intel/vpl-gpu-rt intel-onevpl-26.1.5

on GitHub

This release of Intel® VPL GPU Runtime is part of open source release of media stack. Component revisions included into package:

What’s New

• Decode Surface Allocation Fix: Fixed surface allocation for H264, H265, and VVC decoders —
  minimum surface count changed from dpbSize + 1 + asyncDepth to dpbSize + 2 * asyncDepth to
  prevent surface starvation at async depth > 1.
• Decode Error Code Accuracy: Fixed error code returned for empty work surface in AV1, VVC, and
  legacy decoders — now returns MFX_ERR_INVALID_VIDEO_PARAM instead of MFX_ERR_LOCK_MEMORY for more
  accurate diagnostics.
• AV1 Decode Robustness: Added null pointer check for input parameter in AV1 CheckFourCCProfile
  to prevent potential crash on invalid input.
• VA Profile Coverage: Extended VA_PROFILE bit definitions for broader profile coverage.
• API Alignment: Relaxed frame width and height alignment requirement in API description for
  struct mfxFrameInfo.

Full Changelog: intel-onevpl-25.4.6...intel-onevpl-26.1.5