Commit a11661c
Skip pcNodeIndex work buffer stream when GPU sorting is active (#8492)
The GPU sort + culling path uses interval-based compaction which reads
node visibility directly from intervals, never accessing pcNodeIndex
from the work buffer. Only the CPU sort + culling path needs it.
Convert gpuSorting to a getter/setter and unify the pcNodeIndex stream
decision into _syncNodeIndexStream(), which adds/removes the stream
based on both culling and gpuSorting state.
Made-with: Cursor
Co-authored-by: Martin Valigursky <mvaligursky@snapchat.com>1 parent 78d557c commit a11661c
1 file changed
+44
-11
lines changed| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
124 | 124 | | |
125 | 125 | | |
126 | 126 | | |
| 127 | + | |
| 128 | + | |
| 129 | + | |
| 130 | + | |
| 131 | + | |
| 132 | + | |
127 | 133 | | |
128 | 134 | | |
129 | 135 | | |
130 | 136 | | |
131 | 137 | | |
132 | 138 | | |
133 | | - | |
| 139 | + | |
| 140 | + | |
| 141 | + | |
| 142 | + | |
| 143 | + | |
| 144 | + | |
| 145 | + | |
| 146 | + | |
| 147 | + | |
| 148 | + | |
| 149 | + | |
| 150 | + | |
| 151 | + | |
| 152 | + | |
| 153 | + | |
| 154 | + | |
134 | 155 | | |
135 | 156 | | |
136 | 157 | | |
| |||
226 | 247 | | |
227 | 248 | | |
228 | 249 | | |
229 | | - | |
230 | | - | |
231 | | - | |
232 | | - | |
233 | | - | |
234 | | - | |
235 | | - | |
236 | | - | |
237 | | - | |
238 | | - | |
| 250 | + | |
| 251 | + | |
| 252 | + | |
239 | 253 | | |
240 | 254 | | |
241 | 255 | | |
| |||
249 | 263 | | |
250 | 264 | | |
251 | 265 | | |
| 266 | + | |
| 267 | + | |
| 268 | + | |
| 269 | + | |
| 270 | + | |
| 271 | + | |
| 272 | + | |
| 273 | + | |
| 274 | + | |
| 275 | + | |
| 276 | + | |
| 277 | + | |
| 278 | + | |
| 279 | + | |
| 280 | + | |
| 281 | + | |
| 282 | + | |
| 283 | + | |
| 284 | + | |
252 | 285 | | |
253 | 286 | | |
254 | 287 | | |
| |||
0 commit comments