How many OWC SSDs can you put in RAID0?

[nanofrog]

that second link was *Benchmark* ARC 1221x (SATA) vs ARC 1222x (SAS)
so the 1222x was on their also just widen your browser maybe you will see it
so the reason for the two links one shows with SATA drives and the second showed it with SAS drives the way the site is you have to make sure your browser is wider though to see it

Correct. The second link used different drives with each card (though similarities exist as both = 500GB @ 7200 RPM). The first link used WD 1TB RE3's.

Makes comparing the results more convoluted, as the test platforms aren't similar enough (i.e. change the card out, everything else remains the same in order to limit the variable to one item = Device Under Test).

 

[Honumaui]

Correct. The second link used different drives with each card (though similarities exist as both = 500GB @ 7200 RPM). The first link used WD 1TB RE3's.

Makes comparing the results more convoluted, as the test platforms aren't similar enough (i.e. change the card out, everything else remains the same in order to limit the variable to one item = Device Under Test).

but go back to what he said ! from his quick search scary way to try to find facts ? but oh well so this is what he is claiming

The card seemed to max out at around 600 megabytes/sec read and 350 megabytes/sec write in any raid level (including raid6).

I am trying to say look its not the card even in two dif situations two dif drives SAS and SATA the card performs as it was meant to ! not comparing the results against each other just saying its not limited to a 350 write like he said !
to me its purely about the card not anything else and can it perform better than 350 write ? YES obviously his hero poster is wrong !

most likely the guy he quoted had the wrong HDD in or had not updated the firmware and even you are saying the card is quicker than this ? but he wants to believe one guy on a forum who did not SHOW what discs he had etc..

I am saying here is a benchmark with SAS ! here is one with SATA and the card is not the limiting factor !
does that make more sense now ?

 

[nanofrog]

but go back to what he said ! from his quick search scary way to try to find facts ? but oh well so this is what he is claiming


I am trying to say look its not the card even in two dif situations two dif drives SAS and SATA the card performs as it was meant to !

the platform is the card itself

I don't disagree. I see it as a function of the drives used (they're older models that aren't as fast).

Back then, to get over the 600MB/s range, you'd have had to use more than 8 disks, so the combination of the ports fixed to 8 and the available drives is what generated the throughputs (not as fast as expected).

Now there's faster drives (mechanical and definitely SSD's). And if SSD's were used in a parity based array, they'd slow down as well. I'd need to test that out, with both MLC and SLC drives, which I don't have.

Care to make any drive donations?

But to give you an idea, I can exceed that with WD RE3's on a ARC-1231ML (running 8x members in RAID 5, stripe = 128k), and taking cache out of the equation (disable; cache active, and I can hit just over 1GB/s, as the cache is 2GB), am still getting ~700MB/s. They're newer drives, and are faster due to the higher platter density.

 

[VirtualRain]

but go back to what he said ! from his quick search scary way to try to find facts ? but oh well so this is what he is claiming


I am trying to say look its not the card even in two dif situations two dif drives SAS and SATA the card performs as it was meant to ! not comparing the results against each other just saying its not limited to a 350 write like he said !
to me its purely about the card not anything else and can it perform better than 350 write ? YES obviously his hero poster is wrong !

most likely the guy he quoted had the wrong HDD in or had not updated the firmware and even you are saying the card is quicker than this ? but he wants to believe one guy on a forum who did not SHOW what discs he had etc..

I am saying here is a benchmark with SAS ! here is one with SATA and the card is not the limiting factor !
does that make more sense now ?

I would suggest reading the Storage Forums at Xtremesystems.org if you want insights into the current performance limitations of any SSD and/or RAID card combo. Those guys are not making **** up, they are benchmarking gurus and pushing the limits to the, well, extreme!

That's the source of my insights on the 1222... Search and you will find a couple of posts saying the same. Like most RAID cards, It appears the 1222 is not designed for SSDs.

 

[mangrove]

Does anyone know how many OWC SSDs you can put in a RAID0 before you reach bandwidth saturation?

What would be the maximum number of drives if you use a hardware RAID controller on the 16x PCIe slot? I assume you could use more drives than connecting on the Mac Pro's SATA backplane but I'm not sure what the exact number would be in practice.

Read Lloyd Chamber's article on just that in his Mac Performance Guide where he tests 2 and 3 SSD's in RAID 0. Interesting article. Take a look.http://macperformanceguide.com/Reviews-MacProWestmere-SATA.html