Please recommend an external SSD drive for me

[musicman691]

Originally Posted by Emi

Darryl, the doubt I have about disc cache: if o allocate only 1gb and the session has more, let’s say 3-4 Gb, does the disc cache buffers audio along the way or it only caches 1 gb and the rest doesn’t? If it caches as the session is being played can this produce errors?

Thanks

Not so much errors like 9173/whatever but pops and crackles IF you're using a spinner for your sessions. AND at the same time running with low buffer settings. I ran into just such a situation a couple of months back. But with an internal ssd for sessions there shouldn't be a problem.

[Darryl Ramm]

Originally Posted by Emi

Darryl, the doubt I have about disc cache: if o allocate only 1gb and the session has more, let’s say 3-4 Gb, does the disc cache buffers audio along the way or it only caches 1 gb and the rest doesn’t? If it caches as the session is being played can this produce errors?

Thanks

My understanding is it tries to do something sane but you cannot get blood out of a stone. Memory is cheap if you can add it... fast PCIe/NVMe disk always helps and both together should also interact well, e.g. IIRC it looked like Pro Tools uses fairly efficient asynchronous IO to write modified content out of the disk cache, the sort of stuff that plays well on fast SSD.

Understanding what is going on with overall memory is a little complex on modern demand paged operating systems like OS X/macOS and Windows, and that is made even more complex with black magic like memory compression. It may be possible to increase disk cache beyond initial impressions and have it work OK, or maybe not... may be worth testing, but obviously not great to be running on the edge... and if using lots of VI samples and you have a fast SSD then maybe streaming samples is an option vs caching. Oh and make sure memory pigs like Google Chrome are just not running.

[Emi]

Originally Posted by musicman691

Not so much errors like 9173/whatever but pops and crackles IF you're using a spinner for your sessions. AND at the same time running with low buffer settings. I ran into just such a situation a couple of months back. But with an internal ssd for sessions there shouldn't be a problem.

Yeah, I remember when you had that problem. Told you it was the spinner.

I actually don’t have any problem with errors with ssd and the disk cache is at normal.

[Emi]

Originally Posted by Darryl Ramm

My understanding is it tries to do something sane but you cannot get blood out of a stone. Memory is cheap if you can add it... fast PCIe/NVMe disk always helps and both together should also interact well, e.g. IIRC it looked like Pro Tools uses fairly efficient asynchronous IO to write modified content out of the disk cache, the sort of stuff that plays well on fast SSD.

Understanding what is going on with overall memory is a little complex on modern demand paged operating systems like OS X/macOS and Windows, and that is made even more complex with black magic like memory compression. It may be possible to increase disk cache beyond initial impressions and have it work OK, or maybe not... may be worth testing, but obviously not great to be running on the edge... and if using lots of VI samples and you have a fast SSD then maybe streaming samples is an option vs caching. Oh and make sure memory pigs like Google Chrome are just not running.

Thanks Darryl, I had to double read, and still have doubts about it. I understand is hard to explain what it is actually doing.

I’m running disk cache at normal (so, no buffering) without any issues with ssd’s

[musicman691]

Originally Posted by Emi

Yeah, I remember when you had that problem. Told you it was the spinner.

I actually don’t have any problem with errors with ssd and the disk cache is at normal.

The spinner is NOT the problem. Been through that already.

[Emi]

Originally Posted by musicman691

The spinner is NOT the problem. Been through that already.

What was the problem? Have you discover it?

[musicman691]

Originally Posted by Emi

What was the problem? Have you discover it?

Like I said in the original thread it's a combo of low buffers and low cache setting.

[paul_g]

Originally Posted by Emi

What was the problem? Have you discover it?

My problem existed between the keyboard and chair!

[Emi]

Originally Posted by paul_g

My problem existed between the keyboard and chair!

Lol