On Wed, 2004-12-29 at 18:29, Lee Revell wrote: > On Wed, 2004-12-29 at 17:30 -0800, Fernando Lopez-Lezcano wrote: > > On Wed, 2004-12-29 at 16:52, Lee Revell wrote: > > > On Wed, 2004-12-29 at 16:21 -0800, Fernando Lopez-Lezcano wrote: > > > > On Wed, 2004-12-29 at 15:10, Lee Revell wrote: > > > > > I tried JACK today with vanilla 2.6.10 and had excellent reaults - it > > > > > works with 32 and 64 frames, which previously required Ingo's patches. > > > > > Many of the latency fixes have been going upstream, and it looks like we > > > > > are finally showing some results. I think we may finally have a kernel > > > > > that's usable out of the box for low latency audio. > > > > > > > > > > Can someone else try to verify these results? > > > > > > > > I'll try to asap. A Planet CCRMA fellow dweller just posted some very > > > > rough comparison tests between several kernels and 2.6.10 was doing > > > > quite well. The 0.7.33-04 DESKTOP_PREEMPT was doing much better but was > > > > also less stable (but he was testing it with extremely low buffer > > > > sizes). > > > > > > This was on the CCRMA list? Got a URL? > > > > http://ccrma-mail.stanford.edu/pipermail/planetccrma/2004-December/007341.html > > http://ccrma-mail.stanford.edu/pipermail/planetccrma/2004-December/007343.html > > >From the above post: > > Currently I think that the most usable system is > 2.6.10 (that is if you are willing to run all of your audio software as root, > which is not a good idea as a bad bug could bring your system to its knees) > > This user like many others is confusing the realtime preempt patches > with the realtime LSM. You should point this out (I am not on the CCRMA > list). The current version of the realtime LSM kernel patch applies > cleanly against 2.6.10. I guess he was referring to the kernel he built (without the realtime-lsm). I'll clarify this point in the Planet CCRMA list... I just booted into 2.6.10 + realtime-lsm, not bad at all in some very informal jack tests (run jack at 2x128 + ardour + hydrogen + freqtweak). I'll leave a full build chugging along all night long, should have fresh packages to test/release tomorrow :-) -- Fernando