hard drives

[Martin]

works even better if you put the system on a different disc.
It does make a difference with accuracy of playback over long segments on 4 layers - I know martin had a few issues with this in norway )

We tested this and found out that we did not get any more performance on a raid than on 1 std disk and 1 sata. os on one and cat on the other.
besides, when one disk fucks up, you can keep a backup on the other, so you can (hopefully) still do the show. As for the playback sync problem, the use layer function saved the day.

[samsc]

We tested this and found out that we did not get any more performance on a raid than on 1 std disk and 1 sata. os on one and cat on the other.
besides, when one disk fucks up, you can keep a backup on the other, so you can (hopefully) still do the show. As for the playback sync problem, the use layer function saved the day.

keeping the os separate gives slightly better performance over the long term - minutes -

i dont think using an internal raid is good either. if one disc goes down - you loose everything.

there is no difference between using 1 or 2 raptors - unless you need the extra gb capacity - or unless you want to play back uncompressed movies.

[Jan Opseth]

We tested this and found out that we did not get any more performance on a raid than on 1 std disk and 1 sata. os on one and cat on the other.
besides, when one disk fucks up, you can keep a backup on the other, so you can (hopefully) still do the show. As for the playback sync problem, the use layer function saved the day.

It´s very strange you did´t see any better perf. with Raid0.. I tested this.. with and without... i did see better results with RAID

you should´t be worried the disk´s should crash.. does the OS disk fail your f...ed anyway.. If you shall be totaly backed up You should do Raid 3 or 5 external.. Better yet is SCSI 15k Cheetah´s, I have 2pc of this ones in my Windows computer in Raid 0 and I´m seeing a lot of a difference..
(vs. s-ata), so it´s a difference with RAID.

Disks rearly f... (as long as you don´t have IBM´s) What if your disk in your GrandMA or WH3 fails? THEN you are f..

Jan O..

[samsc]

.. i did see better results with RAID
..

Which Application?

What kindof tests did you do?

[Jan Opseth]

Which Application?

What kindof tests did you do?

I did use Xbench 1.1.3..

[samsc]

I did use Xbench 1.1.3..

catalyst is not a datarate limited application - its a seek rate, and rotational latency limited application and io/ops/second limited

The headline datarate is not relevant.

Whats important is how fast the drive works- up to a point, and then no more.

In my testing, 2 drives was not better than one - playing back 4 movies on 4 layers at 25fps. A single drive was better.


Catalyst is not a normal application wrt disc use. Its quite different.

[Martin]

We tested it with the Apple people, and as you can see from the Catalyst benchmark tests, it´s not a special difference. We also tested raptors V Scsi raid at the workshop yesterday, and only on uncompressed and on 4 layers is there a noticable difference. On standard DVpal, the difference is not visible.

[tylerr]

I assume that when you are adding a SCSI drive you are pulling the power off of the optical drive for the ATI Radeon 9800 pro and for the SCSI drive.

Tyler

[samsc]

I assume that when you are adding a SCSI drive you are pulling the power off of the optical drive for the ATI Radeon 9800 pro and for the SCSI drive.

Tyler

yes.

---

The g5 jam came with a really poor set of power jumpers that hadnt been crimped very well.
---

in my experience - i have older macs with 4 or 5 internal drives- running power jumpers was not a problem - the psu used always seemed capable.
I often have a full set of pci cards as well.

This is not to say this will always be the case.
I tend to leave things on 24/7 the whole time - often with stuff running to soak test hardware.

[Spam Butterfly]

Tyler,

You might need to find yourself a power cable Y-splitter. I misappropriated one from some dead PC kit.

Hugh