search for: c_r

...ments/specrend/specrend.c I could probably port it in R, but I've been hoping to use compiled code for a while and this looks like a good test case. Unfortunately, the only success I've had so far is with the too simple factorial(int *n) function, http://math.acadiau.ca/ACMMaC/howtos/C_R.html Here, however, the code has several functions, some of which use structures as input variables and return several values. I've created a shared library with, > R CMD SHLIB specrend.c and then ran, dyn.load('specrend.so') > .C('spectrum_to_xyz', spec_intens=as...

...s for my software development approach. I hope others find this information useful as well. I have attached the code files in the bottom. Below I provide a summary output, and interpret the results. [All times in table below are in microseconds] m/n/time_.Call_R/time_.Call_C++/.Call_overhead/time_.C_R/time_.C_C++/.C_overhead 100/10/3.23/1.33/*1.90*/16.89/0.96/*15.93* 200/15/5.38/3.44/*1.94*/30.77/2.81/*27.96* 500/20/12.48/10.4/*2.08*/78.12/9.37/*68.75* Interpretation: 1- .Call overhead holds nearly constant, i.e. independent of data size. This is expected since .Call works by passing pointers....

...zr`~(5(mOmBYN{Oo at +|TObKkIbyLxD)cx=Kr3;q#0xq917 at nGn zRBaeLza9{y3Ahr-a7)}RlAhP`?PDp`t*&0j+`|(H7BzjObWF#_ at M;VGCj5K8_T%GE zp)p=0fJ6ei&sCYOGAkAVlLb{CNo?apErHWG0JkInt-MD`yfXby4&|Oc3AdCBOpXG4 zkpOl*!*q3pqm7<AUt3Eq at J&=sQxUZXSBVW(*SqvwHHm=>>_PXbg;OHw4(5PskZUHa z{y0F-mVFRnc_Rs?3SdWEf$hL`lh&1HGe?Zk1n-5H`uCMSC)~3 at e_Gcb$x?<nCtkI* ziK7&JpgK_`ee|m%Kt}?)Zw9R;!&@@>SMpY0y%jQz>>YhbU_Pla>N_329vlu%UH{4< z`2Fjot7Zc>@Ka7s8;QOn4&tupQ(SD|9)JhZWq?YKbKPs+;d7)+h5ZdZ5#Cvr@<Oc- z(uaKo80~w0T~5m0TumgvkbOY-&PiiBM4?~eP?h-dbuKN3ZUVEDFvQnM+{6^CJG_0t...