LSI 9211-8i

[Biker]

also there is the limitation on stripe size which is particularly depressing

Don't think I have ever seen a HBA offering over a 64k stripe....

Anyway this card is going to be the 4k iop's king.... leave the big sequentials to the "proper" raid cards

[XS Janus]

Do you guys have any "E" Intel drives?

[One_Hertz]

hey one hertz, i have an idea. you should prove to us your loading time so that we have something to compare to. you yourself should put up a video of your load times. maybe some kind of storage manager so that we can see the devices used. maybe do it on a 24.7 system without ln2?

prove it!

Sure, but it will be after this Friday (exam on that day).

[Tiltevros]

ok here is some results with 256k file in sequential with 8 i/os per target .. its the best that i can do today after a long bad sad difficult day

system at stock q9550 with 4gb ram @1333mhz and 5xintel X25ms on 9211i8
at the weekend i will try something special if i have the time

sorry for been so late with results.

[Tiltevros]

16k random same setup.

[Tiltevros]

8k random with 8 io/s

[Computurd]

nice results tiltevros! i have made significant improvements with the card and the level load times so in real world testing it is fast. the large sequential for me seem to just not be there LOL. however, it is blazing fast with game loads. this card takes longer to initialize than the 9260. for some reason it likes to hang on the "verifying dmi pool" session of loading for about 15 seconds.that is what make s the boot seem longer. the 92609 did not do that, i am wondering if yours does the same or is it my board? i tried cmos etc but it still hangs there every boot cycle.
first load no mercy loads in 5 and dead air is loading in 3.25. i really wish that i had l4d before i switched arrays so that i could compare the two. i am going to switch the arrays out and test on the 9260 a few things so that we can get a clear comparison.

[Nizzen]

nice results tiltevros! i have made significant improvements with the card and the level load times so in real world testing it is fast. the large se

Why not post results?

pcmark 05 hdd test and pcmark vantage hdd test etc? We want numbers, not ferytales!

[SteveRo]

The numbers are looking pretty good so far!
How about the 4k random numbers 100% read, 100% write? say at least 8 oustanding I/Os up to 64?

[Computurd]

ok 4k 100 percent random 16 outstanding i/o..i have reached higher with the 9260 than that however, so i am going to do some testing in the direction of the 9260 for comparison. i have hit 307 mb/s on 4k random on the 9260-8i, but being able to get that performance outta this product within its price point is outstanding. also curious as to its results with a software raid, using it merely as a pass-through device, however that is not bootable so i will not go in that direction. that is however the way that they received those humongous results in toms hardware testing is a software raid. i am not interested in non bootable raid arrays however. there is the option to use this as an expander for the 9260-8i as well, but since i only have eight disks that would not provide any improvements..unless....
@nizzen i will run pcmark vantage. i am really remiss to put up pcmark05 hdd tests as many people are using dynamic arrays and switching through them as the benchmark progresses, using each array tailored for the test involved. this muddies the water with pcmark05 testing in general so i now shun those tests there is no sense testing against that BS lol

[SteveRo]

Computurd - wow, great results - I think my best 1231ML/8xR0acard 4k 100% random read is ~52Kiops, my best ich10r/6xR0acard 4k 100% random read is ~78.6Kiops.
What is your array again - 8xR0vertex?
Last I saw - I think Nizzen's array was all vertex also.
What are your best 4MB (100 % sequential read) on the 9211?
The 9211 at $233 (provantage) is looking pretty good so far!

[Jor3lBR]

05 hdd tests as many people are using dynamic arrays and switching through them as the benchmark progresses, using each array tailored for the test involved. this muddies the water with pcmark05 testing in general so i now shun those tests there is no sense testing against that BS lol

Computurd, we still want results from PCMark05 because they are valid and it's a great way to compare benchmarks

We don't care about the tweaks people do for HWbot purpose only we know that your run of PCMark05 is UNTWEAKED so please don't bother with lower results.

This is my best PCMark05 run using an untweaked PCMARK05 only thing I did was OC the GPU and CPU and use a single Kingston SSD for Windows Vista32 and PCMark05 is on a dedicated 4 x 32GB Super Talent MLC fresh raid_0 array:

http://service.futuremark.com/compare?pcm05=2151106

I just need better HD general Usage and anti virus performance.

Thanks for your help man.

[Computurd]

that is the only weakness LOL the 4k sequential. and it is a strange one, one that has been holding me back from releasing the full results to be honest. the 4k sequential on this controller in hardware raid mode is actually slower than the 4k random. however, only by a few mb/s. it is like the 4k random and sequential is the same basically. it is a strange phenomenon and one that i have been in contact with lsi support for two days about lol. This card was released last month, and it is on the initial firmware. i have been informed that more firmwares will be forthcoming, and i think that again certain aspects of performance will be improved as they work with the product, much like the 9260-8i and its small file performance. that is why i am resorting to alot of real world tests with this card. i don think that either A. Then benchmarks aren't playing nice with the controller, or B. that they need to tighten up the firmware.

this is the latest correspondence from LSI:

In general, you would be correct that sequential reads and writes should be greater than the random variables. The HBA in use, however, is designed with random IO as it’s primary use and when in use with SSD’s, performance can be seen higher than it would be with sequential. This is partially due to the nature of solid state drives. The same experience would not be found with standard hard disks.

so there is the one issue with the device. however in the real world usage that i am seeing it is blistering fast, and i mean not by a little by a whole lot! it is hard to compare to the lsi because of the difference in the sequential performance when benching. and realistically they both seen the same speed when running applications etc. the difference in real speed may be as finite as a second here or there. we are definitely at the limit of what the system architecture can handle imo. my array is an eight member array of 30gb vertex. i will do further testing and see if i can beat the pcmarkvantage #14 spot that i got with the 9260-8i. but then again i got that result before the small file performance had been fixed on the 9260, so would it really matter if i beat it?. then i will put the 9260-8i back on.
there is one more variable that i have not forseen, and that is the firmware version for my drives is 1.41 right now, but it has shown considerable speed loss compared to version 1.3 and that is for sure, tons of people over at ocz forums have noticed it. so i will be flashing back to 1.3 in order to do testing for pcmark vantage? who knows lol...such is life when your pushing the limits! so many configurations and things to change/test/compare. it is maddening at times, but fun! The only thing i can guarantee you, this is card fast as .

[Computurd]

Decidedd to play with a dynamic disk before i re-flashed the firmware so i set it up as the lsi controller as a pass through device, just for fun of course. i am also gonna set up a software raid0 between the 9260/9211 later with four disks on each array, in raid 0 on each array
should be very interesting. wonder what level load times will be with this???
4k 16 outstanding io 100 percent random

now how bout some 4k with 32 outstanding i/o !!! one hundred percent random

[Computurd]

no mercy 3.28 seconds
dead air less than 3!!!!
i am measuring until the load bar is full form the second i click the "play game" until it is full.

@JOL i have tested with the various pcmarks, however they really grade highly on the sequential performance and the sequential performance of this card just doesn't seem to be there tbh. as far as i see currently it would not benefit your specific need. however, in the future if better firmwares...etc who knows? also i still need flash back to the better firmware on drives that is skewing my results i believe. i will keep you posted. i am far far far from done horsing around with this new gear!!

[Tiltevros]

---------------------------------------------------------------
CrystalDiskMark 3.0 x64 Beta2 (C) 2007-2009 hiyohiyo
Crystal Dew World : http://crystalmark.info/
---------------------------------------------------------------

Sequential Read : 571.120 MB/s
Sequential Write : 234.896 MB/s
Random Read 512KB : 596.694 MB/s
Random Write 512KB : 246.806 MB/s
Random Read 4KB (QD=1) : 19.609 MB/s [ 4787.2 IOPS]
Random Write 4KB (QD=1) : 48.719 MB/s [ 11894.2 IOPS]
Random Read 4KB (QD=32) : 294.857 MB/s [ 71986.5 IOPS]
Random Write 4KB (QD=32) : 224.602 MB/s [ 54834.6 IOPS]


Test : 1000 MB [C: Used 14.6% (53.8/367.8 GB)]
Date : 2009/12/10 10:22:21
OS : Windows 7 Ultimate Edition [6.1 Build 7600] (x64)

[Spoiler]

This card seems like a great buy. Two questions to those who own the card:

1. How is CPU utilization compared to other cards? Not sure if the non Raid chip would cause this to be significantly higher.

2. How much time does this card add while starting @ boot?

[Nizzen]

no mercy 3.28 seconds
dead air less than 3!!!!
i am measuring until the load bar is full form the second i click the "play game" until it is full.

@JOL i have tested with the various pcmarks, however they really grade highly on the sequential performance and the sequential performance of this card just doesn't seem to be there tbh. as far as i see currently it would not benefit your specific need. however, in the future if better firmwares...etc who knows? also i still need flash back to the better firmware on drives that is skewing my results i believe. i will keep you posted. i am far far far from done horsing around with this new gear!!

Do not care about the endresult of pcmark. We just want to see the results of the hdd test suite? Please

[Tiltevros]

nizzen how old r u???

[Tiltevros]

some say that its a crup... some say that it is just for boys...
ALL I KNOW is LSI 9260i8..........
i wonder where is ARECA hahahahahahahahahahahahahhahahahahahahahahahahahah haahhahaahhahahaha

http://www.diskusjon.no/index.php?se...#entry14306468

[Nizzen]

nizzen how old r u???

44 years young

What is wrong with wanting pcmark results? pcmark Vantage hdd suite is a good way to compare other raid setups I think

[One_Hertz]

I felt like testing one of these myself and since it's so cheap I ordered one express from newegg... I could have it tomorrow if I am very lucky, but Canadian newegg kinda blows so I doubt it. Also ordered a 30gb agility to free up another X25-E that's being used by my laptop now, so I will be testing with 4x X25-E.

PCMark is influenced by cache so controllers with cache will probably do better.

[Tiltevros]

nothing is wrong with wanting pcmark results and vantage... but plz with a lot of respect do me a favor... just behave like a 44 mature man and not like 5 . This threat is for benching a Raid controller not being sarkastic... everyone can post some results better than others even i can go to a datacenter and take a benchmark from a 5PT data server but what is the point of it??? i dont think the point is who is best.. or what is best... this is not a threat "BEST THING MONEY CAN BUY" and u know it. I would like to see ur experience with benchies and some advice if u want so to setup or configer or any ideas that we can do this card or to windows to revial the best performance of it. I would like to hear from u with the expirience that u have the cons and the pros for this LSI model. Trying with us. and maybe with the little expirience that we have one day could help u too..
I dont want to be mean to u. I dont like having that kind of speetches but plz respect others of their chooses and the investment that they did or do..

Im really sorry that u r the one that ball hit u
I respect ur knowledge about raid and raid cards but i want u to respect my invest...
hope that everything will be "salt and bread" as we say in Greece.

[Anvil]

Tiltevros

I see you found my benches from earlier this year

Nizzen is a nice guy, he doesn't mean any harm, he's just as interested in getting the best results from a raid card as you and I.

I'll be doing some more benchmarks on my 9260-8i using 8 Intel's sometime this Christmas.

[Tiltevros]

i have the setup with u but i miss only 2 x25m to catch u :P can u help us out plz we really need some extra knowledge plz..!!!!!!!!!!!