This is an old revision of the document!
Sometimes we want to delay a variable number of cycles, e.g. during VSP or when setting up a stable raster.
A typical way of handling a variable delay is to use a clock slide. In the following example, we start somewhere in a given range of cycles, we want to end up exactly at cycle 50, and we've computed the corresponding number of cycles to skip in A.
; delay 19-A cycles sta branch+1 ; 31..41 (A in range 0..10) branch bpl * ; 35..45 lda #$a9 ; 38 lda #$a9 ; 40 lda #$a9 ; 42 lda #$a9 ; 44 lda $eaa5 ; 46 ; at cycle 50
Every additional byte we skip corresponds to one less cycle of delay. Notice that the operands are reinterpreted as opcodes depending on where we land on the slide. For instance, the final
lda $eaa5 is encoded as
ad a5 ea and takes 4 cycles. Skipping the first byte, we get
a5 ea (
lda $ea, 3 cycles). Skipping also the second byte, we get
nop, 2 cycles).
The length of the clock slide depends on the maximum jitter we have to support, and we pay a corresponding penalty in the form of useless waiting cycles. In the example, the maximum supported jitter is 10 cycles (41-31), and the minimum overhead cost is 9 cycles (50-41).
Now here comes the improvement:
Notice that in the latest case (starting at cycle 41), we still execute a single
nop instruction in the clock slide. This is because there is no single-cycle instruction on the 6502. Jumping one byte further would reduce the number of cycles by two. That is, jumping directly to the
at cycle 50 comment would actually get us there one cycle too early.
But we can use the page-crossing penalty of the branch instruction to add an extra cycle in this particular case! Consider:
; delay 18-A cycles sta branch+1 ; 31..41 (A in range 0..10) branch bpl * ; 35..45 .byt $a9 ; 38 lda #$a9 ; 39 lda #$a9 ; 41 lda #$a9 ; 43 lda $eaa5 ; 45 ; page-crossing here! ; at cycle 49
The clock slide is now one byte shorter, and the minimum overhead cost has been reduced to 8 cycles.
The downside is that we now have an alignment requirement. Sometimes it may not be possible to adjust the starting address of the delay code. But note that we can insert dummy bytes just after the branch instruction, as long as we update the computation of A accordingly.