From: nma@12000.org   
      
   On 4/1/2017 3:04 AM, clicliclic@freenet.de wrote:   
      
   > On the /index.htm page you are linking to /reports/rubi_4_11/index.htm,   
   > which latter page did not seem to exist when I looked.   
   >   
      
   I had a small problem and had to upload everything again. It is all   
   there now.   
      
   > I am pleased to see my earlier guess about the integrator performance   
   > ("expect both Sympy and Giac/Xcas to succeed on somewhere between 50%   
   > and 70% of these textbook problems") borne out for Sympy on the   
   > Timofeev suite.   
   >   
      
   > Apparently, this run already implements a distinction between timeout   
   > and other failures. Readers, however, would surely like to know the   
   > time limit actually employed in this run.   
   >   
      
   It is 5 minutes. It is mentioned at the start of each report   
   and at the start of the main index page as well.   
      
   "A time limit of 5 minutes is used for all integrals in each CAS.   
   If the integration does not complete within this time limit   
   then the integral is considered to have failed."   
      
   ALso each F says if it was due to time-out of not now.   
   If it is F(-1), means it is due to time-out.   
      
      
   I think if this is run on PC running Linux with lots of   
   RAM, sympy performance should improve. Sympy was run   
   inside virtual machine, with 2 GB Ram.  I had to run   
   it in Linux, since for time out, had to use signals   
   and timers outside of sympy to control timeout. And   
   this only works on Linux. So will try to run this   
   again on a PC running native Linux.   
      
   Sympy itself does not support timeouts like Maple   
   and Mathematica do, and so user has to implement   
   time-out themselves outside of sympy using signals   
   and catching time-out exceptions, else sympy will   
   just hang without this.   
      
   One nice thing about sympy result I noticed, is that   
   it gives answers as piecwise when needed. This is a   
   nice touch which is not implemented in other CAS I   
   know about. For example Timofeev #50 and many others   
   like it:   
      
   integrate(S('x/sqrt(a**4-x**4)'),x)   
      
   gives   
      
   Piecewise((-I*acosh(x**2/a**2)/2, Abs(x**4/a**4) > 1), (asin(x**2/a**2)/2,   
   True))   
      
   While other CAS do not give Piecewise result. I thought this   
   was nice feature of sympy.   
      
   --Nasser   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)