Suggestions for hard drive, SSD and RAID performance analyst software? I am running in circles trying to determine if my investment in storage devices has paid off! Maybe I am going about it wrong, but it seems I need a starting reference in terms of performance. My W7 score for hard drives is a 5.9, while all other areas score in the mid 7.x.
I am running a Areca 1880i with one Cheetah 15k 16 Mb cache in 'pass-through' for C Boot, RAID 0+1 (UserData) with 4x WD Black SATA3 6Gb/s 7200 64Mb cache 1Tb, and a RAID0 (realtime capture) with 3x Fujitsu SAS 15k 16Mb cache. I do pro adio/video projects in a small commercial setting. I can do either a dual boot or use 2 different computers for studio and home usage.
I need to sort this out, as I am not convinced I am meeting expectations or typical results.
More about :software determine storage performance
Unfortunately, the definition of "performance" depends on what you are trying to do. If you are capturing video, raw sequential writes are important. If you are running a database or editing video, raw sequential access is next to useless and random IOs per second are important.
I personally never bother to go farther than using the AS SSD benchmark app (it works for HDDs, too). If you look at Toms' articles on drives, they use much more intensive and extensive benchmarking tools. Look at a recent article on drives. There are quite a few benchmarks of different types, with different drives leading. The tool used is h2benchw.
The trick is to get a frame of reference. Find some relevant disk reviews, bench your drives with the same tools, and compare.
The most accurate results will be with application-specific tests, as Tom's tried here: http://www.tomshardware.com/reviews/solid-state-drive-w... . The only thing that you care about is how long it takes you to load, edit, and save your projects; 2 GB file copy speeds are irrelevant for you.
Finally, to continue being a wet blanket, most of the threads here that mention Windows experience scores say to ignore what it has to say about disks. And the biggest bang-for-the-buck in video editing is, expensively, an SSD large enough to hold your project and the application's scratch directory.
Were I in your place, I would buy one of the fastest current hard drives, replace the drive subsystem that you doubt with it temporarily, and play. If you notice no change in speed, your arrays aren't doing anything for you. If you say "holy heck, how slow is that!", then your arrays are making your life better.
Let us know how you proceed and what you find. It's not often that we see a real-world performance inquiry, and I for one could learn from what you find.
I just ordered the QCZ synapse 64 for use with my boot Cheetah 15K5 SAS HDD. Should this purchase pay as expected?
I also think that my HDD block size in my RAID0 Capture subsystem should be configured to 1024 or so and not 512Kb. Agree?
I was also told that 3x HDD builds for RAID0 is not enough HDD to fully appreciate RAID0 in I/O intensive applications such as video and audio editing. Should I go to 4 or maybe 5 HDD (currently 3x) in my RAID0 build? Suggestions?
Maybe I need to use my Areca 1880i (2 port/4 SAS each) to build ONE larger RAID0 instead of two RAIDs, (i.e., 8 HDD RAID0). Agree? Of course I could go RAID0+1 with slight reduction in performance but much safer. Opinion?
I currently have 4x Fujitsu SAS 15k 16Mb cache 74Gb HDDs, 4x WD Black 7200 SATAIII 6Gb/s 64Mb cache 1Tb HDDs, and 1x Cheetah 15K5 SAS 15k 16 Mb cache 174Gb HDD to play around with. I am sure I can find addition similar HDDs as needed if advised. Would you invest in additional WD Black 6Gb/s or the older, smaller Fujitsu or even Cheetah 15K5 HDDs? Many say the newer WD 6Gb/s 7200 HDDs are great in RAID configs, and 15k SAS are not required any more. Opinion?
The synapse will probably speed up booting and application launching. Unless you work on one project that will fit into that drive entirely, and work on it over and over again, it won't cache your data, and your execution speed will not be improved. If the editing tool that you use has a large scratch directory, a better use of the device would be to put the scratch directory there.
As to comparing different RAID configurations, I do not have the experience. I have the overall impression that RAID0 is a mistake except as scratch space, because it is vulnerable to loss, and that unless you RAID0 SSDs you won't get much of a practical advantage, because you are still limited by drive access times. But I simply do not have the experience to answer the rest.
Because usage differs user to user it would probably benefit to use multiple benchmark tools to test a drive. Although sparingly with an SSD due to data writes.
Other ones mentioned I've seen are AS SSD, and HD Tune.
In regards to your RAID quest. Not sure what a capture subsystem is, but if you ment block size I would say 128kb is good. If you could swing the array up to 4 HD go for it, I don't think you have to worry about a RAID0+1 as long as you have another solid back up plan.
My personal suggestion is to save the money on the extra drives and put them towards a/more SSD drives. =)
I did build a RAID0 (4x SAS 15k 16Mb cache) for real-time recording/editing of audio and video, and a RAID0+1 for moving projects to after the daily sessions (safety). The 4th HDD I added to the RAID0 system made a huge difference. NO doubt, the number of HDD in RAID0 make a huge improvement, but the odds of failure increase though! That's why I move sessions to my RAID10 system daily. In audio and video projects I/O operations and sequential writes are key. As far as SAS 15K being noisy, yes they are, but can be dampened and/or located in computer closet.
I needed up dual booting W7 64: normal boot and a audio/video tuned boot. Since, my user data is on a separate HDD (RAID10), both boots can access the files such as downloads. This means that I can really tune the AV boot to a mere minimum such as no internet, security, etc.!