Compare commits

..

1 Commits

Author SHA1 Message Date
Adriano 6704d66cd5 feat: kernel JIT batch top-max-per-variant (opt-in)
Nuovo kernel _jit_top_max_per_variant: prange esterno sulle varianti
invece di n_vars chiamate JIT separate via ThreadPoolExecutor.
Wrapper Python top_max_per_variant prepara buffer flat (offsets +
dx_flat/dy_flat/bins_flat) e bg per scala.

Default batch_top=False perche su benchmark realistici (Linux 13 core,
72-180 varianti) ThreadPoolExecutor + kernel singolo che rilascia GIL
e gia ottimale. Path batch_top=True utile come opzione per scenari
con n_vars >>> n_threads o overhead chiamate JIT dominante.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-04 15:35:51 +02:00
2 changed files with 143 additions and 30 deletions
+114
View File
@@ -159,6 +159,63 @@ if HAS_NUMBA:
acc[y, x] = 0.0 acc[y, x] = 0.0
return acc return acc
@nb.njit(cache=True, parallel=True, fastmath=True, boundscheck=False)
def _jit_top_max_per_variant(
spread: np.ndarray, # uint8 (H, W)
dx_flat: np.ndarray, # int32 (sum_N,)
dy_flat: np.ndarray, # int32 (sum_N,)
bins_flat: np.ndarray, # int8 (sum_N,)
offsets: np.ndarray, # int32 (n_vars+1,) prefix sum
bit_active: np.uint8,
bg_per_variant: np.ndarray, # float32 (n_vars, H, W) - 1 per scala
scale_idx: np.ndarray, # int32 (n_vars,) idx in bg_per_variant
) -> np.ndarray:
"""Batch: per ogni variante calcola max score (rescored bg), ritorna
array float32 (n_vars,). Parallelismo prange ESTERNO sulle varianti
elimina overhead di n_vars chiamate JIT separate (avg ~20us per
chiamata su template piccoli) + pool thread Python.
Pensato per fase TOP del pruning quando n_vars >> n_threads.
"""
n_vars = offsets.shape[0] - 1
H, W = spread.shape
out = np.zeros(n_vars, dtype=np.float32)
for vi in nb.prange(n_vars):
i0 = offsets[vi]; i1 = offsets[vi + 1]
N = i1 - i0
if N == 0:
out[vi] = -1.0
continue
si = scale_idx[vi]
inv = nb.float32(1.0 / N)
best = nb.float32(-1.0)
for y in range(H):
for x in range(W):
s = nb.float32(0.0)
for k in range(N):
b = bins_flat[i0 + k]
mask = np.uint8(1) << b
if (bit_active & mask) == 0:
continue
ddy = dy_flat[i0 + k]
yy = y + ddy
if yy < 0 or yy >= H:
continue
ddx = dx_flat[i0 + k]
xx = x + ddx
if xx < 0 or xx >= W:
continue
if spread[yy, xx] & mask:
s += nb.float32(1.0)
s *= inv
bgv = bg_per_variant[si, y, x]
if bgv < 1.0:
r = (s - bgv) / (1.0 - bgv + 1e-6)
if r > best:
best = r
out[vi] = best if best > 0.0 else 0.0
return out
@nb.njit(cache=True, parallel=True, fastmath=True, boundscheck=False) @nb.njit(cache=True, parallel=True, fastmath=True, boundscheck=False)
def _jit_popcount_density(spread: np.ndarray) -> np.ndarray: def _jit_popcount_density(spread: np.ndarray) -> np.ndarray:
"""Conta bit set per pixel: ritorna (H, W) float32 in [0..8].""" """Conta bit set per pixel: ritorna (H, W) float32 in [0..8]."""
@@ -185,6 +242,12 @@ if HAS_NUMBA:
_jit_score_bitmap(spread, dx, dy, b, np.uint8(0xFF)) _jit_score_bitmap(spread, dx, dy, b, np.uint8(0xFF))
bg = np.zeros((32, 32), dtype=np.float32) bg = np.zeros((32, 32), dtype=np.float32)
_jit_score_bitmap_rescored(spread, dx, dy, b, np.uint8(0xFF), bg) _jit_score_bitmap_rescored(spread, dx, dy, b, np.uint8(0xFF), bg)
offsets = np.array([0, 1], dtype=np.int32)
scale_idx = np.zeros(1, dtype=np.int32)
bg_pv = np.zeros((1, 32, 32), dtype=np.float32)
_jit_top_max_per_variant(
spread, dx, dy, b, offsets, np.uint8(0xFF), bg_pv, scale_idx,
)
_jit_popcount_density(spread) _jit_popcount_density(spread)
else: # pragma: no cover else: # pragma: no cover
@@ -198,6 +261,12 @@ else: # pragma: no cover
def _jit_score_bitmap_rescored(spread, dx, dy, bins, bit_active, bg): def _jit_score_bitmap_rescored(spread, dx, dy, bins, bit_active, bg):
raise RuntimeError("numba non disponibile") raise RuntimeError("numba non disponibile")
def _jit_top_max_per_variant(
spread, dx_flat, dy_flat, bins_flat, offsets, bit_active,
bg_per_variant, scale_idx,
):
raise RuntimeError("numba non disponibile")
def _jit_popcount_density(spread): def _jit_popcount_density(spread):
raise RuntimeError("numba non disponibile") raise RuntimeError("numba non disponibile")
@@ -246,6 +315,51 @@ def score_bitmap_rescored(
return np.maximum(0.0, out).astype(np.float32) return np.maximum(0.0, out).astype(np.float32)
def top_max_per_variant(
spread: np.ndarray,
dx_list: list, dy_list: list, bin_list: list,
bg_per_scale: dict,
variant_scales: list,
bit_active: int,
) -> np.ndarray:
"""Wrapper: prepara buffer flat e chiama kernel batch su tutte le varianti.
Parallelismo Numba prange-esterno sulle varianti (n_vars >> n_threads
tipicamente per top-pruning) → meglio del thread-pool Python che paga
overhead di n_vars chiamate JIT separate.
"""
if not HAS_NUMBA or len(dx_list) == 0:
return np.array([], dtype=np.float32)
n_vars = len(dx_list)
sizes = [len(d) for d in dx_list]
offsets = np.zeros(n_vars + 1, dtype=np.int32)
offsets[1:] = np.cumsum(sizes)
total = int(offsets[-1])
dx_flat = np.empty(total, dtype=np.int32)
dy_flat = np.empty(total, dtype=np.int32)
bins_flat = np.empty(total, dtype=np.int8)
for vi, (dx, dy, bn) in enumerate(zip(dx_list, dy_list, bin_list)):
i0 = int(offsets[vi]); i1 = int(offsets[vi + 1])
dx_flat[i0:i1] = dx
dy_flat[i0:i1] = dy
bins_flat[i0:i1] = bn
# bg per variante: indicizzato per scala
scales_unique = sorted(bg_per_scale.keys())
scale_to_idx = {s: i for i, s in enumerate(scales_unique)}
H, W = spread.shape
bg_pv = np.empty((len(scales_unique), H, W), dtype=np.float32)
for s, idx in scale_to_idx.items():
bg_pv[idx] = bg_per_scale[s]
scale_idx = np.array(
[scale_to_idx[s] for s in variant_scales], dtype=np.int32,
)
return _jit_top_max_per_variant(
np.ascontiguousarray(spread, dtype=np.uint8),
dx_flat, dy_flat, bins_flat, offsets, np.uint8(bit_active),
bg_pv, scale_idx,
)
def popcount_density(spread: np.ndarray) -> np.ndarray: def popcount_density(spread: np.ndarray) -> np.ndarray:
if HAS_NUMBA: if HAS_NUMBA:
return _jit_popcount_density(np.ascontiguousarray(spread, dtype=np.uint8)) return _jit_popcount_density(np.ascontiguousarray(spread, dtype=np.uint8))
+21 -22
View File
@@ -40,6 +40,7 @@ from pm2d._jit_kernels import (
score_by_shift as _jit_score_by_shift, score_by_shift as _jit_score_by_shift,
score_bitmap as _jit_score_bitmap, score_bitmap as _jit_score_bitmap,
score_bitmap_rescored as _jit_score_bitmap_rescored, score_bitmap_rescored as _jit_score_bitmap_rescored,
top_max_per_variant as _jit_top_max_per_variant,
popcount_density as _jit_popcount, popcount_density as _jit_popcount,
HAS_NUMBA, HAS_NUMBA,
) )
@@ -239,8 +240,6 @@ class LineShapeMatcher:
self._train_mask = mask_full.copy() self._train_mask = mask_full.copy()
self.variants.clear() self.variants.clear()
# Invalida cache feature di refine: il template e cambiato.
self._refine_feat_cache = {}
for s in self._scale_list(): for s in self._scale_list():
sw = max(16, int(round(w * s))) sw = max(16, int(round(w * s)))
sh = max(16, int(round(h * s))) sh = max(16, int(round(h * s)))
@@ -435,24 +434,9 @@ class LineShapeMatcher:
H, W = spread0.shape H, W = spread0.shape
margin = 3 margin = 3
# Cache template features per angolo (chiave: int(round(ang*20)) =
# bucket di 0.05°). Golden-search ricontratto puo richiedere lo
# stesso bucket piu volte; evita re-warp+gradient+extract (costoso).
# Cache a livello matcher per riusare tra chiamate find() su scene
# diverse: la rotazione del template non dipende dalla scena.
if not hasattr(self, '_refine_feat_cache'):
self._refine_feat_cache = {}
feat_cache = self._refine_feat_cache
cache_scale_key = round(scale * 1000)
def _score_at_angle(off: float) -> tuple[float, float, float]: def _score_at_angle(off: float) -> tuple[float, float, float]:
"""Ritorna (score, best_cx, best_cy) per angolo = angle_deg + off.""" """Ritorna (score, best_cx, best_cy) per angolo = angle_deg + off."""
ang = angle_deg + off ang = angle_deg + off
ck = (round(ang * 20), cache_scale_key)
cached = feat_cache.get(ck)
if cached is not None:
fx, fy, fb = cached
else:
M = cv2.getRotationMatrix2D(center, ang, 1.0) M = cv2.getRotationMatrix2D(center, ang, 1.0)
gray_r = cv2.warpAffine(gray_p, M, (diag, diag), gray_r = cv2.warpAffine(gray_p, M, (diag, diag),
flags=cv2.INTER_LINEAR, flags=cv2.INTER_LINEAR,
@@ -461,10 +445,6 @@ class LineShapeMatcher:
flags=cv2.INTER_NEAREST, borderValue=0) flags=cv2.INTER_NEAREST, borderValue=0)
mag, bins = self._gradient(gray_r) mag, bins = self._gradient(gray_r)
fx, fy, fb = self._extract_features(mag, bins, mask_r) fx, fy, fb = self._extract_features(mag, bins, mask_r)
# LRU semplice: limita cache a ~256 angoli (8 angoli * 32 candidati)
if len(feat_cache) > 256:
feat_cache.pop(next(iter(feat_cache)))
feat_cache[ck] = (fx, fy, fb)
if len(fx) < 8: if len(fx) < 8:
return (0.0, cx, cy) return (0.0, cx, cy)
dx = (fx - center[0]).astype(np.int32) dx = (fx - center[0]).astype(np.int32)
@@ -595,6 +575,7 @@ class LineShapeMatcher:
verify_threshold: float = 0.4, verify_threshold: float = 0.4,
coarse_angle_factor: int = 2, coarse_angle_factor: int = 2,
scale_penalty: float = 0.0, scale_penalty: float = 0.0,
batch_top: bool = False,
) -> list[Match]: ) -> list[Match]:
""" """
scale_penalty: se > 0, riduce lo score per match a scala diversa da 1.0: scale_penalty: se > 0, riduce lo score per match a scala diversa da 1.0:
@@ -678,7 +659,25 @@ class LineShapeMatcher:
kept_coarse: list[tuple[int, float]] = [] kept_coarse: list[tuple[int, float]] = []
all_top_scores: list[tuple[int, float]] = [] all_top_scores: list[tuple[int, float]] = []
if self.n_threads > 1 and len(coarse_idx_list) > 1: # batch_top: usa kernel batch single-call con prange-esterno su
# varianti. Vince su threadpool quando n_vars >> n_threads e quando
# H*W top e' piccolo (overhead chiamate JIT > costo kernel).
if (batch_top and HAS_NUMBA and len(coarse_idx_list) > 4):
dx_l = []; dy_l = []; bn_l = []; vs_l = []
for vi in coarse_idx_list:
var = self.variants[vi]
lvl = var.levels[min(top, len(var.levels) - 1)]
dx_l.append(lvl.dx); dy_l.append(lvl.dy); bn_l.append(lvl.bin)
vs_l.append(var.scale)
scores_arr = _jit_top_max_per_variant(
spread_top, dx_l, dy_l, bn_l, bg_cache_top, vs_l,
bit_active_top,
)
for vi, best in zip(coarse_idx_list, scores_arr.tolist()):
all_top_scores.append((vi, best))
if best >= top_thresh:
kept_coarse.append((vi, best))
elif self.n_threads > 1 and len(coarse_idx_list) > 1:
with ThreadPoolExecutor(max_workers=self.n_threads) as ex: with ThreadPoolExecutor(max_workers=self.n_threads) as ex:
for vi, best in ex.map(_top_score, coarse_idx_list): for vi, best in ex.map(_top_score, coarse_idx_list):
all_top_scores.append((vi, best)) all_top_scores.append((vi, best))