Does PT NT really work only on Digidesign approved PCs?

[JeroenL]

I have a blazing Athlon 600 MHz NT PC with 27GB Maxtor EIDE drive. Is it possible to run PT24 on such a system. I know it is not on Digidesign's compatibility list, but is there any good chance that it might still work?

Thanks,
Jeroen

[Digi Engineering]

Jeroen,

You'll run into problems trying to use the EIDE drive for 24-bit audio. This is due to the fact that bus-master IDE controllers can usually only do DMA to even memory addresses, and editing 24-bit audio will result in regions being placed on odd memory addresses. We don't officially support the use of IDE drives with TDM systems, but it *may* work with 16-bit audio.

There is also a chance the rest of your system will work with PT24, but there's no way to know without trying it. However, for guaranteed performance and support you're better off using a supported machine.

Eric Day
DAE Engineer

[JeroenL]

I have read on the Ensoniq PARIS newsgroup that users have had results using Athlons and fast 7200RPM bus-master EIDE drives, outperforming even Intel/SCSI systems. In that sense it appears that (at least for PARIS) the benefit of SCSI over EIDE is gone.

PARIS is also a 24bit system and still runs perfectly using EIDE. So in what sense is PT using the file system differently?

Will PT in 24bit mode not work because it addresses the bus-master IDE controller directly, such that the addressing problem occurs? I would assume that WindowsNT takes care of that (maybe aligning every 24-bit sample to a 32bit even address) and prevents an application from doing so itself, but I may be wrong here.

So, to make things a bit more clear for me, will 24-bit PT audio not work at all with EIDE or do you simply expect a performance degradation compared to SCSI when doing so?

Sorry, to keep ranting about this, but I am simply trying to understand whether there is any point in trying if EIDE might work for me, or not.

Thanks again,
Jeroen

[Digi Engineering]

>>Will PT in 24bit mode not work because it addresses the bus-master
>>IDE controller directly, such that the addressing problem occurs?

Yes, this is essentially the case. PT uses asynchronous unbuffered IO, which places strict requirements on the buffer address (among other things). SCSI controllers can do DMA to any arbitrary address, but IDE controllers commonly require WORD-aligned (even) addresses. An alternative to unbuffered IO is buffered IO, in which the OS uses an internal buffer for the IO, and then copies the data into the application's buffer afterward. This allows application buffers of any alignment, but imposes a performance penalty which was not seen as acceptable for PT.

>>So, to make things a bit more clear for me, will 24-bit PT audio not
>>work at all with EIDE or do you simply expect a performance degradation
>>compared to SCSI when doing so?

24-bit PT audio will work fine with EIDE as long as playback begins on, and all audio regions are aligned to even-sample boundaries. This means you have to take special care when editing audio as to not cause something to end up on an odd-sample boundary. The consequence of odd-sample playback or audio regions is an error dialog, not simply a performance degradation. The OS function called to do unbuffered IO will fail if the buffer does not conform to the disk controller's requirements, and playback in PT will come to a halt.

Eric Day
DAE Engineer