dave@dptechno.uucp (Dave Lee) (10/09/90)
Does anyone know if the serial port on the new series 400 computers from hp (says HP and apollo on the lable) supports fast baud rates? I would like to run a trailblazer modem at 19.2K for uucp transfers. Using the 9000/300 series fast baud rates could not be sustained in uucp due to a 1 character hardware buffer. (Adding a MUX card helps, however). Anyway, if the new 400 series has better serial port hardware, is it supported in 7.03 unix ? uname -a says -- HP-UX davesys 7.03 A 9000/375 davesys Also, has anyone actually tried uucp at 19.2K on these machines ? Any response, would be helpfull. Email or post at your convience. Thanks a bunch. -- Dave Lee uunet!dptechno!dave
rjn@hpfcso.HP.COM (Bob Niland) (10/10/90)
re: > Does anyone know if the serial port on the new series 400 computers > from hp (says HP and apollo on the lable) supports fast baud rates? Yes. It can be set to rates up to 460K bps. Note: On Series 400, this is for port 1 only. Ports 2&3 (available via K2292 fan-out cable) max out at 19.2K, have single-byte buffering and only software flow control. > I would like to run a trailblazer modem at 19.2K for uucp transfers. > Using the 9000/300 series fast baud rates could not be sustained in uucp > due to a 1 character hardware buffer. The 345/375/400 have a 16-byte FIFO, which helps. > Anyway, if the new 400 series has better serial port hardware, is > it supported in 7.03 unix ? Not officially (it's in 7.05). You can obtain an unsupported patch kit by having your local HP support representative contact me. This shar package also includes centronics patches and tips. > Also, has anyone actually tried uucp at 19.2K on these machines ? I am writing this note over a 19.2K link using HST modems. One end is a 375 using the internal port with 7.05 features turned on. Although it works for ordinary 'cu', it is not 100% reliable for file transfers. ~%put has a 1% error rate, and long uucp (f protocol) transfers often fail. I suspect that this is due to the 16-byte FIFO being smaller than the 32-byte hysteresis normally used for Xon/Xoff flow control. I will be switching back to a mux port shortly. My guess is that the FIFO would provide reliable 9600 bps operation, which is an improvement over the 2400 or so we see on the old 1-byte ports. The new ports also have CTS/RTS bidirectional hardware flow control. If you can use this in your application, and the external device sends less than 16 bytes after detecting RTS high, I would expect very reliable operation and high speeds. Regards, Hewlett-Packard Bob Niland 3404 East Harmony Road Internet: rjn@hpfcrjn.FC.HP.COM Fort Collins UUCP: [hplabs|hpu*!hpfcse]!hpfcrjn!rjn CO 80525-9599