perf(engine): a --gpu render probes the GPU encoders once, not twice (#6638)

* perf(engine,cli): a machine's GPU encoder probe runs once, not on every render

* refactor(engine): one shared probe per process, no answer kept on disk
This commit is contained in:
Miguel Ángel
2026-10-28 08:44:28 +00:00
committed by GitHub
parent e6476c2cd5
commit 8e63faf846
6 changed files with 37 additions and 8 deletions
+1 -1
View File
@@ -186,7 +186,7 @@ export function getEncoderPreset(
}
// Re-export GPU utilities so existing consumers that import from chunkEncoder still work.
export { detectGpuEncoder, type GpuEncoder } from "../utils/gpuEncoder.js";
export { detectGpuEncoder, getCachedGpuEncoder, type GpuEncoder } from "../utils/gpuEncoder.js";
/** The `lockGopForChunkConcat` / `gopSize` pair, shared by every encoder entry point. */
export interface LockedGopOptions {