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

[Boogerlad]

x25-m g2s?

[justin.kerr]

G1's

[lutjens]

You guys had to post this just when I got my 1280ML working on my new motherboard???

Mean....very mean...

[XS Janus]

Ah, things are heating up!

Do real world tests to!!! Not only dry benches.
Please setup, a 4xR0 array on the new card and ICH10 and do some .bat file for automatic app loading and time it, like Napalm did in his thread!
Also timing some software installs form .iso images on the array and file copy times on the array using a >2GB directory full of music and jpegs should give us the clear real world difference between ICH and the new LSI!

Thanks!

[F.E.A.R.]

Question...

Does 6Gb mean 6Gbit or 6GByte???
Cause there is a big difference between Gb an GB

With 6Gbit/s (or ~ 750MByte/s.) you can also buy an "old" controller

On Wikipedia-Sites i found Gb as Gigabit and GB as GigaByte

[lowfat]

Question...

Does 6Gb mean 6Gbit or 6GByte???
Cause there is a big difference between Gb an GB

With 6Gbit/s (or ~ 750MByte/s.) you can also buy an "old" controller

On Wikipedia-Sites i found Gb as Gigabit and GB as GigaByte

6Gbit. That is the drive interface (6Gb/s SAS). Has nothing to do with overall bandwidth of the controller.

[F.E.A.R.]

Oh f***

[Boogerlad]

has anyone tried the new intel controllers that are on pic-e 8 gen2?

[Spoiler]

has anyone tried the new intel controllers that are on pic-e 8 gen2?

?

Not sure if you are aware, but this LSI card is based on the new Intel controller that features PCIe gen2. Still waiting to see any benches though.

[One_Hertz]

Someone has this card over on OCZ forums. They claim over 1GB/s with 5 vertex attached, but it's ATTO so it is not very useful.

[Gogeta]

Someone has this card over on OCZ forums. They claim over 1GB/s with 5 vertex attached, but it's ATTO so it is not very useful.

Link

I'm interested to see if it improves the performance of SATA2 mechanical drives compared to IOP341 controllers. I suppose it depends wholly on the firmware and RAID instruction set used by LSI.

[Boogerlad]

Oh. I got owned then. Thanks for the clarification guys. =)

[Spoiler]

Link

I'm interested to see if it improves the performance of SATA2 mechanical drives compared to IOP341 controllers. I suppose it depends wholly on the firmware and RAID instruction set used by LSI.

Thanks for the link. By the looks of his benchmarks, I don't see anything too amazing. I still would like to see some other setups though.

[F.E.A.R.]

When will the controller appears in europe?

[Nizzen]

When will the controller appears in europe?

It is in stock here in Norway The 8 port version.

If the man on OCZ forums is able to get over 1500 MB/s with 8 Vertex ssd, I buy it.

But i do not think it is goming to happend. Seams Maximum reads: 2,875MB/s
is just BS.

But I wait and see

[F.E.A.R.]

How much is that controller in norway?

Another option to connect 2 (or more) Areca's on a MB to get 3,2GB/s. (or more)

[Nizzen]

MRSAS9260-8I/KIT 5000 norwegian kroner VS Areca 1680ix-8 7000 NOK

[F.E.A.R.]

Hm, 5000 NOK are ~ 576€

The 1680ix 8 costs 553€ in here germany http://geizhals.at/deutschland/a340734.html

And i need not less than 12 Posts (currently 16)

[XS Janus]

looks like 14 second boot initialization time. Not bad.

[Biker]

Got the nice OP with the new LSI card over at the OCZ forums to fire up AS SSD (mostly to check out the latency) but he is pulling some pretty serious throughput there too

Post#119



Results:

[NeedMoMegaHurtZ]

Looks like an interesting controller and a decent price, but I would love to see some real tests, maybe you can coax your "buddy" at the ocz forum to do some.

I wouldn't put too much stock in that AS SSD benchmark, at least for sequential rates. If you look at SteveRo's thread http://www.xtremesystems.org/forums/...d.php?t=231629 , he has 2xmobi (which have about 100read/90write) , but on the areca it's getting 1500/1400 , and 200/160 on ICHR9 (even 2x Acard gets the same sequential #'s as 2xMobi on that benchmark when on the Areca...). One possible explanation is that the Seqential test is too small and all reads and writes are cached. A better , more relevant test for STR would be to do some file copy tests with larger test pattern (e.g. 1 with a large file, multiple tiny ones, at least larger than a few GB , etc...)

Of course other real tests would be welcome, like game level loading times, app loading, maybe even napalm's crazy bench

[Biker]

Totally agree.

Forget the AS SSD sequential.

I was mostly interested in the latency myself which seems to be right up there with Areca (well almost ...)

Lets face it folks we do not have any definitive SSD array benchmarking tools right now...

Either way I can currently get the new 9260-8i in the UK for £382 which is a large saving over the previous gen Areca...

[One_Hertz]

I was mostly interested in the latency myself which seems to be right up there with Areca (well almost ...)

What? It is like worse than Adaptec. 0.3ms for vertex is horrible. The 4k numbers are horrible too.

[Biker]

What? It is like worse than Adaptec. 0.3ms for vertex is horrible. The 4k numbers are horrible too.

The guy has barely got the card and has no experience configuring LSI cards for that matter so that probably explains the .3 read times....

Write access time is excellent.

4k does not look great this run but I don't care about it in this context because i'm sure something is not set optimally in the card / virtual array settings.

Btw what read / write latency were you getting with your Adaptec and how does that contrast with onboard?

[One_Hertz]

The guy has barely got the card and has no experience configuring LSI cards for that matter so that probably explains the .3 read times....

Write access time is excellent.

4k does not look great this run but I don't care about it in this context because i'm sure something is not set optimally in the card / virtual array settings.

Btw what read / write latency were you getting with your Adaptec and how does that contrast with onboard?

0.22 on reads for adaptec. 0.095 onboard. I dont remember the write access times TBH, but those aren't too important anyway.