dymokomi/luce-kernel
Source of release 0.1.1
0.1.1 / src / engine / lanes_compare.lucb
# Generated by tools/lanes.py from its tables; edit those, not this file.
# Compares (mask results) and ternary lane operations.
## Run a compare or ternary instruction; never faults.
func compare_or_ternary(w: Worker*, ins: Instruction) -> usize:
match ins.op:
.eq:
match ins.kind:
.f32: eq_f32(w.masks(ins.dst), w.f32s(ins.a), w.f32s(ins.b))
.f64: eq_f64(w.masks(ins.dst), w.f64s(ins.a), w.f64s(ins.b))
.i32: eq_i32(w.masks(ins.dst), w.i32s(ins.a), w.i32s(ins.b))
.i64: eq_i64(w.masks(ins.dst), w.i64s(ins.a), w.i64s(ins.b))
.u32: eq_u32(w.masks(ins.dst), w.u32s(ins.a), w.u32s(ins.b))
.u64: eq_u64(w.masks(ins.dst), w.u64s(ins.a), w.u64s(ins.b))
.mask: eq_mask(w.masks(ins.dst), w.masks(ins.a), w.masks(ins.b))
_: return w.n
.ne:
match ins.kind:
.f32: ne_f32(w.masks(ins.dst), w.f32s(ins.a), w.f32s(ins.b))
.f64: ne_f64(w.masks(ins.dst), w.f64s(ins.a), w.f64s(ins.b))
.i32: ne_i32(w.masks(ins.dst), w.i32s(ins.a), w.i32s(ins.b))
.i64: ne_i64(w.masks(ins.dst), w.i64s(ins.a), w.i64s(ins.b))
.u32: ne_u32(w.masks(ins.dst), w.u32s(ins.a), w.u32s(ins.b))
.u64: ne_u64(w.masks(ins.dst), w.u64s(ins.a), w.u64s(ins.b))
.mask: ne_mask(w.masks(ins.dst), w.masks(ins.a), w.masks(ins.b))
_: return w.n
.lt:
match ins.kind:
.f32: lt_f32(w.masks(ins.dst), w.f32s(ins.a), w.f32s(ins.b))
.f64: lt_f64(w.masks(ins.dst), w.f64s(ins.a), w.f64s(ins.b))
.i32: lt_i32(w.masks(ins.dst), w.i32s(ins.a), w.i32s(ins.b))
.i64: lt_i64(w.masks(ins.dst), w.i64s(ins.a), w.i64s(ins.b))
.u32: lt_u32(w.masks(ins.dst), w.u32s(ins.a), w.u32s(ins.b))
.u64: lt_u64(w.masks(ins.dst), w.u64s(ins.a), w.u64s(ins.b))
_: return w.n
.le:
match ins.kind:
.f32: le_f32(w.masks(ins.dst), w.f32s(ins.a), w.f32s(ins.b))
.f64: le_f64(w.masks(ins.dst), w.f64s(ins.a), w.f64s(ins.b))
.i32: le_i32(w.masks(ins.dst), w.i32s(ins.a), w.i32s(ins.b))
.i64: le_i64(w.masks(ins.dst), w.i64s(ins.a), w.i64s(ins.b))
.u32: le_u32(w.masks(ins.dst), w.u32s(ins.a), w.u32s(ins.b))
.u64: le_u64(w.masks(ins.dst), w.u64s(ins.a), w.u64s(ins.b))
_: return w.n
.gt:
match ins.kind:
.f32: gt_f32(w.masks(ins.dst), w.f32s(ins.a), w.f32s(ins.b))
.f64: gt_f64(w.masks(ins.dst), w.f64s(ins.a), w.f64s(ins.b))
.i32: gt_i32(w.masks(ins.dst), w.i32s(ins.a), w.i32s(ins.b))
.i64: gt_i64(w.masks(ins.dst), w.i64s(ins.a), w.i64s(ins.b))
.u32: gt_u32(w.masks(ins.dst), w.u32s(ins.a), w.u32s(ins.b))
.u64: gt_u64(w.masks(ins.dst), w.u64s(ins.a), w.u64s(ins.b))
_: return w.n
.ge:
match ins.kind:
.f32: ge_f32(w.masks(ins.dst), w.f32s(ins.a), w.f32s(ins.b))
.f64: ge_f64(w.masks(ins.dst), w.f64s(ins.a), w.f64s(ins.b))
.i32: ge_i32(w.masks(ins.dst), w.i32s(ins.a), w.i32s(ins.b))
.i64: ge_i64(w.masks(ins.dst), w.i64s(ins.a), w.i64s(ins.b))
.u32: ge_u32(w.masks(ins.dst), w.u32s(ins.a), w.u32s(ins.b))
.u64: ge_u64(w.masks(ins.dst), w.u64s(ins.a), w.u64s(ins.b))
_: return w.n
.mul_add:
match ins.kind:
.f32: mul_add_f32(w.f32s(ins.dst), w.f32s(ins.a), w.f32s(ins.b), w.f32s(ins.c))
.f64: mul_add_f64(w.f64s(ins.dst), w.f64s(ins.a), w.f64s(ins.b), w.f64s(ins.c))
_: return w.n
.lerp:
match ins.kind:
.f32: lerp_f32(w.f32s(ins.dst), w.f32s(ins.a), w.f32s(ins.b), w.f32s(ins.c))
.f64: lerp_f64(w.f64s(ins.dst), w.f64s(ins.a), w.f64s(ins.b), w.f64s(ins.c))
_: return w.n
.clamp:
match ins.kind:
.f32: clamp_f32(w.f32s(ins.dst), w.f32s(ins.a), w.f32s(ins.b), w.f32s(ins.c))
.f64: clamp_f64(w.f64s(ins.dst), w.f64s(ins.a), w.f64s(ins.b), w.f64s(ins.c))
.i32: clamp_i32(w.i32s(ins.dst), w.i32s(ins.a), w.i32s(ins.b), w.i32s(ins.c))
.i64: clamp_i64(w.i64s(ins.dst), w.i64s(ins.a), w.i64s(ins.b), w.i64s(ins.c))
.u32: clamp_u32(w.u32s(ins.dst), w.u32s(ins.a), w.u32s(ins.b), w.u32s(ins.c))
.u64: clamp_u64(w.u64s(ins.dst), w.u64s(ins.a), w.u64s(ins.b), w.u64s(ins.c))
_: return w.n
.smooth:
match ins.kind:
.f32: smooth_f32(w.f32s(ins.dst), w.f32s(ins.a), w.f32s(ins.b), w.f32s(ins.c))
.f64: smooth_f64(w.f64s(ins.dst), w.f64s(ins.a), w.f64s(ins.b), w.f64s(ins.c))
_: return w.n
.fit01:
match ins.kind:
.f32: fit01_f32(w.f32s(ins.dst), w.f32s(ins.a), w.f32s(ins.b), w.f32s(ins.c))
.f64: fit01_f64(w.f64s(ins.dst), w.f64s(ins.a), w.f64s(ins.b), w.f64s(ins.c))
_: return w.n
_: return w.n
return w.n
func eq_f32(d: u8[], a: const f32[], b: const f32[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] == b[i])
func eq_f64(d: u8[], a: const f64[], b: const f64[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] == b[i])
func eq_i32(d: u8[], a: const i32[], b: const i32[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] == b[i])
func eq_i64(d: u8[], a: const i64[], b: const i64[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] == b[i])
func eq_u32(d: u8[], a: const u32[], b: const u32[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] == b[i])
func eq_u64(d: u8[], a: const u64[], b: const u64[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] == b[i])
func eq_mask(d: u8[], a: const u8[], b: const u8[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] == b[i])
func ne_f32(d: u8[], a: const f32[], b: const f32[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] != b[i])
func ne_f64(d: u8[], a: const f64[], b: const f64[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] != b[i])
func ne_i32(d: u8[], a: const i32[], b: const i32[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] != b[i])
func ne_i64(d: u8[], a: const i64[], b: const i64[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] != b[i])
func ne_u32(d: u8[], a: const u32[], b: const u32[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] != b[i])
func ne_u64(d: u8[], a: const u64[], b: const u64[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] != b[i])
func ne_mask(d: u8[], a: const u8[], b: const u8[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] != b[i])
func lt_f32(d: u8[], a: const f32[], b: const f32[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] < b[i])
func lt_f64(d: u8[], a: const f64[], b: const f64[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] < b[i])
func lt_i32(d: u8[], a: const i32[], b: const i32[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] < b[i])
func lt_i64(d: u8[], a: const i64[], b: const i64[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] < b[i])
func lt_u32(d: u8[], a: const u32[], b: const u32[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] < b[i])
func lt_u64(d: u8[], a: const u64[], b: const u64[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] < b[i])
func le_f32(d: u8[], a: const f32[], b: const f32[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] <= b[i])
func le_f64(d: u8[], a: const f64[], b: const f64[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] <= b[i])
func le_i32(d: u8[], a: const i32[], b: const i32[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] <= b[i])
func le_i64(d: u8[], a: const i64[], b: const i64[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] <= b[i])
func le_u32(d: u8[], a: const u32[], b: const u32[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] <= b[i])
func le_u64(d: u8[], a: const u64[], b: const u64[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] <= b[i])
func gt_f32(d: u8[], a: const f32[], b: const f32[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] > b[i])
func gt_f64(d: u8[], a: const f64[], b: const f64[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] > b[i])
func gt_i32(d: u8[], a: const i32[], b: const i32[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] > b[i])
func gt_i64(d: u8[], a: const i64[], b: const i64[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] > b[i])
func gt_u32(d: u8[], a: const u32[], b: const u32[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] > b[i])
func gt_u64(d: u8[], a: const u64[], b: const u64[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] > b[i])
func ge_f32(d: u8[], a: const f32[], b: const f32[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] >= b[i])
func ge_f64(d: u8[], a: const f64[], b: const f64[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] >= b[i])
func ge_i32(d: u8[], a: const i32[], b: const i32[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] >= b[i])
func ge_i64(d: u8[], a: const i64[], b: const i64[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] >= b[i])
func ge_u32(d: u8[], a: const u32[], b: const u32[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] >= b[i])
func ge_u64(d: u8[], a: const u64[], b: const u64[]):
assert(a.length == d.length and b.length == d.length)
for i in 0..<d.length: d[i] = (u8)(a[i] >= b[i])
func mul_add_f32(d: f32[], a: const f32[], b: const f32[], c: const f32[]):
assert(a.length == d.length and b.length == d.length and c.length == d.length)
for i in 0..<d.length: d[i] = a[i].mul_add(b[i], c[i])
func mul_add_f64(d: f64[], a: const f64[], b: const f64[], c: const f64[]):
assert(a.length == d.length and b.length == d.length and c.length == d.length)
for i in 0..<d.length: d[i] = a[i].mul_add(b[i], c[i])
func lerp_f32(d: f32[], a: const f32[], b: const f32[], c: const f32[]):
assert(a.length == d.length and b.length == d.length and c.length == d.length)
for i in 0..<d.length: d[i] = lib.lerp(a[i], b[i], c[i])
func lerp_f64(d: f64[], a: const f64[], b: const f64[], c: const f64[]):
assert(a.length == d.length and b.length == d.length and c.length == d.length)
for i in 0..<d.length: d[i] = lib.lerp64(a[i], b[i], c[i])
func clamp_f32(d: f32[], a: const f32[], b: const f32[], c: const f32[]):
assert(a.length == d.length and b.length == d.length and c.length == d.length)
for i in 0..<d.length: d[i] = b[i] if a[i] < b[i] else (c[i] if a[i] > c[i] else a[i])
func clamp_f64(d: f64[], a: const f64[], b: const f64[], c: const f64[]):
assert(a.length == d.length and b.length == d.length and c.length == d.length)
for i in 0..<d.length: d[i] = b[i] if a[i] < b[i] else (c[i] if a[i] > c[i] else a[i])
func clamp_i32(d: i32[], a: const i32[], b: const i32[], c: const i32[]):
assert(a.length == d.length and b.length == d.length and c.length == d.length)
for i in 0..<d.length: d[i] = b[i] if a[i] < b[i] else (c[i] if a[i] > c[i] else a[i])
func clamp_i64(d: i64[], a: const i64[], b: const i64[], c: const i64[]):
assert(a.length == d.length and b.length == d.length and c.length == d.length)
for i in 0..<d.length: d[i] = b[i] if a[i] < b[i] else (c[i] if a[i] > c[i] else a[i])
func clamp_u32(d: u32[], a: const u32[], b: const u32[], c: const u32[]):
assert(a.length == d.length and b.length == d.length and c.length == d.length)
for i in 0..<d.length: d[i] = b[i] if a[i] < b[i] else (c[i] if a[i] > c[i] else a[i])
func clamp_u64(d: u64[], a: const u64[], b: const u64[], c: const u64[]):
assert(a.length == d.length and b.length == d.length and c.length == d.length)
for i in 0..<d.length: d[i] = b[i] if a[i] < b[i] else (c[i] if a[i] > c[i] else a[i])
func smooth_f32(d: f32[], a: const f32[], b: const f32[], c: const f32[]):
assert(a.length == d.length and b.length == d.length and c.length == d.length)
for i in 0..<d.length: d[i] = lib.smooth(a[i], b[i], c[i])
func smooth_f64(d: f64[], a: const f64[], b: const f64[], c: const f64[]):
assert(a.length == d.length and b.length == d.length and c.length == d.length)
for i in 0..<d.length: d[i] = lib.smooth64(a[i], b[i], c[i])
func fit01_f32(d: f32[], a: const f32[], b: const f32[], c: const f32[]):
assert(a.length == d.length and b.length == d.length and c.length == d.length)
for i in 0..<d.length: d[i] = lib.fit01(a[i], b[i], c[i])
func fit01_f64(d: f64[], a: const f64[], b: const f64[], c: const f64[]):
assert(a.length == d.length and b.length == d.length and c.length == d.length)
for i in 0..<d.length: d[i] = lib.fit01_64(a[i], b[i], c[i])