Luminous Landscape Forum

Equipment & Techniques => Computers & Peripherals => Topic started by: PeterAit on November 24, 2009, 08:00:16 pm

Title: Benchmarking RAID speed
Post by: PeterAit on November 24, 2009, 08:00:16 pm
I have an external eSATA RAID 0 unit that I use for storing my images and LR catalogs. Some file operations are not as fast as I think they should be - for example, a 180 MB file takes 15 seconds to save from Photoshop. Maybe that's an OK time, but I'd like to be sure that the RAID is performing as it should. Can anyone suggest some way to benchmark the system to verify? I am running Vista 64.  Thanks.
Title: Benchmarking RAID speed
Post by: Farmer on November 24, 2009, 08:15:05 pm
Quote from: PeterAit
I have an external eSATA RAID 0 unit that I use for storing my images and LR catalogs. Some file operations are not as fast as I think they should be - for example, a 180 MB file takes 15 seconds to save from Photoshop. Maybe that's an OK time, but I'd like to be sure that the RAID is performing as it should. Can anyone suggest some way to benchmark the system to verify? I am running Vista 64.  Thanks.

Copy the same file from another drive and see how long it takes.

Photoshop write speeds of PSD are not particularly snappy - it has to do with compression being used.

Also bear in mind that an eSATA raid only utilises a single SATA channel to cover both drives as opposed to a mainboard or RAID controller setup that uses two channels (one to each drive).  This will have some impact on overall performance as well.

The last time I used Benchmarking software was PCMark which was quite decent.
Title: Benchmarking RAID speed
Post by: rovanpera on November 24, 2009, 09:01:22 pm
And the less you have data in the drive the faster it is...

Title: Benchmarking RAID speed
Post by: PeterAit on November 24, 2009, 09:05:12 pm
Quote from: Farmer
Copy the same file from another drive and see how long it takes.

Photoshop write speeds of PSD are not particularly snappy - it has to do with compression being used.

Also bear in mind that an eSATA raid only utilises a single SATA channel to cover both drives as opposed to a mainboard or RAID controller setup that uses two channels (one to each drive).  This will have some impact on overall performance as well.

The last time I used Benchmarking software was PCMark which was quite decent.

Thanks for the suggestions. I am using a 2-channel eSATA so it should be quite fast. Also, wouldn't copying from another drive most likely be limited by that drive?

I'll take a look at PCMark.
Title: Benchmarking RAID speed
Post by: fike on November 25, 2009, 10:25:35 am
Are you using RAID 0, RAID 1, or RAID 0+1?

oops....reread your entry.  RAID 0, I see.  

I would have to say that I agree with the earlier poster who suggests copying the same file to another internal drive.  Also, as another poster mentioned, the compression that PSD files go through will slow performance.  So, I would suggest copying files for your testing instead of saving from PS.  

I am not a big fan of those benchmarking tests.  They tell you  nothing about the comparative speed of your hardware in real-world apps.  

Is the eSATA device operating in SATA or SATA2 mode?  Do you know?
Title: Benchmarking RAID speed
Post by: alain on November 25, 2009, 02:11:25 pm
Hi

I found a small program that can test disk write speed for specific size.  If there's some write caching, it will take it in account.  Continous tests don't do that.

Small HDD test software (http://www.raymond.cc/blog/archives/2008/02/28/measure-actual-hard-disk-perfomance-under-windows/)
Title: Benchmarking RAID speed
Post by: PeterAit on November 25, 2009, 05:05:35 pm
Quote from: fike
Are you using RAID 0, RAID 1, or RAID 0+1?

oops....reread your entry.  RAID 0, I see.  

I would have to say that I agree with the earlier poster who suggests copying the same file to another internal drive.  Also, as another poster mentioned, the compression that PSD files go through will slow performance.  So, I would suggest copying files for your testing instead of saving from PS.  

I am not a big fan of those benchmarking tests.  They tell you  nothing about the comparative speed of your hardware in real-world apps.  

Is the eSATA device operating in SATA or SATA2 mode?  Do you know?

The purpose of a benchmark in this case is to run it and compare my results with the results obtained by the vendor I bought the RAID from (MacGurus, excellent support). Assuming they know how to set it up for optimum performance, then if our results match I know I am getting the best performance from the unit. That's all I care about, not comparing the speed of my hardware to other setups.

As for SATA or SATA II, I don't know. I would guess SATA II because when I made the purchase I asked for the fastest interface card they had.

Title: Benchmarking RAID speed
Post by: Gambit on November 25, 2009, 05:47:12 pm
if your using a mac you can use disktester from mac performance guide .com
Title: Benchmarking RAID speed
Post by: John.Murray on November 25, 2009, 06:19:49 pm
iometer

http://www.iometer.org/ (http://www.iometer.org/)
Title: Benchmarking RAID speed
Post by: fike on November 25, 2009, 06:58:59 pm
The Resource Monitor can be useful for looking at system performance during demanding tasks.  It can be found under the Task Manager on the Performance Tab.
Title: Benchmarking RAID speed
Post by: DarkPenguin on November 25, 2009, 07:48:04 pm
Quote from: Joh.Murray
iometer

http://www.iometer.org/ (http://www.iometer.org/)

This is what we used at seagate/xio....
Title: Benchmarking RAID speed
Post by: mmurph on November 26, 2009, 11:55:45 am
http://smallnetbuilder.com/ (http://smallnetbuilder.com/)

These folks have many tests using iozone and brute force copy, eSata, gigabit network, etc.

If they have your product you can compare their results. Otherwise you can get a general feel for what is realistic, plus tips and I ideas on tuning (jumbo frames, etc.)

I can't navigate it - or read the charts - on my iPhone, so I'll just post the link right now.

"Happy happy" to those in the US today, to quote Phish.