New Reply

why isn't this a machine sized integer

 
Share Thread
Feb27-12, 10:03 AM   #1
 
Blog Entries: 2

why isn't this a machine sized integer


In:list5 = Import["Composites.csv"]; (*imports a list of odd integers each less than 1000000*)
Timing[f=Compile[{{Caa,_Integer},{S0,_Integer},{S1,_Integer},{Co,_Integer}},
Module[{xCo=Co,xS0=S0,xS1=S1,Temp},While[Temp=Mod[6 xS1-xS0-6,Caa];xCo>0&&Temp>=1,
xS0=xS1;xS1=Temp;xCo--];
xCo]];
count = Length[list5];
list3=Reap[While[count>0,
Caa= list5[[count]];Sow[Caa];Co=Round[(Caa-2)/2];S0=2;S1=3;
So = f[Caa,S0,S1,Co]; Sow[So];
count--]]]
results in red msg:
{3827} in position 1 should be a machine size Integer,
{13067} in position 1 should be a machine size Integer, etc
These are integers from the imported list. I am at loss as how to correct this error. Please help.
PS when I switched the positions of Caa and Co in the compiled function, I get in red {1913} in position 1 should be a machine sized integer, etc.
PhysOrg.com science news on PhysOrg.com

>> City-life changes blackbird personalities, study shows
>> Origins of 'The Hoff' crab revealed (w/ Video)
>> Older males make better fathers: Mature male beetles work harder, care less about female infidelity
Feb27-12, 10:27 AM   #2
 
Is this for a course?

What language is this, what machine and what OS?

machine sized ints vary with cpu and language usage.

For C/C++ a machine sized int +-32768 or 16-bit integer

For Java an int is a 32-bit value.

Also check your csv file to see if any numbers are greater than the 1000000 mentioned in the comment. Thats if this program had worked before but now does not.
Feb27-12, 10:34 AM   #3
 
Mentor
Quote by jedishrfu View Post
Is this for a course?

What language is this, what machine and what OS?

machine sized ints vary with cpu and language usage.

For C/C++ a machine sized int +-32768 or 16-bit integer
For the past dozen years or so, most C or C++ compilers consider an int to be 32 bits.
Quote by jedishrfu View Post

For Java an int is a 32-bit value.

Also check your csv file to see if any numbers are greater than the 1000000 mentioned in the comment. Thats if this program had worked before but now does not.
Feb27-12, 12:33 PM   #4
 
Blog Entries: 2

why isn't this a machine sized integer


Opps. Sorry but this is code for Mathematica. I should have made that clear in the title
PS I understand that integers less than 2^31 may qualify as machine readable. I think the {} may be making the number not be an "integer" but if that is the problem, how do I get rid of them, i.e. let the element be the number per se and not --{3427}--.
Feb27-12, 12:39 PM   #5
 
noticed an extra space in your code is that an artifact of display on this forum or is it really there.

,{S1,_Integer},{Co,_I nteger}},
__________________^_________
maybe this space is causing the trouble
Feb27-12, 01:03 PM   #6
 
Blog Entries: 2
Quote by jedishrfu View Post
noticed an extra space in your code is that an artifact of display on this forum or is it really there.

,{S1,_Integer},{Co,_I nteger}},
__________________^_________
maybe this space is causing the trouble
Thanks, that seems to be a problem with cut and paste from Mathematica into this forum. It is not a space appearing in my code, so that is not the problem. Interesting that when Bill Simpson provided the same code in my earlier thread (Sow within While Loop(mathematica)), that a space was apparently added also. I used the compiled function f with numbers bigger than in my list so size of the numbers is not the problem. It has to be something with the adding of Import and Part of list functions to the code.
Feb27-12, 01:24 PM   #7
 
Remember that long list of warnings and cautions I gave you about Compile[] and one of those was that all integers needed to be less than 32 bits. It looks like you either have a small syntax problem or you have stepped over that limit.

By trial and error see if you can create a MUCH shorter list Composites.csv that still demonstrates this error, post that here and I'll see if I can figure out what you have done. I don't want to have to dig through thousands of numbers to discover where your problem is. But as I also said, when you ran into troubles with a Compile I might easily not be the person to figure it out for you.

And yes, that annoying space and other formatting problems show up in text pasted onto the forum. I mentioned that before. That isn't the problem, just as long as you remember to watch for and fix that. I think that inserting a few spaces after commas or operators sometimes makes the problem hide.
Feb27-12, 01:53 PM   #8
 
Blog Entries: 2
Quote by Bill Simpson View Post
Remember that long list of warnings and cautions I gave you about Compile[] and one of those was that all integers needed to be less than 32 bits. It looks like you either have a small syntax problem or you have stepped over that limit.

By trial and error see if you can create a MUCH shorter list Composites.csv that still demonstrates this error, post that here and I'll see if I can figure out what you have done. I don't want to have to dig through thousands of numbers to discover where your problem is. But as I also said, when you ran into troubles with a Compile I might easily not be the person to figure it out for you.

And yes, that annoying space and other formatting problems show up in text pasted onto the forum. I mentioned that before. That isn't the problem, just as long as you remember to watch for and fix that. I think that inserting a few spaces after commas or operators sometimes makes the problem hide.
Trust me, the numbers are all odd numbers less than 32 bits. Could it be that the syntax is "text" instead of a "number"? What I posted is the complete mathematical code that returns this error message. To check if Mathematical was reading the elements as numbers, I tried 2 list5[[25]] and got out: 6854 which was twice the last element, but I am not sure that mathematica always treats the elements as numeric and not just when coupled with multiplication. As for a short list to import, try to import 43,51,59,67,75 and see what output you get.
Feb27-12, 07:02 PM   #9
 
Quote by ramsey2879 View Post
Trust me, the numbers are all odd numbers less than 32 bits.
After decades and decades of experience I don't trust me, let alone anyone else or software.

Quote by ramsey2879 View Post
Could it be that the syntax is "text" instead of a "number"? What I posted is the complete mathematical code that returns this error message. To check if Mathematical was reading the elements as numbers, I tried 2 list5[[25]] and got out: 6854 which was twice the last element, but I am not sure that mathematica always treats the elements as numeric and not just when coupled with multiplication. As for a short list to import, try to import 43,51,59,67,75 and see what output you get.
Using Excel 5 I created a column of 43,51,59,67,75 as you directed and saved it as Composites.csv. When I ran your code I got:

CompiledFunction:: cfsa: Argument {75} at position 1 should be a machine-size integer.

and got that for each integer. That looks EXACTLY like the error message you are reporting.

I then inserted

Print[FullForm[list5]];

immediately after your list5=Import[...] and discovered that a column of integers comes in as a list of lists when you are expecting list5 to be just a list so I am guessing that your input from Excel is not in the form that you are trusting it to be.

Try inserting

list5 = Flatten[list5];

after your list5=Import[...] and see if you get a better result.
Feb28-12, 01:44 AM   #10
 
Admin
Spaces are inserted every 50 characters into strings without white spaces by the forum software.
New Reply

Similar discussions for: why isn't this a machine sized integer
Thread Forum Replies
To understand how spherical harmonics apply to integer spin but not half integer spin Quantum Physics 10
Angular momentum - integer or half-integer Quantum Physics 2
Proof Question: Prove integer + 1/2 is not an integer Calculus & Beyond Homework 4
Material in machine tool body (milling machine) Materials & Chemical Engineering 0