Sounddesigner wrote: ↑Tue Jan 21, 2020 10:25 pm
Definitely for Asio and Wave drivers use dsp 2.
O.k.,- that´s what I knew already from here ...
- XITE-1 DSP Structure.gif (20.07 KiB) Viewed 5195 times
Sounddesigner wrote: ↑Tue Jan 21, 2020 10:25 pm
I don't remember if it's necessary for ADAT, etc to be placed on dsp 2 or not but Valis makes a great point that it's still best to have everything on the same dsp.
I understood, soft- and hardware I/Os can go to DSP#1 and
soft-I/Os again and VDAT can go to DSP#2.
Now, it matters what ADAT and Zlink I/Os are .. soft- or hardware ?
Sounddesigner wrote: ↑Tue Jan 21, 2020 10:25 pm
It's been a longtime since SonicCore spoke on these things and since we've talked much about them here so my memory has faded a little regarding ADAT etc.
Yeah,- and it might have changed ... maybe for SCOPE7 (and OS beyond Win7) ... no ?
Sounddesigner wrote: ↑Tue Jan 21, 2020 10:25 pm
I just remember... that the older ones were for drivers and for very old SCOPE plugins that might be incompatible with newer dsp's (like VDAT) and that dsp 2 was for drivers.
I don´t understand why all this cannot be handled by the SCOPE system itself in the way once it recognizes a "driver" dependend module being loaded in the environment, placing it automatically on the required (requested ?) DSP.
Sounddesigner wrote: ↑Tue Jan 21, 2020 10:25 pm
Newer dsp's are'nt as efficient with drivers and may suffer some compatibility issues with some older plugins.
I dunno if that´s the real truth.
It´s possibly because the ancient PCI card based SCOPE system worked w/ these old SHARC 60Mhz DSP chips that way.
Newer SHARC DSPs do I/O routing/handling in Apollo, Solaris or whatelse hardware FX devices well.
I know XITE is more complex, but when old SHARC DSPs were replaced by new and much faster ones, the I/O assignment and routing would still work,- but backwards compatibiliy w/ some old devices would have been gone.
This might not be a big issue if we got more new devices replacing the old ones 100% in sound and functionality,- possibly improve sound and features in some cases,- p.ex. B2003.
I also understood "old" 60MHz DSPs #3-6 are for communication between DSP-"slots" in XITE-1D as well as the "chip-arrays" (3 DSPs in each slot) in XITE-1.
Means, these work as a virtual S/TDM cable and there might come up a different data traffic depending on it´s a XITE-1D or XITE-1.
In addition, they do the communicaton between devices running on these DSPs and when these require host computer´s RAM and/or CPU cycles.
SAT connection errors come from these DSPs alway w/ the message
"no more SAT connections possible between DSP # X and # Y" and such.
I only know early PCIe 1.0/1.1 standard is already 250MB/sec. @ 70MHz and old SHARC runs @60MHz,- which in theory reduces bandwidth.
So,- for XITE-1, w/ 3 fast (333MHz) SHARC DSPs in each of it´s 4 DSP slots, such 60MHz SHARC might be a bottleneck once devices use host computer´s RAM and CPU inside a large project´s combo of loaded devices.
Just only my guess though ...
Bud