2 MB or 8 MB cache on external drive?

[besmusic]

Is there a realized benefit in PTLE using 002r for the 8MB cache? Any recommendations?

[where02190]

The apps workings, wave draws, file caching, etc are a bit faster with the 8mb cache drives. there's little difference in price these days, and it can't hurt. The WD Caviars seem the fastest, the IBM/Hitachi's almost as fast and much quieter.

[hurdy gurdy]

I second the notion:
Those WD drives are pretty hard to beat for the price. Quiet, fast, reliable, and just dang cheap and rugged.
L.A.Branville

[3JDamon]

It's usually best to disable the drive cache for Protools, but the 8mb cache is better for other operations.

[besmusic]

3JDamon,
Can you explain why? Obviously conflicting advice above. I know digi says there's no benefit...how do you disable the cache just when working in PTLE? Have you noticed a specific benefit in doing this?

thx so much,
Pete

[lemix]

Hmmm...

It's usually best to disable the drive cache for Protools, but the 8mb cache is better for other operations

Aren't you referring to the Windows implemented allocation of read/write cache here ??
How can one disable that 8Mb of RAM, which is actually residing on the HD's controller board ??
And why ??
Allen...please...

cheers,

[tele_player]

I had the same question, but then I did some searching.
It's hard to find documentation, but I read something on the MS website recently which suggested the Windows disk drivers know how to enable/disable the buffer on the EIDE drives.

[lemix]

OK,

But why though ?? Has anyone here actually compared disc performance between two identical models ( say...WD 80Gb ) with 2 and 8 Mb on board ??
I'm really interested in knowing the results.

cheers,

[tele_player]

3JDamon claims (I think) to have done enough testing to be convinced. I haven't, I leave it enabled, no problem on my sessions so far.

Intuitively, I think it should be helpful. Practically, it's really hard to devise a test to actually quantify the performance effect, since we don't know exactly how PTLE accesses the disk in normal use, and the DAE buffer setting is also involved.

But the 8mb buffer on the drive should decouple (somewhat) OS file writes from the actual commitment to disk, which should make file writing very fast, up to a point. As the write rate approaches the physical disk write rate, any benefit should diminish.

Enough theory and speculation. Install a drive, try it both ways, use whatever works.

[lemix]

Allright...let's do music...have a Guinness...or a few...