On Friday 09 November 2001 09:35, Maarten de Boer wrote: > Roger Larsson <[EMAIL PROTECTED]> wrote: > > They are not bad either... (sub 10 ms) > > I think it is bad. I want the <3 ms :-) > I will be added latency in the processing (fft, ifft) > and it all adds up. > > > > Something is definitely very wrong here. I am running kernel > > > 2.4.13 patched with preempt-kernel-rml-2.4.13-2.patch > > > > And you have enabled it in .config? (checking the obvious...) > > Yeah, I triple checked the obvious (also: did I run lilo, did > lilo use the right vmlinuz) > > > But wait... an AMD.... > > They have some 3DNow specific kernel optimizations that causes trouble - > > they need to disable preemtion (uses fp registers...) > > Hmm.. Does anybody have more information about this? Could I disable these > optimizations compiling for another CPU (386, or maybe even Pentium > Classic?) I suppose the performance penalty that would give would be > acceptable. Latency is more important. >
Check if you have this in .config CONFIG_X86_USE_3DNOW=y > Would Andrew Morton's LL patches also be affected by this? > Yes, if this is the problem. /RogerL -- Roger Larsson Skellefte� Sweden
