From: Ioannis V. <no...@ho...> - 2002-07-27 04:02:25
|
> -----Original Message----- > From: min...@li... > [mailto:min...@li...] On Behalf Of > Sternbach, William [IT] > Sent: Friday, July 26, 2002 10:04 PM > To: 'Bob Friesenhahn' > Cc: 'min...@li...' > Subject: RE: [Mingw-users] Program to calculate the MIPS of > your computer (how fast it is). > > > Bob, > > I guess I was most interested in a "C" language 32 bit multiplication. > > My benchmark program is customized on purpose to consider 1 > instruction > as a "C" language 32 bit multiplication. > This multiplication probably translates into many assembler > instructions. > But I am more interested in how many multiplications per > second my computer > can do (than how many assembler instructions it can do), > perhaps you can say my program measures MMPS. > MIPS stands for Million Instructions per second. > MMPS stands for Million Multiplications per second. > > Writing this benchmark was a fun experience for me. > I'm going to run try compiling it and running it on several different > platforms: > Sun Ultra Spark, Mainframe, PC, Apple Macintosh, etc. > > It will be interesting to compare the results to find out > which is the best > platform > to do number crunching programs which do alot of math. Well benchamrking the CPU cannot be reliable with this program especially in a multitasking OS. I agree that you can find differences among compilers though but only on the subject you are interested in & measure, in this case loops with some multiplications. Ioannis * Ioannis Vranos * Programming pages: http://www.noicys.cjb.net * Alternative URL: http://run.to/noicys |