Threaded Mode | Linear Mode

**Katie Wasserman** · 04-17-2014, 04:07 PM

Quote:You might want to try using Everyready Lithium AA Batteries. They will last a lot longer than standard AA batteries. They cost more, but I can usually find them on sale for about half price several times a year and I stock up on them at the lower price. I use tem in my HP-200LX and in my Omnibook OB-430.

I was using these in my HP-200LX as well, they do last a lot longer. However they have a very flat discharge voltage profile, dropping to zero very rapidly at the end of their life. This results in the low battery indication being shown for a very short time before the cells are completely dead. I didn't like that and switched back to alkaline cells.

Bill (Smithville NJ) · 04-17-2014, 07:33 PM

Quote: just did the test (167699497) on a HP-200LX running GW-Basic and Basica. Both took about 9.3-9.5 seconds. So UBASIC is quite a bit faster.

I've never heard of UBASIC. With this speed difference. I need to check it out.

Okay, I repeated the test on a pocket computer - Sharp PC-1600. I had to slightly modify the program - Sharp's basic didn't like multiple statements after the IF/THEN in line 190. Plus used the TIME function to calculate the elapsed time.

For 167699497, it took about 100 seconds.

I was going to try the PC-1262, but can't find my unit. Probably in the bottom of a very messy drawer.

Bill

Sylvain Cote · (This post was last modified: 04-17-2014 10:26 PM by Sylvain Cote.)

I just ran the program on a HP-75C and it took 40.737 seconds for X=167699497.

program changes ...
1 DELAY 0
5 REAL X, F, M, T1, T2
15 T1=TIME
900 T2=TIME @ DISP
all PRINT's replaced by DISP
all : replaced by @

Sylvain

[edit: add program changes]

hp41cx · 04-17-2014, 10:36 PM

Bench

Guenter Schink · 04-18-2014, 06:19 PM

(04-17-2014 10:36 PM)hp41cx Wrote: Bench

With UBASIC on the 200LX without any change except for time measurement and dim(8) 6.76 sec.

A minor improvement to 6.36sec. when using shortened arithmetic e.g. A+=X instead of A=A+X.

Günter

Guenter Schink · 04-18-2014, 09:12 PM

(04-16-2014 08:23 PM)Guenter Schink Wrote: Your example of 167699497 is calculated in about 2.4 seconds

Günter

Now I have transformed the code into a structured form, no "GOTOs" any more Smile

. Then I use some special operations of UBASIC. But the algorithm is the same still.

Code:

 100   input "X=";X

  110   clr time

  120   M=int(sqrt(X))

  130   F=0

  140   data 2,1,2,2

  145    ' take care of divisors 2-7

  150   for I=1 to 4

  160   read S:gosub *Divide

  170   next I

  180   data 4,2,4,2,4,6,2,6

  185    'Now calculate with numbers >7 omitting all that can be divided by 2,3 and 5

  190   while X>1

  200    for I=1 to 8

  210    read S

  220    gosub *Divide

  230    next I

  240    restore 180

  250    if F>M then

  260    :    F=X

  270    :    S=0

  280    :    gosub *Divide

  290    endif

  300   wend

  310   ET=time1000/1000     'time1000 gives milliseconds after last "clr time"

  320   print:print "Execution Time: "

  330   print ET;

  340   end

  345   'Now start factoring

  350   *Divide

  360   F+=S                'this is shorter and faster for F=F+S

  370   while not X@F        'while there is no remainder

  380       X\=F            'this is short for x=x\f where "\" denotes integer division

  390        M=int(sqrt(X))

  400        print F;

  410   wend

  420   return

This program needs 1.53 sec for 167699497 on the 200LX.

Günter

Dave Britten · 04-18-2014, 09:51 PM

(04-18-2014 09:12 PM)Guenter Schink Wrote:
(04-16-2014 08:23 PM)Guenter Schink Wrote: Your example of 167699497 is calculated in about 2.4 seconds

Günter

Now I have transformed the code into a structured form, no "GOTOs" any more . Then I use some special operations of UBASIC. But the algorithm is the same still.

Code:

100 input "X=";X 110 clr time 120 M=int(sqrt(X)) 130 F=0 140 data 2,1,2,2 145 ' take care of divisors 2-7 150 for I=1 to 4 160 read S:gosub *Divide 170 next I 180 data 4,2,4,2,4,6,2,6 185 'Now calculate with numbers >7 omitting all that can be divided by 2,3 and 5 190 while X>1 200 for I=1 to 8 210 read S 220 gosub *Divide 230 next I 240 restore 180 250 if F>M then 260 : F=X 270 : S=0 280 : gosub *Divide 290 endif 300 wend 310 ET=time1000/1000 'time1000 gives milliseconds after last "clr time" 320 print:print "Execution Time: " 330 print ET; 340 end 345 'Now start factoring 350 *Divide 360 F+=S 'this is shorter and faster for F=F+S 370 while not X@F 'while there is no remainder 380 X\=F 'this is short for x=x\f where "\" denotes integer division 390 M=int(sqrt(X)) 400 print F; 410 wend 420 return

This program needs 1.53 sec for 167699497 on the 200LX.

Günter

That's pretty impressive. I definitely need to experiment with UBASIC this weekend.

For the record, I loaded my GW-BASIC version into QBASIC and ran it with no modifications, and it took about 12 seconds, or ~50% longer.

Still can't find any place to download Swift!BASIC that I've seen mentioned in The Palmtop Paper archive.

ckreling · 05-07-2014, 01:36 AM

(04-06-2014 06:13 PM)rprosperi Wrote: [quote='Katie Wasserman' pid='8499' dateline='1396768062']
I almost completely forgot about the Sharp OZ-7000 with the OZ-707 Scientific Computer Card inserted. This is a very capable BASIC ...

So nice to see this discussion. I have a pretty large Sharp Wizard collection (Museum) some new in box. I love hunting down elusive cards for the wizards.

As you mentioned the work you did - I recently acquired a prudential 'PRUPACRATER' card. After reading your info I looked at the back and low and behold it is an OZ-770 64KB with a special label from Prudential on the front. Thanks so much for sharing that information.

Paul Berger (Canada) · 05-07-2014, 02:44 AM

I ran Dave's prime factor program on an Panasonic HHC running snapBASIC I replaced the first data statement with an array because the restore function in this dialect restores all data statements in the program. Time functions are different you read the clock by peeking at 5 sequential bytes in memory. Replaced all the INT functions with one called FIX as INT will only work with integer values and they are limited to a signed 16 bit number. Replaced SQR with SQRT, SQR returns the square of the argument.

In interpreted mode 167699497 takes almost exactly 3 mins despite having a considerable faster CPU than the HP handhelds. I think the difference is the HP handhelds are running a CPU that is optimized for floating point math, whereas the HHC is running a 1 MHz 6502, a general purpose processor and a very simple one at that.

I wanted to try this as a compiled program but before I can do that I need to figure out a way to get the compiled program off the machine and burn it into a EPROM as the design of the system does not allow the execution of compiled programs from RAM. The design of this machine is, interesting but the keyboard on it is pretty unfortunate. It works well but the keys are crowded and the keytops are small rectangles arranged in straight columns. The key legends are printed on the top of the machine around the keys instead of on the keytops and are pretty busy, it is not a nice keyboard to type on.

Csaba Tizedes · 01-04-2019, 04:52 PM

I found Xerxes' Benchmarks for 8-Queens problem and I found a TI-74 assembly routine on that page, but as I know the assembly programming on this machine is not available. Is there any information/docs about how these units are programmed in assembly?

Also the same information available for CASIO FX-850P/880P ?

Thanks!
Csaba

Csaba Tizedes · 01-04-2019, 07:56 PM

(04-06-2014 08:26 PM)Alvaro Wrote: The Casio Fx-850P has a User Defined Function

Really? How? This chocolate-sized marvel was my first computer and I do not really remember this feature... But this was 25 years ago...

JurgenRo · 01-04-2019, 10:32 PM

(03-30-2014 12:58 PM)Sylvain Cote Wrote: Most if not all of your questions can be answered by the HHC BASIC Comparison Sheet from Marcus
Sylvain

Sylvain, you are an unbelievable source of information! Smile

Thanks for sharing this comparison data sheet!
Greetings,
Juergen

xerxes · 01-05-2019, 06:31 PM

(01-04-2019 04:52 PM)Csaba Tizedes Wrote: I found Xerxes' Benchmarks for 8-Queens problem and I found a TI-74 assembly routine on that page, but as I know the assembly programming on this machine is not available. Is there any information/docs about how these units are programmed in assembly?

Also the same information available for CASIO FX-850P/880P ?

Please check this discussion for assembly programming on the TI-74: http://www.vcfed.org/forum/showthread.ph...-the-TI-74

AFAIK there is no way for the FX-850P/880P.

Csaba Tizedes · 01-06-2019, 02:28 PM

(01-05-2019 06:31 PM)xerxes Wrote: Please check this discussion for assembly programming on the TI-74: http://www.vcfed.org/forum/showthread.ph...-the-TI-74

Thanks, I have checked and downloaded tons of information. Really valuable source!

Sukiari · 01-07-2019, 08:14 AM

(04-13-2014 05:26 AM)Waon Shinyoe Wrote: IMHO SHARP G850 series is the best one though it aims at education market. It supports BASIC, on-calc C and ASM and machine code. (Though the soft keyboard is annoying...)

I agree, it’s a great little machine, and it’s a big shame that there is basically nothing like these around any more.

toml_12953 · 01-07-2019, 08:40 PM

(04-08-2014 02:53 AM)Joe Horn Wrote: Although that wasn't the OZ-707, it was the OZ-770 (only 64K of RAM), almost identical to the OZ-707 except for having only 64K of RAM, and an IQ-775 (64K of EPROM and 32K of RAM) in which you had put the BASIC language.

How much memory does the OZ-707 have? I thought it only had 32K RAM.

Tom L

**rprosperi** · 01-08-2019, 02:45 AM

(01-07-2019 08:40 PM)toml_12953 Wrote: How much memory does the OZ-707 have? I thought it only had 32K RAM.

The OZ-707 BASIC Card had 32K RAM and was sold at retail as one of the consumer application cards for the Wizard family and had a label on the front with digits and scientific function names to see through the clear touchpad on the OZ-7000, 7200, 8000, etc. (the product line used model designation OZ-NNN in US, and IQ-NNN in ROW).

The OZ-770 had 64K RAM and had a blank face (as seen above) was sold only through VAR channels for vertical solution developers.

damaltor · 01-09-2019, 01:04 PM

I would rather stick with casio, the fx-880p is what i would select. the pb-2000 with the basic module is decent, too, the pb-1000 has a nice basic os, but the worst possible housing.

all three possibilities are not exactly cheap, though the fx-880p is a very nice device which might be "worth it" for whatever it is needed for.

alternatively, the older casio basic models are nice too but have exceptionally small displays. the good point is that i made a usb interface a while ago which allows transfer from the basic programs to and from the calculator, and also allows use with the tools made by Marcus to compile/decompile the basic programs.

Erwin · 01-09-2019, 05:24 PM

(03-30-2014 12:57 PM)Bill (Smithville NJ) Wrote: Hi Namir,

With the exception of the display, I'm sure Valentin would rightfully say the HP-71B meets and exceeds your requirements.

So I'll give you my favorite pocket computer: Sharp PC-1600.

Has four line display.
Can expand memory to about 77kb free for basic programs.
You can have memory file systems for storing programs.
Has RS-232C port,
Has Analog input port.
Best Feature: Z80 processor.
Well written Technical Manual that details all the machine routines that can either be called from Basic or used in Z80 programs.

Now for the downside:
All numeric variables are reals - there is no DEFINT command.
Variable names are two characters or less.
No support for separately defined functions.
Subroutines are the standard Basic GOSUB type.

For me, the downsides pale compared with the ability to program in Z80 machine code, save programs to memory modules as files, and ability to easily transfer files to/from the PC using the RS-232.

I do some Z80 programming (just for fun), using a cross compiler on the desktop PC and then transferring to the PC-1600. A lot of fun.

Bill

I owned a PC1600 and sold it last year complete with disc drive, printer and all manuals. I programmed Z80 Assembler in the past and Basic and it was very fine. But:

Bad to programm the display nativley
Display was bad readable - bad contrast
A/D converter was nice but too much internal noise of the machine
Keys are very soft

This all compared to my HP71B - who lacks a bigger display

regards
Erwin

toml_12953 · 01-11-2019, 03:32 AM

(01-08-2019 02:45 AM)rprosperi Wrote:
(01-07-2019 08:40 PM)toml_12953 Wrote: How much memory does the OZ-707 have? I thought it only had 32K RAM.

The OZ-707 BASIC Card had 32K RAM and was sold at retail as one of the consumer application cards for the Wizard family and had a label on the front with digits and scientific function names to see through the clear touchpad on the OZ-7000, 7200, 8000, etc. (the product line used model designation OZ-NNN in US, and IQ-NNN in ROW).

The OZ-770 had 64K RAM and had a blank face (as seen above) was sold only through VAR channels for vertical solution developers.

Thanks. The way Joe kept saying "only 64K" when talking about the 770 made it sound like the 707 had more than the 770.

Tom L