nm

8679 Reputation

19 Badges

11 years, 274 days

MaplePrimes Activity


These are replies submitted by nm

@Preben Alsholm 

I also use windows 10, pro version with 128 GB. This is strange that C_R does not see the crash and I see it and we are both on windows 10. Something else is different.

But good you are able to reproduce it, even if on windows 11. so it is not just me. I do not think the windows version makes a difference. There is something else going on.

I noticed using Maple 2024, I get many more server crashes on same PC than I had on Maple 2023.2.1 and I have not made any changes really to my PC or the software I use. I do not have any special anti-virus software, other than what windows 10 itself has.  (This is different from the random hangs which happens much more frequent).

I wish Maple would generate a log file when the server crashes. May be a core dump somewhere to use for analysis to see the cause of the crash?

But nothing is generated as far as I know. Even the pop-up message does not give any indication of the cause of the server crash. This is not good.  

@C_R 

Thanks for trying, This is very strange, Why some get this crash and some do not.  I have no idea what is going on. May be someone will figure it out.

I'll try to turn on tracing and see if I can find something new.

@C_R 

I noticed you do not have Physics updates installed on your system. This could be the difference., I do not know. Try with Physics updates installed with latest version and see.

@ecterrab 

FYI, I've added two movies in my post.  I see you are using mac OS. I am on windows. This could be the difference.

my ini file has only one command which just prints the process id. Nothing else.

Please see updated movies. Another person above was also able to reproduce the crash on windows.

@acer 

" Does it crash if you force only method=contour? "

Yes, It still crashes.

u:=Int(2/3/alpha^(3/2)*sin(1/2*3^(1/2)*ln(alpha))*sin(alpha)*3^(1/2),alpha=0..x,method=contour);
value(u);

#server crash

@acer 

How does one check "without Physics update"? Other than removing it from libname, how to check? I tried below removing Physics update lib from libname, but crash still happens.

 


 

26028

interface(version);

`Standard Worksheet Interface, Maple 2024.0, Windows 10, March 01 2024 Build ID 1794891`

Physics:-Version();

`The "Physics Updates" version in the MapleCloud is 1715 and is the same as the version installed in this computer, created 2024, April 3, 20:27 hours Pacific Time.`

restart;

24896

libname

"C:\Users\Owner\maple\toolbox\2024\Physics Updates\lib", "C:\Program Files\Maple 2024\lib"

libname:="C:\\Program Files\\Maple 2024\\lib"

"C:\Program Files\Maple 2024\lib"

u:=Int(2/3/alpha^(3/2)*sin(1/2*3^(1/2)*ln(alpha))*sin(alpha)*3^(1/2),alpha = 0 .. x)

Int((2/3)*sin((1/2)*3^(1/2)*ln(alpha))*sin(alpha)*3^(1/2)/alpha^(3/2), alpha = 0 .. x)

value(u);


 

Download another_server_crash_on_int_maple_2024_april_4_2024.mw

@acer 

That is correct. verified on Maple 2024

 

integrand:=(2*x^2022+1)/(x^2023+x);
convert(integrand,parfrac);  #kernel crash

windows 10

@ecterrab 

Ah!  I found the problem. few days ago I asked about expansion of exp here Why-Maple-Likes-To-Write-Exp2a-As-expa2-

So I used the suggestion in one of the answers

"There's also a way to suppress the expansion of exp (or any other function) that maintains the suppression until you turn it off; so it's suitable for use in an initialization file:

expand(expandoff()): expandoff(exp):"

And edited my Maple ini file and added the above. Now that you mentioned ini file, I removed the above and the problem is now gone. 

So that was the cause. Here is worksheet that shows it is fixed now after removing the above from my ini file

restart;

142864

interface(version);

`Standard Worksheet Interface, Maple 2024.0, Windows 10, March 01 2024 Build ID 1794891`

Physics:-Version();

`The "Physics Updates" version in the MapleCloud is 1708 and is the same as the version installed in this computer, created 2024, March 27, 16:20 hours Pacific Time.`

expand(expandoff()): expandoff(exp):

integrand:=(d*x)^m/(a+b*arctanh(c*x^n))^2;

(d*x)^m/(a+b*arctanh(c*x^n))^2

int(integrand,x);

Error, (in int/gparse/gmon) too many levels of recursion

restart;

142864

integrand:=(d*x)^m/(a+b*arctanh(c*x^n))^2;

(d*x)^m/(a+b*arctanh(c*x^n))^2

int(integrand,x);

2*x*(c*exp(n*ln(x))-1)*(c*exp(n*ln(x))+1)*exp(m*(ln(d)+ln(x)-((1/2)*I)*Pi*csgn(I*d*x)*(-csgn(I*d*x)+csgn(I*d))*(-csgn(I*d*x)+csgn(I*x))))/(b*c*n*exp(n*ln(x))*(-b*ln(1-c*exp(n*ln(x)))+b*ln(c*exp(n*ln(x))+1)+2*a))+int(-2*exp(m*(ln(d)+ln(x)-((1/2)*I)*Pi*csgn(I*d*x)*(-csgn(I*d*x)+csgn(I*d))*(-csgn(I*d*x)+csgn(I*x))))*(c^2*m*(exp(n*ln(x)))^2+c^2*n*(exp(n*ln(x)))^2+(exp(n*ln(x)))^2*c^2-m+n-1)/(b*c*n*exp(n*ln(x))*(-b*ln(1-c*exp(n*ln(x)))+b*ln(c*exp(n*ln(x))+1)+2*a)), x)

 


Download int_gparse_gmon_error_FIXED_maple_2024_march_27_2024.mw

I had no idea that disabling expansion of exp will cause such internal issue. Thanks for everyone's comments and answers.

@Carl Love 

opps. fixed it now. Was sleepy and did not notice.

@Ronan 

Thanks. But I tried RealDomain and it did not work for me. The code is still on my screen. Here it is 

35788

restart;

35788

n:=3;m:=2;
eqx:=x^(n/m)=a;
use RealDomain in PDEtools:-Solve(eqx,x); end use;
F:=map(X->eval(eqx,X),[%]);
map(X->evalb(X),F);

3

2

x^(3/2) = a

x = a^(2/3), x = (1/4)*a^(2/3)*(1+I*3^(1/2))^2, x = (1/4)*a^(2/3)*(I*3^(1/2)-1)^2

[a = a, (1/16)*4^(1/2)*(a^(2/3)*(1+I*3^(1/2))^2)^(3/2) = a, (1/16)*4^(1/2)*(a^(2/3)*(I*3^(1/2)-1)^2)^(3/2) = a]

[true, false, false]

 

Download why_real_domain_does_not_work_for_me.mw

What is the difference between what I did above and what you did for it to have worked with you?

The strange thing, is that in Mathematica by default it works in complex domain. Maple should be the same. But Mathematica gave x=a^(2/3) as only solution without saying real domain.

It looks like when I used solve() with real domain, it works. But when using PDEtools:-Solve with real domain, it does not work!  is this a bug?

@vv 

I know about simplify(size) ofcourse but I stopped using it long time ago. I had mixed results from it. Sometimes it gives worst result than just simplify. Here is just one example out of 100's I have

138396

restart;

138396

A:=int(sin(x)^(1/3),x)

-((3/2)*I)*2^(2/3)*(-I*((exp(I*x))^2-1)/exp(I*x))^(1/3)+(1/2)*(int(-2/(-((exp(I*x))^2-1)^2*exp(I*x))^(1/3), x))*2^(2/3)*(-I*((exp(I*x))^2-1)/exp(I*x))^(1/3)*(-((exp(I*x))^2-1)^2*exp(I*x))^(1/3)/((exp(I*x))^2-1)

MmaTranslator:-Mma:-LeafCount(simplify(A))

66

MmaTranslator:-Mma:-LeafCount(simplify(A,size))

114

You see the simplify size gives result almost twice as large are just simplify in this example !

I understand that simplification is not easy in CAS. I think it is one of the hardest things to get right, but also simplification in CAS is one of the most important core functionalities. So important to get it right.

It is 50% science and 50% art and there is no one algorithm to follow each time as what is simpler can by a fuzzy concept.

I am finding the best thing is to reduce as much calling simplify and its friends or always check after calling simplify to see if the result should be used or not using some more tests. 

Download simplify_size.mw

@vv 

It looks like my method is not general and does not work for all cases. May be a iterative solutions as you have is simpler than trying to find short cut.

@Thomas Richard 

I did not see the check to turn off infinite scroll and neither did the OP. May it should be off by default?

But when I turned it off, how does one go to "next page"? I see no next page arrow any where to click on.

So user can only now see the first page.

Is this by design or is there a trick to go to next page while infinite scroll is off?

https://www.maplesoft.com/Applications/TopApplications.aspx

@sursumCorda 

However, Mathematica by default does not use LeafCount to rank the complexity

But in https://reference.wolfram.com/language/ref/ComplexityFunction.html

"With the default setting ComplexityFunction->Automatic, forms are ranked primarily according to their LeafCount"

Any way. my simple implementation uses LeafCount for now to keep things simple. A more advanced version can accept user defined function to use for comparison.

@sursumCorda 

thanks for the workarounds. I ended up adding simplify on each result in my code after each call to  solve(identity,....

I am testing this now as there many of them in my code.

Luckily, it seems adding simplify gives back same solution as Maple 2023.2.1. At least on the ones I noticed so far.

Something internal in solve(identity... seems to have changed to cause this much bigger result to be generated in Maple 2024.  Result is still correct but much larger than before.

 

1 2 3 4 5 6 7 Last Page 1 of 71