Email: Password: Remember Me | Create Account (Free)

Back to Subject List

Old thread has been locked -- no new posts accepted in this thread
???
02/14/07 15:47
Read: times


 
#132923 - slightly less
Responding to: ???'s previous message
Erik Malund said:
silabs goes as high as 100MHz at 1 clock = 100 times faster than "the standard"

Slightly less as the instruction cycle per instruction is different for the 1- (and 4-) clockers. However, the SiLabs perform remarkably well, according to my findings it's the best what you can get among the 1-clockers.
My "benchmark" indicates something around 90-times improvement over the 12MHz/12clocker "vanilla", that's around 23 times faster than the 'S8253 would be at 22.118400MHz. These figures are slightly overoptimistic as they assume branch cache hit each time, so the real number would be maybe around 80 times for the vanilla. And of course it depends on the particular instruction mix and maybe other circumstances (e.g. external external memory access), we all know the value of benchmarks...

JW


List of 68 messages in thread
TopicAuthorDate
Speed Recommendation            01/01/70 00:00      
   Time for processing            01/01/70 00:00      
   Get a 1X Chip            01/01/70 00:00      
      another 485 issue            01/01/70 00:00      
         slightly less            01/01/70 00:00      
            who gives a hoot            01/01/70 00:00      
               maybe yes, maybe not...            01/01/70 00:00      
   Choose the appropriate tool!            01/01/70 00:00      
      Don't have much of a choice!            01/01/70 00:00      
         drop the "coding standards"            01/01/70 00:00      
         Yikes.            01/01/70 00:00      
            down memory lane            01/01/70 00:00      
               a possible compromise...            01/01/70 00:00      
                  Well...            01/01/70 00:00      
               The other end of the spectrum ...            01/01/70 00:00      
                  afer quite a few iterations....            01/01/70 00:00      
                     The guide to writing unmaintainable code says:            01/01/70 00:00      
         That's the kind of nonsense...            01/01/70 00:00      
   put some load on the hardware            01/01/70 00:00      
      Jan, oh Jan            01/01/70 00:00      
      9-bit - more work for the hardware!            01/01/70 00:00      
         true, but...            01/01/70 00:00      
            It is...            01/01/70 00:00      
               calculate the ISR after making the ISR skip            01/01/70 00:00      
   Optimizing C            01/01/70 00:00      
      just one example            01/01/70 00:00      
         The advantage is...            01/01/70 00:00      
            If the thinking had preceded the coding ...            01/01/70 00:00      
               you miss the point            01/01/70 00:00      
                  then what's the chance of having him to C-craft?            01/01/70 00:00      
                     zero            01/01/70 00:00      
                     Job security, of course.            01/01/70 00:00      
      Examples are not equivalent            01/01/70 00:00      
         Wasn\'t saying that they were.            01/01/70 00:00      
            I just tried SDCC...            01/01/70 00:00      
               Hm, can you post the assembly output ?            01/01/70 00:00      
                  Keil does this            01/01/70 00:00      
                     Interesting.            01/01/70 00:00      
                     Optimizing and Keil/SDCC/Raisonance            01/01/70 00:00      
                        Retort            01/01/70 00:00      
                           Register Banks...            01/01/70 00:00      
                              benchmarks            01/01/70 00:00      
                              tools differ            01/01/70 00:00      
                              Clarification            01/01/70 00:00      
                                 comments            01/01/70 00:00      
                                    not to write things twice            01/01/70 00:00      
                                       no need            01/01/70 00:00      
                                          And how does this produce two incarnations?            01/01/70 00:00      
                                             like this            01/01/70 00:00      
                                                hehe            01/01/70 00:00      
                                                   this is no better confusion-wise            01/01/70 00:00      
                                                      what is "efficient"            01/01/70 00:00      
                                                         ok is \'safe\' better?            01/01/70 00:00      
                                                            How to convey our ideas to the mcu?            01/01/70 00:00      
                                                               Carefull now, you could end with Prolog :)            01/01/70 00:00      
                                                                  my kids            01/01/70 00:00      
                                                            Huge code != huge time            01/01/70 00:00      
                                                               This comes under the heading of "famous last words            01/01/70 00:00      
                        the question I always raise on this            01/01/70 00:00      
                           Here's Why...            01/01/70 00:00      
                  SDCC            01/01/70 00:00      
                     Thanks. Also interesting.            01/01/70 00:00      
                     SDCC peephole optimiser conflict            01/01/70 00:00      
                  Raisonance            01/01/70 00:00      
   help me ,I NEED SOME HELP ON MY PROJECT            01/01/70 00:00      
      that is NOT 'help'            01/01/70 00:00      
      Well that is not the way to get it!            01/01/70 00:00      
      start a new thread!            01/01/70 00:00      

Back to Subject List