702 replies to this topic

[tsiozos]

Thanks sentaro21. I'm updating.

[CalcLoverHK]

Hi sentaro21,

 

I recently heard about C:Basic (compiler) but not fully understand.

 

Therefore, I have few questions:

• What features are added?
• Does it support all the calculators that C.Basic also support?
• What is the current progress? (How much % is done, 0-100%) If 0%, which year/time will you start this compiler part?
• In theory, how is the speed of C:Basic compared to C.Basic and Casio Basic?

Thanks for reading my questions. Keep up the good work!

 

Regards,

CalcLoverHK

[tsiozos]

Sentaro21, I'm having a request if it's possible. Can we have the nimble matrices as unsigned?
I'm programming an algorithm for big number calculations (addition, subtraction etc) and I store each digit in a nimble matrix cell. But when i store any digit larger than 8 it makes it negative (understandable since it uses the highest order bit as sign) and causes problems. I don't want to use a byte matrix since it will double the needed storage (we're talking large numbers here with 2000 digits and such).
Is it possible to have some kind of flag, say like Mat a.Q to signify that a is an unsigned nibble matrix?
Thanks

Edited by tsiozos, 16 August 2019 - 11:43 AM.

[sentaro21]

@CalcLoverHK

The progress of the C;Basic compiler is still around 10%.

More details will be given again, 

but not all commands can be supported initially. It starts with speeding up basic commands and formulas.

In the real version, more than 3x speed,

In the integer version it aims at over 10x speed.

 

 

@tsiozos

Ok!

I updated the trial version by changing 4 bit matrix to unsigned from signed.

If signed 4 bit are required, please use 8 bit.

 

2.23 beta for Graph 35+USB/35+EII/75/85/95 (SD) (trial version)

https://pm.matrix.jp...ASIC223beta.zip

        -Changed 4bit matrix specification to unsigned from signed.

 

1.23 beta for Graph90+E/fx-CG10/20/50 (trial version)

https://pm.matrix.jp...ICCG123beta.zip

 

Edited by sentaro21, 17 August 2019 - 02:54 AM.

[tsiozos]

Thank you very much! I'm updating right now.

[tsiozos]

Ok sorry to bother you every now but it seems I hit another bug with integers in real mode and For loops:
'#CBDBL
For 1->i% to 6
Disp i%
Next

This code makes the first loop i% takes the value of 1 then quits.
OTOH, the same loop in integer mode works fine. Also if i remove the % it works fine. If I increase the i% variable outside the loop with 1+i%->i% also works just fine. So, my guess is that integer variables confuse For loops or something.
Latest version of C.Basic.

[sentaro21]

@tsiozos

I'm very thankful for bug reports every time.

"For~Next" bug had before, but I was able to fix it.

With this update, the executing speed of the loop itself will be about 3 times faster by using the integer variable the control variable in double mode.

 

This update will make 2.23 the official version.

 

2.23 beta for Graph 35+USB/35+EII/75/85/95 (SD)

https://pm.matrix.jp...ASIC223beta.zip

        -Changed 4bit matrix sspecification to unsigned from signed.

 

1.23 beta for Graph90+E/fx-CG10/20/50

https://pm.matrix.jp...ICCG123beta.zip

-Fixd the bug that not worked when the control variable of For~Next is different from the execution mode.

(Example) '#CBDBL

  For 1-> I% To 6

   Disp I%

  Next

 

-Changed 4bit matrix specification to unsigned from signed.

[tsiozos]

Thank you very much for the bug corrections. Makes my simulation much faster indeed!

One last little bug before you release the final 2.23:

sprintf function doesn't seem to work correctly when printing an integer variable with the %d sequence:

#CBDBL
1->i%
Disp sprintf("%d", i%)  -------> prints some weird number (probably some double bit equivalent) 1072693248

but
#CBDBL
1->i%
Disp sprintf("%f", i%)   -----------> prints correctly the value but as a double 1.00000

[sentaro21]

If it is different from the execution mode, the argument of sprintf must specify the data type.

#CBDBL
1->i%
Disp Sprintf("%d", %i%)

Edited by sentaro21, 18 August 2019 - 07:53 AM.

[tsiozos]

But that's what i do and it doesn't work correctly.

Edited by tsiozos, 18 August 2019 - 08:07 PM.

[sentaro21]

But that's what i do and it doesn't work correctly.

How is the output result?

[tsiozos]

#CBDBL
1->i%
Disp sprintf("%d", i%)

 

I get

1072693248

[sentaro21]

Sorry,there was a bug left in fixing For~Next in 2.23.

If use an integer variable for the control variable,

Fixed a bug that the end value and step value were expression evaluation in integer mode. 

'#CBDBL
6->E
For 1-> I% To E
 Disp I%
Next

The variable E should be evaluated with the double variable, but it was evaluated with the integer variable E%.

 

So, It was fixed and re-uploaded.

 

2.23 beta for Graph 35+USB/35+EII/75/85/95 (SD) 

https://pm.matrix.jp...ASIC223beta.zip

 

1.23 beta for Graph90+E/fx-CG10/20/50

https://pm.matrix.jp...ICCG123beta.zip

-Fixd the bug that not worked when the control variable of For~Next is different from the execution mode.

(Example) '#CBDBL

  For 1-> I% To 6

   Disp I%

  Next

 

-Changed 4bit matrix specification to unsigned from signed.

 

 

 

@tsiozos

#CBDBL
1->i%
Disp sprintf("%d", %i%)

Please add % before i%.

 

 

 

Edited by sentaro21, 19 August 2019 - 06:23 AM.

[tsiozos]

Thank you very much. It's unbelievable how fast you fix things!
Updating right now.

[tsiozos]

Ok I'm officially impressed.
I've made an algorithm that multiplies very large numbers. I ran it with two 47-digit numbers.
Casio Basic: ~44 sec
C. Basic: 0.5 sec. Yes: half a sec!!

Amazing. We're in C++ territory here.
I'll upload the code to try it yourself.

[sentaro21]

@tsiozos

About 80 times speed is amazing!!.

I'm looking forward to the program code.

[tsiozos]

Posted here:

 

https://community.ca...-vs-casiobasic/

[sentaro21]

Thanks for new entry.

[tsiozos]

Just an idea: is it possible to make C.Basic extensible? For example if i write my own functions in a G1A file can i make C.Basic load and execute them?

Edited by tsiozos, 21 August 2019 - 05:42 AM.

[sentaro21]

I think it would be very interesting if it could be extended with external g1a files.

Although it is difficult with the current specifications,

I will consider whether there is a possibility of realization.

[tsiozos]

Great. This would be great.

[Krtyski]

Hi, tsiozos and sentaro21

 

In a mean while until sentaro21 will be able to get some good idea, I think modified LARGEMUL could be used as subroutine.

 

I've wrote some subroutine which requires arguments and return, eg "Input Box" which I posted here before. Based on this experience, maybe following approach help.

 

For example, Mat X and Mat Y are defined in main routine, then call LARGEMUL, Prog "LARGEMUL" to output the result using Mat X and Y (global variables). In this case there is no return.

 

What do you think?

Edited by Krtyski, 22 August 2019 - 01:10 PM.

[tsiozos]

Hi Krtyski.

Yes it will work. And the result can be written to a Str global variable.

[CalcLoverHK]

Hi sentaro21,

 

Seems I found a bug about memory management: In Main Mem, you can't go to SMEM when there is no file in main memory.

This can be easily fixed by creating at least one program.

 

Cheers,

CalcLoverHK

[sentaro21]

@tsiozos

@Krtyski

Sorry, it's hard to use as an external library of add-in now.

It will be a future extension.

 

 

@CalcLoverHK

Thanks!

I confirmed and it will be fixed in the next update.

[tsiozos]

Ok understandable, never mind. I'm looking forward for the future updates.

Edited by tsiozos, 26 August 2019 - 01:15 PM.

[CalcLoverHK]

Hi sentaro21,

I am recently testing my programs in C.Basic, but C.Basic has some problems on my CG50 emulator. Almost every time the program runs to a loop and exit by pressing AC/on, it stopped running and do not show the Break message. It also cannot be closed, so I have to use Task Manager to mandatory close it. I don't know why it happens, but it is annoying that I have to use Casio Basic editor (it has no problems at all). Can you investigate it?

Cheers,
CalcLoverHK

[sentaro21]

I confirmed a problem with the AC/on key on emulator.

Please wait for a while.

[sentaro21]

Here is new updated version.

The main feature of this update is that the monochrome version can be used to acquire voltage.

 

2.24 beta  for 9860G/9860GII series.

https://pm.matrix.jp...ASIC224beta.zip

        -Added BatteryStatus command that return battery voltage.(Thanks to Simon's document sample program.)

         VARS-F3(extd)-F6(BatteryStatus)

        (Example) BatteryStatus

         retrun value is 516, it is 5.16V

         *Supported all models.

        -Fixed the response of the Getkey command when overclocking to 88MHz or higher in 9860G Slim.

        -Changed specification that return value of the Getkey in 9860G Slim.

         [HELP] ->11

         [LIGHT]->21

 

1.24 beta for CG10/20/50/Graph90+E.

https://pm.matrix.jp...ICCG124beta.zip

        -Fixed the BatteryStatus command bug that not worked in comlex mode.

        -Changed the return value of the BatteryStatus command up 9% in CG50/Graph90+E.

 

--Same update--

        -Fixed the bug that not change to the storege memory from main memory in file mode F6-F6-F4(->SMEM).

        -Fixed Load command bug that can not load the string variable.

         (Example) Load "Sample"->Str 1

        -Added battery voltage display in version popup of setup.

        -Added specification that CSV text file can be edit.(trial ver.)

         F2(edit)

        -Added specification that plain text file can be edit.(trial ver.)

         Shift+[X^2] (extended)

 

 

 

@CalcLoverHK

Please try CBCG124E.G3A in the Zip file.

 

 

[CalcLoverHK]

Hi sentaro21,

 

I forgot to tell you that if you do the memory management bug procedure vice versa (a.k.a. SMEM→Main Mem), this bug still happen. My bad.

 

You can reset all in calculator and produce it. Again, it can be fixed by creating at least one program.

 

And another problem in my CG50 emulator, when I try to open an unsupported file and pressing EXIT, it stucks so long time.

 

Cheers,

CalcLoverHK

Edited by CalcLoverHK, 29 August 2019 - 01:10 PM.

[sentaro21]

Thanks for more bug reports. 

I'll fixed next update.

Please wait a little.

[CalcLoverHK]

Hi Krtyski,

Regarding the question you ask me in IRC channel:

Why do you use emu­lator, not real CG50?

- Because fx-9860G series, fx-CG series and even fx-5800P calculators are not available in Hong Kong. Well, you can still find them, but it is extremely rare. Most importantly, we, Hong Kong students, are not allowed by HKEAA to use such calculators in exams (No.1 reason is they have access to 26 letters and you know  , but I can't think why HKEAA doesn't know that fx-9860GII and fx-CG50 have Exam Mode). We normally use fx-3650PII and/or fx-50FHII calculators. You can look at HKEAA-permitted calculators list at http://www.hkeaa.edu...cal/CAL2019.pdf. Plus the expensive price of them (I'm in Form 3!), I prefer the emulator to the real one at this time.

 

And thanks for mentioning the problem I have. It sometimes stucks in my emulators, but I don't know if the real one also has this problem?

 

Cheers,

CalcLoverHK

Edited by CalcLoverHK, 29 August 2019 - 02:55 PM.

[sentaro21]

Here is updated for fx-CG Manager.

CBCG124E.G3A was deleted because there was still a bug.

 

1.24 beta for CG10/20/50/Graph90+E. (re-uploaded)

https://pm.matrix.jp...ICCG124beta.zip

        -Fixed the bug that key input is hung-up in fx-CG Manager.

Edited by sentaro21, 30 August 2019 - 03:39 AM.

[tsiozos]

Ok, question:
I built the serial cable with the MAX232 chip and I'm buying a usb to serial adapter for my laptop for testing. If I use
OpenComport38k 7,0,x
what do I need to put in "x"?
The manual says 0: not interrupt, 1: interrupt but I don't quite understand that.
I just need to send bytes to the pc running Hyperterminal on Windows.

Edited by tsiozos, 30 August 2019 - 06:30 AM.

[sentaro21]

When it is 1,

to enable AC/on key break. 

When it is 0,

serial command is wait until coming valid data.

 

I do n’t know if the communication works,

I hope it goes well.

[sentaro21]

Updated these add-ins to match the battery voltage of C.Basic.

 

for fx-9860G/GII/Slim (SH3)

https://pm.matrix.jp/Ftune_120.zip

 

for fx-9860GII (SH4A)

https://pm.matrix.jp/Ftune2_120.zip

 

for Graph 35+E II (SH4A)

https://pm.matrix.jp/Ftune3_220.zip

 

for fx-CG50 / Graph 90+E

https://pm.matrix.jp/Ptune3_022.zip

[Krtyski]

sentaro21

 

many thanks for the updating all version of over clocking tools.

 

In my Japanese site, I revised updated info.

[Krtyski]

Hi, CalcLoverHK

 

Thank you for the answer!

 

Surprisingly I didn't know those programmable calcs are not available in HK, one of important financial center in Asia.

It's long time ago when I was student and at that time, as well as HK in all exams did not allow us to use any type of calculators, even in graduate school entrance exam.

 

Casio's largest market is for students in North America and some part of Europe, they said it's "Gakuhan" (https://edu.casio.co...kuhan/index.php). So in mother country of Casio, programmable calculator market is quite small. Due to this business size in Japan, some models such like fx-9860G Slim, fx-CG400 and fx-CG500 have not been available. I bought them in Amazon USA, eBay, etc.

 

Anyway, problems in key-in on C.basic has not been occurred on real calculators as far as I know. Between emulator and real one, timing of key input is slightly different. So sentaro21 adjusted the timing for emulator. Now C.Basic detect emulator automatically and change the timing and then single C.Basic add-in file supports both now.

 

 

Hi Krtyski,

Regarding the question you ask me in IRC channel:

Why do you use emu­lator, not real CG50?

- Because fx-9860G series, fx-CG series and even fx-5800P calculators are not available in Hong Kong. Well, you can still find them, but it is extremely rare. Most importantly, we, Hong Kong students, are not allowed by HKEAA to use such calculators in exams (No.1 reason is they have access to 26 letters and you know  , but I can't think why HKEAA doesn't know that fx-9860GII and fx-CG50 have Exam Mode). We normally use fx-3650PII and/or fx-50FHII calculators. You can look at HKEAA-permitted calculators list at http://www.hkeaa.edu...cal/CAL2019.pdf. Plus the expensive price of them (I'm in Form 3!), I prefer the emulator to the real one at this time.

 

And thanks for mentioning the problem I have. It sometimes stucks in my emulators, but I don't know if the real one also has this problem?

 

Cheers,

CalcLoverHK

Edited by Krtyski, 01 September 2019 - 11:52 PM.

[CalcLoverHK]

Hi Krtyski and sentaro21,

 

Thanks for great updates. But tomorrow (9/2) is the starting of school days, so I can't try it yet. I will try to keep active on UCF.

 

As I said we use fx-3650PII and/or fx-50FHII in exams, but they have very limited Casio Basic (?→:◢, RTL, Goto, Lbl, If statement, For statement, While statement, Break) and cannot be nested more than one. The differences between fx-3650PII and fx-50FHII are that fx-3650PII has calculus calculations, while fx-50FHII has more programming bytes (680 bytes), 23 formulas (you cannot show the procedures because this is H (Hong Kong) varient) and 40 scientific constants. We normally use fx-50FHII in normal calculations and formulas, but when you choose M2 (Mathematics Extension Module 2 - Algebra & Calculus) as an elective course, you have to use fx-3650PII.

 

Buying both two calculators is the best situration in case of storing more programs or being uncertained about choosing M2 or not. However, the price is a big problem. fx-50FHII costs 175 HKD and fx-3650PII costs 155 HKD (These prices are only for school massive order, not to mention market price). Adding together is $330. Well, $330 for only 1070 bytes, 8 programming slots and very limited Casio Basic is not a good idea because the performance-price ratio is too low. Moreover, we are in 2019, but both calculators are released in 2013. No further upgrades for six years. I think it is time for Casio to upgrade and migrate those old models because they are similar.

 

Hope Casio cares about Hong Kong market.

 

Cheers,

CalcLoverHK

[CalcLoverHK]

Hi sentaro21,

 

Is there a way to treat the undefined value?

 

Example:

Undefined(1÷0)=>"Undefined ERROR"

It will show "Undefined ERROR" instead of showing the error message.

The "Undefined(" command is for not showing the error message. It will return 1 or 0 depend on the expression.

 

Cheers,

CalcLoverHK

Edited by CalcLoverHK, 03 September 2019 - 06:53 AM.