> >The cdrecord README clearly states that you _need_ working DMA > >in order to write DVDs. If you have burnfree ON the drive just > >dramatically slows down. > > > > > If the DVD unit were set to 4x, but data underrun were reducing the > effective transfer speed to 2x,
Well, this I can agree with 100%, but I can't see that this is equivalent to the referred statement. There is a huge difference between "having burnfree ON" and "suffering from buffer underruns," as well as between "*just* dramatically slows down" and "cause *effective* speed reduction," don't you think? Of course whenever recording suffers from buffer underrun, an idle spin has to be taken, which effectively slows down the recording, but the question is for how much. This was discussed already, given that DVD+ performs as expected at 2.4x (as on requestor's system), 3x seems to be commonly observed value for 4x media in the lack of DMA, not 2x as user reported. Also note that given that buffer capacity is enough to sustain several rotations, the unit is capable to follow the wire transfer speed without abrupt performance degradation (because proportion of idle spins is minimized). I mean you don't normally end up in situation when performance degrades by factor of N, whenever maximum transfer rate between host and logical unit is only slightly lower than required. Instead, you observe effective recording speed being slightly/somewhat less than the actual wire speed. Well, this naturally does not exclude the possibility that there might be firmware implementations out there, which would exhibit abrupt performance degradation, in which case one'd say "if recording suffers from buffer underruns, you *can* observe dramatic performance degradation," but you can't say "it *just* does it." > would cdrecord report the speed set or > the speed achieved? It prints both as far as I understand. Requestor should be able to clarify which one, intended velocity in the beginning or average summary at the end, he was referring to. Cheers. A.
