I plan to use STM32 as benchmark to profile against. see testing/profile_stm32.txt for profiling scripts.
These are the initial results. I believe verify_ircapture performs much worse than before currently(performance improvement should not be ~2x). parport trunk rev 1675 ====================== step 336 mem 160 erase 3076 flash fill 32754 verify_ircapture disable step 114 mem 96 erase 1547 flash fill 15564 To run tests copy & paste this into telnet: poll off set before [flush_count] step set step_count [expr [flush_count]-$before] set before [flush_count] mdb 0 128 set mem_count [expr [flush_count]-$before] set before [flush_count] flash erase_address 0x8000000 0x10000 set erase_count [expr [flush_count]-$before] set before [flush_count] flash fillb 0x8000000 0x55 0x10000 set flash_fill_count [expr [flush_count]-$before] puts "counts" ; puts "step $step_count" ; puts "mem $mem_count" ; puts "erase $erase_count" ; puts "flash fill $flash_fill_count" -- Øyvind Harboe Embedded software and hardware consulting services http://consulting.zylin.com _______________________________________________ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
