On Thu, 2021-02-25 at 19:05 +0800, Steven Liu wrote: > Artem Galin <artem.ga...@gmail.com> 于2021年2月24日周三 下午11:46写道： > > > > Hello, > > > > > > > > Please find the proposal for the following two projects for GSoC this year. > > > > The FATE project had been proposed earlier already but I don't know why it > > didn't happen. > > > > I previously got the feedback from Thilo Borgmann and would be happy to get > > feedback from the community as well. > > > > > > > > """"""""""""""""" > > > > > > > > FATE tests for HW acceleration > > > > > > > > Description: FFmpeg's FATE regression test system do not include validation > > of hardware acceleration. This project is about to include these > > hardware-specific regression tests to the FATE system. Since this project > > requires access to corresponding hardware, it is not expected to cover the > > whole range of hardware accelerators there are but to initialize some > > useful tests for available hardware, most likely to include CPU > > acceleration on Intel or AMD at least. Access to other hardware can also be > > provided during the project, depending on the students requirements. > > > > > > > > Expected results: Implement tests for the different hardware decoding, > > encoding and transcoding capabilities there are. > > > > > > > > Prerequisites: Good C and Makefile coding skills, basic familiarity with > > Git. > > > > > > > > Qualification Task: TBD > > > > > > > > Mentor: Artem G (artem.galin [at] intel [dot] com) > > > > > > > > Backup Mentor: TBD > > > > > > > > """"""""""""""""" > > > > > > > > > > > > Title: oneAPI Video Processing Library (oneVPL) for CPU & GPU hardware > > accelleration (https://www.oneapi.com/) > > > > > > > > Description: Add support for oneVPL > > > > > > > > The oneAPI project is developing a common cross-platform API for video > > processing, including CPU and GPU hardware acceleration. > > > > > > > > At least Intel's MediaSDK (https://github.com/Intel-Media-SDK/MediaSDK < > > https://github.com/Intel-Media-SDK/MediaSDK>;), providing the currently used > > QSV API used for HW acceleration on Intel, will transition to implement the > > oneAPI API, called oneVPL (https://github.com/oneapi-src/oneVPL). > > > > > > > > Therefore, this project aims to get FFmpeg ready for this new API to ensure > > future availability of not only Intel's HW acceleration after the > > transition to oneVPL but also for possible other implementations of the > > oneAPI from other vendors. > > > > > > > > > > > > Expected results: Implement a oneVPL library wrapper into FFmpeg. > > > > Validate that media pipelines work with no regressions by supported feature > > set and performance. > > > > > > > > > > > > Prerequisites: Good C and Makefile coding skills, basic familiarity with > > Git. > > > > > > > > Qualification Task: TBD > > > > > > > > Mentor: Artem G (artem.galin [at] intel [dot] com) > > > > > > > > Backup Mentor: TBD > > > > > > > > > > > > """"""""""""""""" > > > > > > > > Thanks, > > > > Artem. > > _______________________________________________ > > ffmpeg-devel mailing list > > ffmpeg-devel@ffmpeg.org > > https://ffmpeg.org/mailman/listinfo/ffmpeg-devel > > > > To unsubscribe, visit link above, or email > > ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe". > > Do you mean you want base on this patch: > http://ffmpeg.org/pipermail/ffmpeg-devel/2020-September/269905.html
I refreshed the patches and prepared a new patchset since then. The new series will be sent out after the new oneVPL release. We are discussing internally if this is a good proposal for GSOC Thanks Haihao > _______________________________________________ > ffmpeg-devel mailing list > ffmpeg-devel@ffmpeg.org > https://ffmpeg.org/mailman/listinfo/ffmpeg-devel > > To unsubscribe, visit link above, or email > ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe". _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-devel To unsubscribe, visit link above, or email ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
