Skip to content
Permalink

Comparing changes

Choose two branches to see what’s changed or to start a new pull request. If you need to, you can also or learn more about diff comparisons.

Open a pull request

Create a new pull request by comparing changes across two branches. If you need to, you can also . Learn more about diff comparisons here.
base repository: whitequark/libfx2
Failed to load repositories. Confirm that selected base ref is valid, then try again.
Loading
base: 9603d4529c4d
Choose a base ref
...
head repository: whitequark/libfx2
Failed to load repositories. Confirm that selected head ref is valid, then try again.
Loading
compare: afaae3ce52fc
Choose a head ref
  • 2 commits
  • 1 file changed
  • 1 contributor

Commits on May 20, 2019

  1. Fix some comment typos.

    @whitequark
    whitequark committed May 20, 2019
    Copy the full SHA
    6bb8e07 View commit details

Commits on May 21, 2019

  1. Incorporate even better nop optimization by 8051Enthusiast. 

    https://twitter.com/8051Enthusiast/status/1130803396229967872
    @whitequark
    whitequark committed May 21, 2019
    Copy the full SHA
    afaae3c View commit details
Showing with 27 additions and 25 deletions.
  1. +27 −25 firmware/library/delay.c
52 changes: 27 additions & 25 deletions firmware/library/delay.c
View file
Original file line number Diff line number Diff line change
@@ -2,41 +2,43 @@
#include <fx2regs.h>
#include <bits/asmargs.h>

// Spin for exactly min(24, DP*4) cycles (i.e. min(96, DP*16) clocks), including the call of
// this function. The implementation is a bit complicated, but the calculations in the callee
// Spin for exactly max(24, DP*4) cycles (i.e. max(96, DP*16) clocks), including the call of
// this function. The implementation is a bit complicated, but the calculations in the caller
// are simpler.
//
// This was originally optimized by @whitequark, and then improved by @8051Enthusiast.
void delay_4c(uint16_t count) __naked {
count;
__asm
; (ljmp delay_4c) ; 0c > 4c
mov acc, dph ; 3c
cjne a, #0, 00000$ ; 4c > 11c [A]
mov acc, dpl ; 3c
subb a, #(24/4+1) ; 2c
inc a ; 1c
jnc 00003$ ; 3c > 20c [B]
_ASM_RET ; 4c > 24c
mov a, dph ; 2c
jnz 00001$ ; 3c > 9c [A]
mov a, dpl ; 2c
add a, #-(20/4+1) ; 2c
jc 00000$ ; 3c > 16c [B]
_ASM_RET ; 4c > 20c
00000$:
mov a, dpl ; [A] > 2c
jz 00002$ ; 3c > 16c [C]
cjne a, #0, 00004$ ; [B] > 4c > 20c [H]
_ASM_RET ; 4c > 24c
00001$:
djnz acc, 00001$ ; [C] > 4Lc > (16+4L)c [D]
; [E] > 1016c > (16+1024H+4L)c [G]
mov a, dpl ; [A] > 2c
jz 00003$ ; 3c > 14c [C]
00002$:
djnz acc, 00002$ ; [C] > 4Lc > (14+4L)c [D]
; [E] > 1016c > (14+1024H+4L)c [G]
00003$:
nop ; [D] > 1c
mov acc, #-2 ; 3c
djnz dph, 00001$ ; 4c > (16+8+4L) [E]
nop ; 1c
mov acc, #-((28+4)/4) ; 3c > (28+4L) [F]
djnz dph, 00002$ ; 4c > (14+8+4L) [E]
mov a, #-((24+4)/4) ; 2c > (24+4L) [F]
; [G] > 1c
; 3c
; 4c > (16+8+1024H+4L)c [E]
; 1c
; 3c > (28+1024H+4L)c [F]
00003$:
djnz acc, 00003$ ; [B] > 4Lc
; 4c > (14+8+1024H+4L)c [E]
; 2c > (24+1024H+4L)c [F]
00004$:
djnz acc, 00004$ ; [H] > 4Lc
_ASM_RET ; 4c > (24+4L)c
; [F] > 992c
; [F] > 996c
; 4c > (1024H+4L)c
__endasm;
}
@@ -156,7 +158,7 @@ void delay_us(uint16_t count) __naked {
count;
__asm;
; (mov dptr, #?) ; 3c
; (ljmp delay_us) ; 4c
; (lcall delay_us) ; 4c
mov a, #17 ; 2c
push acc ; 2c
lcall _delay_us_overhead
@@ -194,14 +196,14 @@ void delay_ms(uint16_t count) __naked {
mov dptr, #1000 ; 3c
lcall _delay_us_overhead

sjmp 00000$ ; 3c
sjmp 00000$ ; 3c => 17c

00002$:
dec sp

// epilog
pop ar6
pop ar7
_ASM_RET ; 4c => 17c
_ASM_RET
__endasm;
}