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

[SteveRo]

Does anyone have a LSI MegaRAID 9210-8i yet?
Looks like it might be a good controller?
Maybe less $ also.
see - http://www.tomshardware.com/reviews/...l#xtor=RSS-182

[Biker]

i'm after only one thing...... X58 compatibility

Speaking of compatibility.... this card is not compatible with the DFI X58 UT T3eH8

Works fine with the EP-45UD3P though, but it does drop my GFX to 8x as not enough lanes available to run at 16x 8x...

[Spoiler]

Does anyone have a LSI MegaRAID 9210-8i yet?
Looks like it might be a good controller?
Maybe less $ also.
see - http://www.tomshardware.com/reviews/...l#xtor=RSS-182

As with anything that looks awesome.....

It's either not out yet or it really was too good to be true.

[justin_c]

I ordered the 9260-8i a few days ago. Hunting for X25-M G2s.

[Biker]

LSI 9260-8i RAID 0 with 2 X25-M G2's 128k stripe:



Not sure about the 4k results but the sequential is looking good

[SteveRo]

LSI 9260-8i RAID 0 with 2 X25-M G2's 128k stripe:



Not sure about the 4k results but the sequential is looking good

I agree, looking good except the 4k read number looks low.
Run iometer 4KB random read 100%, queue=1 for 1-2 min (yes, I know, queue of 32-64 is a more realistic load).
For a single 80GB G1 - using these settings in iometer- I was getting 93 MBps - this (I think) on an nvidia 680i controller.

[XS Janus]

I see Tom's listened and learned from their last half baked article!
They vindicate themselves well with 3.4GB/sec results!

[SteveRo]

[QUOTE=XS Janus;3985055]I see Tom's listened and learned from their last half baked article!
They vindicate themselves well with 3.4GB/sec results!

Yes - impressive! - I wonder when that 9210 controller will be available?

[XS Janus]

They didn't give any access time info, though...

Hmmm, what if ICH11 controller will be on par with LSi 9210!
Niiice

[saint-francis]

They didn't give any access time info, though...

Hmmm, what if ICH11 controller will be on par with LSi 9210!
Niiice

WTF is ICH11 and when can we expect to see it? Googleing it produces about 50,000,000 hits worth of garbage.

[XS Janus]

Dude, ICH10 -->ICH11 (*hint*-Intel-*hint*)

But I'm just hopping here. I have no idea or clue haw fast it will be.
Seeing how ich10 caps at 600MB/s we need it to be quite a bit faster that that!

[saint-francis]

I know that it is the supposed successor to ICH10 and I know it's Intel. I just don't know anything else. Is there even going to be an ICH11? WTF is P55/P57? Is it just ICH10 without a NB? I have seen no news at all about a new south bridge. If you google it you come up with a million hits on conversations on forums much like this thread here.

[SteveRo]

Does anyone have a LSI MegaRAID 9210-8i yet?
Looks like it might be a good controller?
Maybe less $ also.
see - http://www.tomshardware.com/reviews/...l#xtor=RSS-182

Who is going to be the first to run a complete set of tests on the LSI MegaRAID 9210-8i controller?
It is now commercially available as the INTEL RS2BL080 .
See - http://www.valleyseek.com/product/show/168571
Looks like this controller should be faster than the 9260 at raid 0?

[Speederlander]

This nearly laser-like focus by everyone on XS on sequential access speeds is kind of funny.

[Serra]

This nearly laser-like focus by everyone on XS on sequential access speeds is kind of funny.

Speeder, you and I both know that access time only affects things when opening programs, accessing non-sequential media (ie. many games), and increasing the overall response time of a system. Get off your high horse and understand that all anyone wants to do is stream blu-ray content from one hard drive array to the other already! [/sarcasm]


Still, there is something to be said for sequential scores. On a platter based system, provided I know the number and spin speed of the drive I don't need to see access time benchmarks because the controller really isn't going to change it by more than 0.1ms on a test, and that's within standard deviation anyway (unless something is wrong). Well, we could go into it a little deeper, but I'm not sure the mob really ever cared about getting that deep.

But if we want to try to max out a card or get the best performance we're using SSD's anyway, and suddenly it's a whole new ball game. I know some people turn up their noses at it, but I think that there is something to be said for IO Meter results these days.

And where are the RAID-6 results? RAID-1 results with platter-based drives (I would love to see this one to see how it's implemented)?

[XS Janus]

SteveRo: why do you think it will be faster than LSi 9260?

[lowfat]

SteveRo: why do you think it will be faster than LSi 9260?

take a look at the review that he linked

[SteveRo]

SteveRo: why do you think it will be faster than LSi 9260?

I am wondering if it might be faster based on info/results in the Toms review -

http://www.tomshardware.com/reviews/...id,2388-3.html

See your own post above - http://www.xtremesystems.org/forums/...5&postcount=82.

About the 9210 controller Toms says - "The 9210-8i, aka Intel RS2BL080, offers the same number of SAS/SATA ports as the 9260-8i, and the PCB has exactly the same dimensions, but the product is somewhat different, and much simpler overall. It does not feature a powerful RAID-on-Chip solution with XOR acceleration for RAID 5 and RAID 6; it does not come with cache memory; and it doesn’t feature a battery backup unit. However, that strip-down may be helpful in terms of throughput, as there is no processing unit and no cache memory that might be trying to improve performance while effectively limiting it in our scenario (as was the case with Adaptec’s RAID 5805). Of course, it also helps with pricing, we're sure. We don’t have much information on the 9210-8i, as it has not yet been officially released. It also utilizes PCI Express 2.0 (eight lanes) as one of the first products on the market, and it is ready for SAS/600 drives as well.".

[lowfat]

I am wondering if it might be faster based on info/results in the Toms review -

http://www.tomshardware.com/reviews/...id,2388-3.html

See your own post above - http://www.xtremesystems.org/forums/...5&postcount=82.

About the 9210 controller Toms says - "The 9210-8i, aka Intel RS2BL080, offers the same number of SAS/SATA ports as the 9260-8i, and the PCB has exactly the same dimensions, but the product is somewhat different, and much simpler overall. It does not feature a powerful RAID-on-Chip solution with XOR acceleration for RAID 5 and RAID 6; it does not come with cache memory; and it doesn’t feature a battery backup unit. However, that strip-down may be helpful in terms of throughput, as there is no processing unit and no cache memory that might be trying to improve performance while effectively limiting it in our scenario (as was the case with Adaptec’s RAID 5805). Of course, it also helps with pricing, we're sure. We don’t have much information on the 9210-8i, as it has not yet been officially released. It also utilizes PCI Express 2.0 (eight lanes) as one of the first products on the market, and it is ready for SAS/600 drives as well.".

It also costs a bit more

[Computurd]

Question, if that there card has no ROC, why is there that big ass heatsink on there?

[Biker]

Who is going to be the first to run a complete set of tests on the LSI MegaRAID 9210-8i controller?
It is now commercially available as the INTEL RS2BL080 .
See - http://www.valleyseek.com/product/show/168571
Looks like this controller should be faster than the 9260 at raid 0?

That Intel card you linked looks exactly a 9260-8i which i suspect it is in all but name...

[Spoiler]

Question, if that there card has no ROC, why is there that big ass heatsink on there?

I'm looking for the link right now, but I heard the 9210 does have some type of raid chip. It just doesn't have hard Raid 5 support. Not sure if that makes sense or not. It sounds to me that it should be more effective at Raid 0 and 1 setups.

After thinking about it, it's got to have something on there (controller) if its using 512MB of onboard memory. I can't wait to see some more in depth reviews of the card. (Even if it is from Toms Hardware).

Also available at Provantage: http://www.provantage.com/intel-rs2bl080~7ITSP0R3.htm
It looks like you are paying for the Intel name though.

[Computurd]

Here Tom's is a bit wrong - it is a 2008-based (an abridged version of 2108) card. RAID 5 is software only, but otherwise the same chip. W/o RAID 5 it does not need additional memory (built-in is enough). W/o RAID 5 overhead should be a bit faster in RAID 0.

Think will come in several incarnations - PCIe x4/x8 and different number of ports and drives supported (HW only supports up to 12 with 10-drives max. array - with advanced sort/firm-ware may support more - see Supermicro site - they've announced several SAS 6G UIO cards based on the same chip)

atrrib to stasdm at ocz...he dudg that up...no cache no raid 5 or 6 on the 9210....even though it is technically faster the feature set is way cut down

[XS Janus]

Hmm, yes I thought of that... But the link to Intels controller SteveRo linked states

512MB embedded memory

and Tom said that 9210-8i it doesn't have cache.
The Provantage link even says it supports Raid5,6...
So no I'm confused...
Is the RS2BL080 same as 9260-8i or as 9210-8i?

[smallfish]

LSI Corporation (NYSE: LSI) is a leading provider of innovative silicon, systems and software technologies that enable products which seamlessly bring people, information and digital content together. The company offers a broad portfolio of capabilities and services including custom and standard product ICs, adapters, systems and software that are trusted by the world’s best known brands to power leading solutions in the Storage and Networking markets. More information is available at www.lsi.com.

(1) Estimated MSRP. Actual prices may vary.
(2) 9260DE-8i and 9280DE-8e include SafeStore encryption services.