ack/mach/i80/libem/rst.s

#
.sect .text
.sect .rom
.sect .data
.sect .bss
.sect .text

! Which resets we install are determined by statistical analysis of Star
! Trek. When changing these, make sure to update the i80 table to match.
!     97 call .floadn2
!     41 call .floadn4
!     34 call .fload4
!     28 call .fstoren2
!
! Also:
!     48 call .cmps_mag

.define .rst_init
.rst_init:
    lxi h, .floadn2
    lxi d, 0x0008
    call copy
    lxi h, .floadn4
    call copy
    lxi h, .fload4
    call copy
    lxi h, .fstoren2
    call copy
    lxi h, .cmps_mag
    jmp copy

! Copies eight bytes from HL to DE.
copy:
    mvi c, 8
.1:
    mov a, m
    stax d
    inx h
    inr e
    dcr c
    jnz .1
    ret
Experimentally use rst 1 to optimise frame variable loads. Reduces Star Trek from 43892 to 43205 bytes. 2019-02-06 23:06:32 +00:00			`#`
			`.sect .text`
			`.sect .rom`
			`.sect .data`
			`.sect .bss`
			`.sect .text`

Instead of using parameterised rsts for stack access, add a huge swathe of automatically built helper tools. Star Trek goes up from 40243 to 40779 bytes, but should be a lot faster. 2019-02-13 21:45:22 +00:00			`! Which resets we install are determined by statistical analysis of Star`
			`! Trek. When changing these, make sure to update the i80 table to match.`
			`! 97 call .floadn2`
			`! 41 call .floadn4`
			`! 34 call .fload4`
			`! 28 call .fstoren2`
Call cmps rather than cmpu for bge... make cmps and cmpu rsts. Apply various other tiny optimisations. 2019-02-14 23:38:39 +00:00			`!`
			`! Also:`
			`! 48 call .cmps_mag`
Instead of using parameterised rsts for stack access, add a huge swathe of automatically built helper tools. Star Trek goes up from 40243 to 40779 bytes, but should be a lot faster. 2019-02-13 21:45:22 +00:00
Experimentally use rst 1 to optimise frame variable loads. Reduces Star Trek from 43892 to 43205 bytes. 2019-02-06 23:06:32 +00:00			`.define .rst_init`
			`.rst_init:`
Instead of using parameterised rsts for stack access, add a huge swathe of automatically built helper tools. Star Trek goes up from 40243 to 40779 bytes, but should be a lot faster. 2019-02-13 21:45:22 +00:00			`lxi h, .floadn2`
			`lxi d, 0x0008`
			`call copy`
			`lxi h, .floadn4`
			`call copy`
			`lxi h, .fload4`
			`call copy`
			`lxi h, .fstoren2`
Call cmps rather than cmpu for bge... make cmps and cmpu rsts. Apply various other tiny optimisations. 2019-02-14 23:38:39 +00:00			`call copy`
			`lxi h, .cmps_mag`
Instead of using parameterised rsts for stack access, add a huge swathe of automatically built helper tools. Star Trek goes up from 40243 to 40779 bytes, but should be a lot faster. 2019-02-13 21:45:22 +00:00			`jmp copy`

			`! Copies eight bytes from HL to DE.`
			`copy:`
			`mvi c, 8`
			`.1:`
rst 1 now takes an eight-bit parameter, not a 16 bit one. Reduces Star Trek again from 43205 to 42339 bytes. 2019-02-07 07:45:20 +00:00			`mov a, m`
Instead of using parameterised rsts for stack access, add a huge swathe of automatically built helper tools. Star Trek goes up from 40243 to 40779 bytes, but should be a lot faster. 2019-02-13 21:45:22 +00:00			`stax d`
Experimentally use rst 1 to optimise frame variable loads. Reduces Star Trek from 43892 to 43205 bytes. 2019-02-06 23:06:32 +00:00			`inx h`
Fix nearly all the comparisons, a lot of which were broken. Using a rst for unsigned comparisons is surprisingly not that useful due to marshalling overhead; it's only four bytes to do inline (plus jc), or six for a constant. Also add some useful top optimisations. Star Trek goes from 39890 to 39450 bytes. 2019-02-15 20:30:27 +00:00			`inr e`
Instead of using parameterised rsts for stack access, add a huge swathe of automatically built helper tools. Star Trek goes up from 40243 to 40779 bytes, but should be a lot faster. 2019-02-13 21:45:22 +00:00			`dcr c`
			`jnz .1`
			`ret`