While simulating the Cygnal C8051F020 microcontroller, I found the 'states count' is not accurate. Branching operations have a different cycle count depending on if the branch is taken and this is not correctly simulated. Is this an error? If so, will it be fixed?
hi could please explain clearly so that i can provide some solution to you john
For example, the following loop results in two different timings for the processing of the DJNZ operation of the Cygnal processors: MOV R1, #2 LOOP: NOP DJNZ R1, LOOP When R1 != 0, the timing is 3 cycles for the DJNZ operation, whereas when R1 = 0 the timing for the DJNZ is 2 cycles. This is a feature of the micros from Cygnal where execution time is optimized. You can verify this via thier datasheets. I see that the DCYG.DLL library provided with C51 has exported routines that deal with state timing and such. I thought maybe the DLL incorrectly processed the Cygnal operations. I'm trying to use the simulator for accurate timing measurements.