Someone sent me this test code (little modified by me):
Code: Select all
DefType.f
v=3568464845
y=3447
x=v/y
MessageRequester("Calc",StrF(x),0)
Regards
André
*** German PureBasic Support ***
Value range of Floats ?
-
BackupUser
- PureBasic Guru
- Posts: 16777133
- Joined: Tue Apr 22, 2003 7:42 pm
Restored from previous forum. Originally posted by Andre.
Someone sent me this test code (little modified by me):
What is the exact possible value range of float variables? (Manual says "unlimited".)
Regards
André
*** German PureBasic Support ***
Someone sent me this test code (little modified by me):
Regards
André
*** German PureBasic Support ***
-
BackupUser
- PureBasic Guru
- Posts: 16777133
- Joined: Tue Apr 22, 2003 7:42 pm
Restored from previous forum. Originally posted by geoff.
I don't think this is a range problem.
I think it's a string to float conversion bug for the 3568464845.
Single floats give an accuracy of about 6 decimal digits in the number range 1E-35 to 1e35
(ie 1 followed by 35 zeros at the high end)
Doubles give an accuracy of about 15 decimal digits in the number range 1e-300 to 1e300.
The exact values depend on the implementation used.
On a modern processor with integrated hardware co-processor there is negligible speed penalty in using doubles for all floating point calculations.
I don't think this is a range problem.
I think it's a string to float conversion bug for the 3568464845.
Single floats give an accuracy of about 6 decimal digits in the number range 1E-35 to 1e35
(ie 1 followed by 35 zeros at the high end)
Doubles give an accuracy of about 15 decimal digits in the number range 1e-300 to 1e300.
The exact values depend on the implementation used.
On a modern processor with integrated hardware co-processor there is negligible speed penalty in using doubles for all floating point calculations.