Quantic Kernel r6
Official-Ayrton990
released this
03 May 23:55
·
1879 commits
to QK-ten
since this release
Update is out!
Changelog:
- Partially fixed dt2w, some users reporting not working on PA but the rest of the ROMs works
- Upstreamed to linux stable 4.14.178
- Added a new CFS thread for freeing GPU memory. Improves performance when tearing down processes and lets UI to submit work to the GPU. (more info here: 24c27d8)
- Offload mementry destroy work to separate thread, this is done to reduce the time spent by user applications in waiting for memory to be freed. (more info here: 0dc7bcd)
- Use optimized spinning loop for arm64, this makes about 20% performance improvement on the CPU multithread load. (more info here: 9abedaf)
- Fix race between runtime distribution and assignment (more info here: 0d521cd)
- Align rq->avg_idle and rq->avg_scan_cost, The only difference between the two is that rq->avg_scan_cost is computed using a pure division rather than a shift (more info here: f3721e1)
- Get rid of the big mutex lock on ashmem (more info here: 3ec8f26)
- Enable ZRAM to 3GB and use lz4 as compressor.
- Most of the new features were backported.