User Tools

Site Tools


blog:pushbx:2022:1023_95lx_sending_at_57600_baud_line_history_optimisations_third_95lx

95LX sending at 57600 baud, line history optimisations, third 95LX

2022-10-16

Today during the file transfer between the HP 95LX (running PD ZModem version 1.26) and the Linux box (running Debian's sz/rz) I tried out sending from the 95LX at 57600 baud. I used the command line zm -88 -c1 -b57600 -h -d sz *.com. It worked, although the resulting Bytes per second rate was below 2000 (up from less than 1000 for receiving with 9600 baud), so not as much of a speed up as one could hope. But I am happy that it works at all, and sending multiple files using a single command also works very well. I did notice that the labels within the PDZM window do not seem to be readable; there's a bunch of numbers but no visible labels that indicate what those numbers are. I seem to recall from using PDZM on the server and the DOS box that there are labels there, and the space is certainly enough to hold such labels.

To add to yesterday's remarks on the lDebug line history comments, there are two ways we could optimise history handling:

First, when using the separate history segment, then the high array entries are offsets. That means when they are used as displacements, the base that we displace from is zero. That means we could conditionally disable all the displacing.

Second, the word [history.first] actually holds a constant value currently. As there is no runtime configuration that can change this, the variable accesses could be replaced by immediates.

2022-10-23

The day before yesterday rr made me aware of another offer to buy an HP 95LX. Imagine my surprise when the PLZ (german postal code) was the same as the one I live in! The same day I contacted the seller and we arranged a meeting that occurred yesterday. Compared to Menden I had a very short way to travel, just taking one bus for 14 minutes and then walking another 10.

The device turned out to be in very good condition, good screen hinges, backup battery tray included, main battery tray not damaged, all tests ran fine. Later I even ran the serial port loopback test, using a screwdriver bit as an impromptu loopback wire. That was easy to set up because the RX and TX pins of the serial port are right next to each other, the middle two pins of the four total in the port.

Merely the batteries had nearly no capacity left and were mismatched, which was quickly fixed by buying new ones from a local store. (One CR 2032 for the backup battery and two AA format NiMH rechargeable batteries for the main batteries.)

So that's the third HP 95LX in my "collection". I'm very happy about its condition, and it can serve as a backup in case the first one breaks down. No power or serial cables or memory cards were included in the offer yesterday, though I do not expect problems using the ones I do own already with this device. The power supply certainly works, as I already tested yesterday.

You could leave a comment if you were logged in.
blog/pushbx/2022/1023_95lx_sending_at_57600_baud_line_history_optimisations_third_95lx.txt · Last modified: 2022-10-23 17:49:07 +0200 Oct Sun by ecm