New LSI 9200 series controllers: 6Gb/s, 2.88 GB/s seq. reads, 1.87 GB/s seq. writes

[SteveRo]

XS Janus might be right.
More specs on the INTEL RS2BL080 located here - http://www.intel.com/Assets/PDF/prodbrief/322235.pdf
512MB cache running at 800Mhz with LSI onboard proc running at 800 Mhz supporting raid 0, 1, 5 and 6.
This doesn't sound like the same (9210/RS2BL080) controller Tom's tested.
I'm beginning to wonder if this is the 9260 under the INTEL label?

[SteveRo]

Compare the pictures from here - http://www.newegg.com/Product/ShowIm...e%20-%20Retail

and here - http://www.intel.com/Assets/PDF/prodbrief/322235.pdf

Looks like the same to me.

[tiro_uspsss]

sorry, seem to have missed it - what XOR do these use? is it the dual core Intel like in the Areca 16XX series?

[lutjens]

XS Janus might be right.
More specs on the INTEL RS2BL080 located here - http://www.intel.com/Assets/PDF/prodbrief/322235.pdf
512MB cache running at 800Mhz with LSI onboard proc running at 800 Mhz supporting raid 0, 1, 5 and 6.
This doesn't sound like the same (9210/RS2BL080) controller Tom's tested.
I'm beginning to wonder if this is the 9260 under the INTEL label?

I saw this myself and scratched my head for sometime. I think Tom's is in error...as the card on Intel's site as the RS2BL080 is clearly the 9260. The remaining question in my mind is which card achieved the higher benchmark...the 9210, which Tom claimed as the RS2BL080 (in error) or the RS2BL080, which visually and technologically is the 9260?

[SteveRo]

sorry, seem to have missed it - what XOR do these use? is it the dual core Intel like in the Areca 16XX series?

I think the proc is a single core, from INTEL the proc is an LSI SAS2108 - see - http://www.intel.com/Assets/PDF/prodbrief/322235.pdf

[SteveRo]

I just ordered a 9260-8i and a 4th acard - hoping for compatability!
If all comes together well - expecting 1400/1200 seq. read/write and some very high random numbers.
What 9260 firmware and controller settings seems to be working best for SSDs?

[XS Janus]

Looks like the latest one kicks some good ass.
Computrd, posted his results. The only "bad" thing I see is worse access time than Areca. Maybe the one Tom tested, but wrongfully reported as 9210, is better at this than a real 9260...
hopefully we learn soon!

[F.E.A.R.]

i would like to see a setup with Acard 9010 (2 or more) + 9210
is it true that 9210 is faster than 9260? what´s the reason???

[SteveRo]

i would like to see a setup with Acard 9010 (2 or more) + 9210
is it true that 9210 is faster than 9260? what´s the reason???

Based on what I have seen so far, it might be that the 9010 never went into production or perhaps has not gone into production YET. Tom's is the only mention of it that I have seen - and some of the data Tom's provided may not have been completely accurate re the 9010.

[tiro_uspsss]

I think the proc is a single core, from INTEL the proc is an LSI SAS2108 - see - http://www.intel.com/Assets/PDF/prodbrief/322235.pdf

mm.. wouldnt that mean darn long R6 rebuild times?

[lutjens]

I just ordered a 9260-8i and a 4th acard - hoping for compatability!
If all comes together well - expecting 1400/1200 seq. read/write and some very high random numbers.
What 9260 firmware and controller settings seems to be working best for SSDs?

I also have one on the way to hopefully uncork my SSDs...I'd have bought the Intel version of the card if it was available. I'm hoping for some interesting results...

[Biker]

I just ordered a 9260-8i and a 4th acard - hoping for compatability!
If all comes together well - expecting 1400/1200 seq. read/write and some very high random numbers.
What 9260 firmware and controller settings seems to be working best for SSDs?

These settings are best for lowest latency:



I use a 128k stripe personally but it might be better to go larger depending on usage

ps. now you have a replacement you can sell me your 1231ML

[SteveRo]

These settings are best for lowest latency:



I use a 128k stripe personally but it might be better to go larger depending on usage

ps. now you have a replacement you can sell me your 1231ML

I will let you know if I decide to replace the 1231ML.
I have high hopes for the 9260-8i - you don't sound to encouraging.

[Computurd]

oh lord once you get this card fired up you will be so frickin blown away...anyone wanna buy a areca 1680-ix?

[Spoiler]

I'm still awaiting more info about the 9210 card.

[SteveRo]

I'm still awaiting more info about the 9210 card.

Other than the Tom's review can you find any other mention of it?

[Spoiler]

Other than the Tom's review can you find any other mention of it?

No, but I am seeing a number of different HBA and Raid controllers coming out with 6 Gb/s SAS /SATA technology.

http://www.supermicro.com/SAS2/SAS2.cfm?pg=AOC

and

http://www.wiredzone.com/itemdesc.as.../AOC-USAS2-L8i

It looks like its based of of LSI's SAS 2008 IC.

[Biker]

I will let you know if I decide to replace the 1231ML.
I have high hopes for the 9260-8i - you don't sound to encouraging.

In what way do I not sound encouraging?

The card is awesome, fast and so far reliable.

The only downside is slightly higher latency than the Areca due to SAS overhead.

How can this level of available card bandwidth be discouraging:

[Spoiler]

I'm sure the card is pretty awesome. I would love to have one too, but please...why do we have to see benchmarks using 50MB?

[Biker]

I'm sure the card is pretty awesome. I would love to have one too, but please...why do we have to see benchmarks using 50MB?

50mb was used to illustrate the card bandwidth capabilities on my setup not how fast my array is.

[XS Janus]

The card is great!
Now all we need is a non SAS version (special SSD, version perhaps*hint LSi*), and 600MB/s SSDs.
That way we could use just 4 of those SSD, and go with a cheaper card!

[One_Hertz]

I don't see why people say this card is great... Sure the max sequential speeds are high, but the small file speeds aren't that good at all. The access times are also awful like from an Adaptec. What is so great? The areca 1231 is better but even on it the small file speeds and the access times are still not on par with ICH.

[Spoiler]

I don't see why people say this card is great... Sure the max sequential speeds are high, but the small file speeds aren't that good at all. The access times are also awful like from an Adaptec. What is so great? The areca 1231 is better but even on it the small file speeds and the access times are still not on par with ICH.

As far as I know, everyone thinks this card is great because it's the only readily available card that supports PCIe 2 and performs decently well in sequential reads/writes.

I need to take a better look at access times on these cards and see what the overall impact would be on the work that general users/gamers do. Besides benchmarks what kind of real noticeable difference am I going to see when using a LSI 9260 versus an Areca 1231? I suppose it all depends on what people are doing too.

A more specific question would be:

What kind of applications/tasks would show a noticeable difference when going from an Areca Controller access time of .015 ms to .1 or .2 of an LSI card?

[One_Hertz]

What kind of applications/tasks would show a noticeable difference when going from an Areca Controller access time of .015 ms to .1 or .2 of an LSI card?

If you are going to go all "value" and "noticeable difference" on me, then my question is:

What kind of applications/tasks would show a noticeable difference when going from a single X25-M to a large SSD raid on an LSI card?

The answer is - almost no apps/tasks... Certainly nothing that I do.

The thing is... faster IS faster whether it is easily noticeable or not.

[Biker]

This card is great because of high sequential bandwidth and pci-e 2.0 specification.

It is well documented and discussed to death that it's strong points are not 4k or access times.

Can we move on now please?