Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

branchless #16

Open
wants to merge 2 commits into
base: master
Choose a base branch
from
Open

branchless #16

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
b12c7eb
branchless
alion02 Dec 18, 2024
f3f3781
oops screwed up the signs
alion02 Dec 18, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
41 changes: 17 additions & 24 deletions src/day18.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -44,25 +44,26 @@ unsafe fn inner1(s: &[u8]) -> u32 {
let mut ptr = s.as_ptr().cast::<i8x16>();
let lut = &LUT;

static mut MAP: [i8; 73 * 72 / 8] = [-1; 73 * 72 / 8];
static mut MAP: [i8; 73 * 72 / 8] = [0; 73 * 72 / 8];

let map = MAP.as_mut_ptr();
for i in 0..23 {
map.add(72 / 8 + i * 72 / 8 * 3)
.cast::<i8x32>()
.write_unaligned(i8x32::from_array([
0, 0, 0, 0, 0, 0, 0, 0, -128, 0, 0, 0, 0, 0, 0, 0, 0, -128, 0, 0, 0, 0, 0, 0, 0, 0, -128, -1, -1, -1,
-1, -1,
!0, !0, !0, !0, !0, !0, !0, !0, !-128, !0, !0, !0, !0, !0, !0, !0, !0, !-128, !0, !0, !0, !0, !0, !0,
!0, !0, !-128, !-1, !-1, !-1, !-1, !-1,
]));
}
map.add(69 * 72 / 8).cast::<i8x32>().write_unaligned(i8x32::from_array([
0, 0, 0, 0, 0, 0, 0, 0, -128, 0, 0, 0, 0, 0, 0, 0, 0, -128, 0, 0, 0, 0, 0, 0, 0, 0, -128, -1, -1, -1, -1, -1,
!0, !0, !0, !0, !0, !0, !0, !0, !-128, !0, !0, !0, !0, !0, !0, !0, !0, !-128, !0, !0, !0, !0, !0, !0, !0, !0,
!-128, !-1, !-1, !-1, !-1, !-1,
]));

macro_rules! bts {
macro_rules! btr {
($idx:expr) => {
asm!(
"bts dword ptr[{map} + {offset}], {idx:e}",
"btr dword ptr[{map} + {offset}], {idx:e}",
map = in(reg) map,
idx = in(reg) $idx,
offset = const 72 / 8,
Expand All @@ -82,8 +83,8 @@ unsafe fn inner1(s: &[u8]) -> u32 {
let chunk: u32x4 = _mm_madd_epi16(chunk, u16x8::from_array([72, 1, 72, 1, 72, 1, 72, 1]).into()).into();
let p1 = chunk[0];
let p2 = chunk[1];
bts!(p1);
bts!(p2);
btr!(p1);
btr!(p2);
ptr = ptr.byte_add(step as usize);
}

Expand All @@ -94,29 +95,21 @@ unsafe fn inner1(s: &[u8]) -> u32 {
asm!(
"30:",
"lea {next:e}, [{pos} + 1]",
"bts dword ptr[{map}], {next:e}",
"jc 20f",
"btr dword ptr[{map}], {next:e}",
"mov word ptr[{front} + {j} * 2], {next:x}",
"inc {j:l}",
"20:",
"adc {j:l}, 0",
"lea {next:e}, [{pos} + 72]",
"bts dword ptr[{map}], {next:e}",
"jc 20f",
"btr dword ptr[{map}], {next:e}",
"mov word ptr[{front} + {j} * 2], {next:x}",
"inc {j:l}",
"20:",
"adc {j:l}, 0",
"lea {next:e}, [{pos} - 1]",
"bts dword ptr[{map}], {next:e}",
"jc 20f",
"btr dword ptr[{map}], {next:e}",
"mov word ptr[{front} + {j} * 2], {next:x}",
"inc {j:l}",
"20:",
"adc {j:l}, 0",
"lea {next:e}, [{pos} - 72]",
"bts dword ptr[{map}], {next:e}",
"jc 20f",
"btr dword ptr[{map}], {next:e}",
"mov word ptr[{front} + {j} * 2], {next:x}",
"inc {j:l}",
"20:",
"adc {j:l}, 0",
"cmp {i:l}, {k:l}",
"jne 20f",
"mov {k:e}, {j:e}",
Expand Down
Loading