Pelles C forum
C language => Tips & tricks => Topic started by: TimoVJL on August 12, 2017, 09:23:20 AM
-
Fork of QD (2.3.17) (https://github.com/scibuilder/QD) dll
c_test.c from it & QD.dll & QD.lib import library.
EDIT: missing headers and libs added.
EDIT QDSSE.dll for 32-bit.
-
That looks very interesting, thanks ;-)
The *.lib files are just import libs. How useful is that, if you need the DLL anyway?
-
Because it is C++ code, dll is obvious choice.
-
thanks you :)
-
hello TimoVJL
is there a way for you to compile the 32-bit version using sse?, then you won't have to worry about fpu_fix_start/fpu_fix_end.
I edited the makefile in QD-master\src and added -mfpmath=sse -msse2 to CPPFLAGS = -O2 -mfpmath=sse -msse2
but that's when using gcc, not sure how you would do that in PellesC.
btw, that's a very small-sized dll, only 115k, mine compiled with gcc was 1.2 meg even after stripping.
-
It was compiled with msvc 2010 with dynamic dlls.
It uses msvcp100.dll.
I added SSE version, with cl option -arch:SSE too.
-
unfortunately it's still using the fpu.
-
TimoVJL, did you have to edit any of the source files in order for vs to compile the project?
I have visual studio 2013-pro and also later community editions but am not able to setup a dll project that will compile
I have an older project, qd-2.3.7 and the 32-bit version compiles ok but if I replace the source files with the newer version then I have problems, something like max is not defined in the namespace std and can't convert qd_read to int.
-
No.
build cmd attached.
-
thank you
but still the same problem when using vs-2013
src\qd_real.cpp(446) : error C2039: 'max' : is not a member of 'std'
src\qd_real.cpp(446) : error C2440: '=' : cannot convert from 'qd_real' to 'int'
I have vs-2010 somewhere.
-
so you have to add
#if _MSC_VER >= 18
#include <algorithm>
#endif to dd_real.cpp and qd_real.cpp
-
you are the man :)
works like a charm, thank you.
-
tank you TimoVJL :)
having a C version means we can have a static lib instead of a dll, will keep an eye on your progress.
-
hello TimoVJL :)
just for fun I tried to compile the source files with gcc and there are no complaints, however if I try to compile to dll then there are three errors
undefined reference to `c_dd_error'
undefined reference to `c_dd_div_d_dd'
undefined reference to `dd_real_eps'
-
thank you
-
hello TimoVJL
I tested c_dd_dll_ms and here's what I found
c_dd_add ok
c_dd_sub ok
c_dd_mul ok
c_dd_div ok
c_dd_sqr ok
c_dd_sqrt ok
c_dd_write ok
c_dd_read ok
the trig, log and exp functions give wrong results, c_dd_atan gives right result only for arg=1, c_dd_tan is missing.
sorry for the bad report, I just thought that you might like to have some feedback. :)
-
hello TimoVJL
I didn't think about this earlier, but this type of arithmetic is sensitive to the FPU setting, remember that in the original code there were two functions to set the FPU flags, I bet that's where the problem is, or if you can coerce the compiler to use sse2 instructions then you should be ok.
-
this type of arithmetic is sensitive to the FPU setting, remember that in the original code there were two functions to set the FPU flags, I bet that's where the problem is, or if you can coerce the compiler to use sse2 instructions then you should be ok.
But don't forget LDMXCSR ;)
-
hello TimoVJL
you are getting close, however there's a bug in dd_to_digits, if dd_variable.x[0] = 0.1 and dd_variable.x[1] = -5.551115123125983e-018 then you get an error and crash.
(dd_to_digits): can't compute exponent.
the trigs seem to be ok, but only made a very simple test.
-
there's a problem with dd_cos, if dd_x = 1/n , n=11, 12, 13 ... then dd_cos fails, for dd_x=1/11 the result is
dd_x.x[0] dd_x[1]
cos(1/11) 0.9958706137005629 0.9958706137005629
it should be something like
cos(1/11) 0.9958706137005629 -4.772639498746044e-017
[edit] dd_log is not working
-
I did a loop from -70 to 70 and used mpfr for verification, for log/exp the loop was from 1 to 70
sin, cos, tan, atan, log and exp are ok, not sure about the hyperbolic functions.
-
dd_atanh(dd_tanh(x)) from 1 to 19 ok, above 19 dd_atanh says "Argument out of domain."
-
we really need a good test suite to check the functions in the supported range, the hyperbolic functions are trickier to check.
-
I made a small test and the trigs -not including hyperbolic- , log and exp seem to be ok
I have been doing my testing in freebasic because I am more familiar with it than C.