just an 11/94 box without any peripherals, I used TU58EM.
First I toggled the TU58 boot code from EK-OTU58-UG-004 Appendix A in.
Every time TU58EM reports a packet error. Running at 9600 Bd. The flags
-t or -T does not improve that.
Then I toggled the TU58 boot code from EK-OTU58-UG-002 Appendix A in.
Again, no luck. But then I Used the -T flag ... I get the RT11 . prompt.
And it works at 9600 Bd and at 19k2.
I can repeat the experiment running at a higher baudrate to see whether
RT11 can cope with the higher speed ... I would be surprised (positively)
if RT11 can handle 115.2 kBd.
- Henk
As I said, I don't use RT-11 over TU58EM, only XXDP to run diagnostics.
I know
that XXDP is not interrupt driven, it uses simple polled I/O. Almost for
certain
RT-11 DD driver uses interrupt driven I/O. So processing interrupts at
115.2K
input rate on a DL11-W on a slower CPU (like my 11/34) is going to fail.
I don't use any of the switches to add artificial timing delays (-t or
-T) on my
11/34 with the DL11-W at 115.2Kb, and it boots/runs XXDP diags just fine.
Now my 11/44 running the TU-58 interface at just the standard 38.4Kb rate
seems slow in comparison (at least for XXDP diagnostic loading).
For any real OS running I have Emulex UC17/18 Unibus SCSI interfaces.
Don