the problem with benchmarks is that every manufacture can write code that makes their processor look good and others look bad. if you are a 'c' programmer more can be done with a good compiler, but not often can a 'c' compiler create code that is tighter or faster that a good assembly programmer. don't get mad walter here is your plug. writing and debugging in 'c' is much faster than coding in assembly. what must be done is to use the best tools available to get the job done with the least amount of effort. and don't be fooled by clock speeds. which is faster at a simple instruction like incrementing a register, a 4MHz PIC or a 12MHz 8051? answer neither, they both take 1uS to complete the instruction, see a pic uses 4 clock cycles per instruction and the 8051 uses 12 cycles. if you have a benchmark the does a lot of multiplying a 8051 will beat a pic because the 8051 have a mult instruction and the pic must run a software multiply, see how you can make any processor look good. i review the project requirements then choose the processor and language to best complete the job. no use trying to stuff a square peg in a round hole. michael -----Original Message----- From: Bob Blick To: PICLIST@MITVMA.MIT.EDU Date: Thursday, March 19, 1998 3:51 AM Subject: benchmarks >A friend asked me about the performance of pics, and other than the usual >"gee, they are darn fast!" I was unable to give him any solid answers. >Being a C programmer, he was interested in SPEC95 or related ones. > >Has anyone here seen any results, or seen source for some standard >benchmarks that are know to run on pics? Walter? Clyde? > >Thanks, > >Bob > >http://www.bobblick.com/ >