PowerBASIC and the Cambridge Test

[TimoVJL]

DELETED

[Gunther]

OBJ file consist many parts, ...

I know that OBJ files consist of many parts, because I've read this book.

... so a bit funny comment

By now I know that you have a very strange sense of humor. What's so funny about that? I have compiled the assembly language source with JWASM and TASM 4.0; both in 2 passes. With the command line
switch /q for TASM. This suppresses the parts that are not necessary for linking. The size of the OBJ JWASM file is 345 bytes and 375 bytes for TASM. And that with this ridiculously small source. I'm sorry,
but that' s not funny to me.

But once again: That's not the issue here. The thread is about PowerBASIC and the Cambridge test. There can be no doubt about this, because this topic is quite clearly named in the title.

[HSE]

Hi Gunther!

The OBJ file generated by JWASM has a size of 343 bytes. Other assemblers obviously insert a whole lot of junk.

Like Timo say different OBJ sizes result in same EXE size. In fact, the OBJ file generated by ML has a size of 335 bytes.

Just to play, I builded JwasmR, and produce the 343 bytes OBJ, and is the same.

A little intriguing, but not important to enter Cambridge 

Regards, HSE.

[six_L]

Hi,Gunther
i got another result.
x1+x2=2*log(3,20)= 5.453666055721684 ;real10
;----------------------------------------------------
2*log(3,20)=[2*log(2,20)]/[1*log(2,3)]

Code Select Expand

local @tmpIValue:QWORD
local ResultReal10:REAL10

invoke RtlZeroMemory,addr buff,sizeof buff

invoke RtlZeroMemory,ADDR OutBuf,sizeof OutBuf
finit
;int 3
mov @tmpIValue,2
fild @tmpIValue
mov @tmpIValue,20
fild @tmpIValue
fyl2x ;st0=20,st1=2 --> st0=2*log(2,20)

        fld1   
mov @tmpIValue,3
fild @tmpIValue
fyl2x ;st0=3,st1=1 --> st0=1*log(2,3)
fdiv st(1),st(0) ; -->[2*log(2,20)]/[1*log(2,3)]
fxch st(1) ; st1 <--> st0
lea rax,ResultReal10
        fstp TBYTE PTR [rax] ;st0 --> addr ResultReal10

invoke FpuFLtoA64,addr ResultReal10,15,ADDR OutBuf,SRC2_DIMM or SRC1_REAL or STR_REG ;STR_SCI

what's wrong?

regard.

[Gunther]

Like Timo say different OBJ sizes result in same EXE size. In fact, the OBJ file generated by ML has a size of 335 bytes.

Just to play, I builded JwasmR, and produce the 343 bytes OBJ, and is the same.

This is certainly correct for such small OBJ files. Otherwise, there may already be a difference.

A little intriguing, but not important to enter Cambridge 

Of course. But wait a bit: in a few days I'll post on the Oxford test, which is equally quite challenging. I don't want anyone to be able to say that I'm taking sides unilaterally with Cambridge.

[Gunther]

Dear six_L,

Hi,Gunther
i got another result.
x1+x2=2*log(3,20)= 5.453666055721684 ;real10
;----------------------------------------------------
2*log(3,20)=[2*log(2,20)]/[1*log(2,3)]

At first glance, I would say that there is something wrong with the calculation. For example, I think that 1*log(2,3) = log(2,3). If you want, send me a PM with your email address. I'll then send you a PDF file with
the correct derivation of a formula with which you can calculate any logarithms to any bases. If others are interested in it too, I can also publish it here in the forum.

But I think that there is no great general interest in that. Therefore, let's do it individually between both of us. 

[HSE]

i got another result.
x1+x2=2*log(3,20)= 5.453666055721684 ;real10
what's wrong?

Nothing wrong! It's that.

Using qWord's algorithm in BASIC

Note for Gunther: perhaps is more clear to say "result using x1 is ..."

updated files.

[TimoVJL]

DELETED

[Gunther]

HSE,

Nothing wrong! It's that.

Using qWord's algorithm in BASIC

right. It's the result for x1 + x2. But in principle, it doesn't matter at all. All we have to show is that 9*(SQR(3))^x - 3^x = 20. That's exactly what my program does - nothing more, nothing less.

But it's good that you've clarified the question. However, I had trouble because you did not change the old comment function header of GetSolution. And why is the new function called logbx? Is it
just a name or does it have a deeper meaning?

Note for Gunther: perhaps is more clear to say "result using x1 is ..."

That' s a matter of taste, see my remark about what we just need to show.

[Gunther]

Sure i do that, before that, i had to make a funny gmail address for that

What's with all your blather? Are you bored?

[HSE]

However, I had trouble because you did not change the old comment function header of GetSolution.

I don't even noted that 

And why is the new function called logbx? Is it just a name or does it have a deeper meaning?

Means "logarithm of base x". It's the name in use in SmplMath from years (2011, I think).

Take me a while to find how to extract results:

Code Select Expand

           **************************************
           * Result x1    = 2.92994704143585433 *
           * Result x2    = 2.52371901428582975 *
           * Result x1+x2 = 5.45366605572168408 *
           **************************************
           2 * logbx(3,20) = 5.45366605572168408


[Gunther]

HSE,

Means "logarithm of base x". It's the name in use in SmplMath from years (2011, I think).

ah, qword's material. Since I'm not a big fan of macros, I didn't look into it. Are you the current package maintainer?

Take me a while to find how to extract results:

Why? The approach is quite simple after all. Correctly combine the function with its inverse function and logarithmize this expression. That's all. The rest is just a matter of skills.

[HSE]

Gunther,

Are you the current package maintainer?

That is qWord suggestion, but everyone can make a fork.

Why? The approach is quite simple after all.

Sorry, I was talking about the program itself. The program look like a console program using print. After years using print in console assembly programs take me a lot of effort to realize that PowerBasic print is a graphic function   

The instruction for console output is stdout and work exactly like print if console output is not captured. 

Code Select Expand

PRINT
PRINT"           **************************************"
PRINT"           * Result x1    ="; STR$(x1, 18);" *"
PRINT"           * Result x2    ="; STR$(x2, 18);" *"
PRINT"           * Result x1+x2 ="; STR$(x1x2, 18); " *"
PRINT"           **************************************"
PRINT
PRINT"            2 * logbx(3,20) = "; STR$(x3,18)
PRINT


Code Select Expand

STDOUT"           **************************************"
STDOUT"           * Result x1    =";
STDOUT STR$(x1, 18);
STDOUT " *"
STDOUT"           * Result x2    =";
STDOUT STR$(x2, 18);
STDOUT" *"
STDOUT"           * Result x1+x2 =";
STDOUT STR$(x1x2, 18);
STDOUT" *"
STDOUT"           **************************************"
STDOUT
STDOUT"           2 * logbx(3,20) = ";
STDOUT STR$(x3,18)
STDOUT

[six_L]

Hi,HSE

Nothing wrong! It's that.

thanks your confirm.

Hi,Gunther

right. It's the result for x1 + x2. But in principle, it doesn't matter at all. All we have to show is that 9*(SQR(3))^x - 3^x = 20. That's exactly what my program does - nothing more, nothing less.

i can't run the codes becauseof no 16bit system. At first glance, the paper " expo.pdf " is so remarkable, written 
carefully / perfectly / earnestly / attractively . it Leads me to calculate again. You are always admirable while done everything

regard.

[Gunther]

HSE,

Sorry, I was talking about the program itself. The program look like a console program using print. After years using print in console assembly programs take me a lot of effort to realize that PowerBasic print is a graphic function   

The instruction for console output is stdout and work exactly like print if console output is not captured. 

that's remarkable. Can you give a simple working example for PB 3.5 and PBCC respectively? Steve aka Hutch and I would be very interested in this.