Very good point, Steve. Thanks for continuing my argument. In our host
we do limit to 4 sample segments for parameters, so we have not worried
about writing the segment in the SIMD register. But this would work
great, and definitely be faster that interleaved operation on channel
mappings that do not have a multiple of 4 channels.
Steve Harris wrote:
On Wed, Jun 04, 2003 at 01:09:21PM -0400, RonKuper@xxxxxxxxxxxx wrote:
It probably seems like hair-splitting, but I'm pretty sure now that we've established with our customers that we render gain at a control rate of 1 sample, they might not accept any less going forward. So for us anyway, and for any other host or plugin that renders sample by sample, the loss of interleave would pose a potential problem.
No, because the vector x vector operation is plenty fast enough - memory bandwidth will kill you long before that does.
You can load the gain segment, run 4 samples of each channel, repeat. Assuming your gains are the same for all channels. If thier not it makes no difference wether the data is interleaved or not.
This will probably even give you better cache uasage than the interleaved case as the prefetch for channel 1 can be happening while you process channels 2-6.
- Steve
---------------------------------------------------------------------- Generalized Music Plugin Interface (GMPI) public discussion list Participation in this list is contingent upon your abiding by the following rules: Please stay on topic. You are responsible for your own words. Please respect your fellow subscribers. Please do not redistribute anyone else's words without their permission.
Archive: //www.freelists.org/archives/gmpi Email gmpi-request@xxxxxxxxxxxxx w/ subject "unsubscribe" to unsubscribe
-- Mike Berry Adobe Systems
---------------------------------------------------------------------- Generalized Music Plugin Interface (GMPI) public discussion list Participation in this list is contingent upon your abiding by the following rules: Please stay on topic. You are responsible for your own words. Please respect your fellow subscribers. Please do not redistribute anyone else's words without their permission.
Archive: //www.freelists.org/archives/gmpi Email gmpi-request@xxxxxxxxxxxxx w/ subject "unsubscribe" to unsubscribe