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

[chinobino]

Hey Tiltevros,

I was looking at your CrystalDiskMark 2.2 sequential read performance of 181.9 MB/s thinking, '4 drives only giving 2 drives worth of performance!'

Then I saw your result of 307.5 MB/s for CrystalDiskMark 3.0 Beta 1 - it looks likes they've fixed a bug

Downloaded CrystalDiskMark 3.0 Beta 1 and ran it on the 780i using Windows 7 x64 and 4 x 80 GB Intel x25-m's (G2's, firmware = 02G9) in RAID0 [edit]128k stripe (software RAID, OS drive);

---------------------------------------------------------------
CrystalDiskMark 3.0 x64 Beta1 (C) 2007-2009 hiyohiyo
                   Crystal Dew World : http://crystalmark.info/
---------------------------------------------------------------

           Sequential Read :  699.984 MB/s
          Sequential Write :  304.597 MB/s
         Random Read 512KB :  508.305 MB/s
        Random Write 512KB :  282.222 MB/s
    Random Read 4KB (QD=1) :   21.263 MB/s [  5191.2 IOPS]
   Random Write 4KB (QD=1) :   49.615 MB/s [ 12113.1 IOPS]
   Random Read 4KB (QD=32) :   79.148 MB/s [ 19323.3 IOPS]
  Random Write 4KB (QD=32) :  109.260 MB/s [ 26674.9 IOPS]

  Test : 1000 MB [C: Used 61.0% (181.8/298.0 GB)]
  Date : 2009/12/03 20:10:58
    OS : Windows 7 Ultimate Edition [6.1 Build 7600] (x64)

You can really see the cache kick in hard on your LSI 9260i4 (you said LSI 9240i4 the 2nd time but I assume that was a typo) when reading & writing small files.

My main concern with using any PCI-e RAID card is the added latency to access times, particularly random access times, in which Intel SSD's are supposed to excel.

I ran HD Tune Pro 3.5's random access test and would be interested to compare the results to yours. I know the variability of the test makes a comparison pointless to a degree, I would just like a ball park comparison.



I am wondering if the LSI 9260i4 is only likely to show benefits (over software RAID) in sequencial reading & writing - with access times being roughly the same?

[Tiltevros]

tell me why a single drive intel x25-m G2 is faster in 4kb than ur raid and u will understand what im saying about LSI areca and all the hardware Raid controllers
Im not saying that Raid sux but.. Raid controllers are not for home use... or for gaming. Yes they give us performance but not in the way that the controller does..

[chinobino]

tell me why a single drive intel x25-m G2 is faster in 4kb than ur raid and u will understand what im saying about LSI areca and all the hardware Raid controllers
Im not saying that Raid sux but.. Raid controllers are not for home use... or for gaming. Yes they give us performance but not in the way that the controller does..

Hi Tiltevros - you make a very good point about the single drive performance compared to my RAID setup with small files.

I also would like to know the answer...

Is it hardware (780i SATA controller + SSD's) or software (I am using the default Windows 7/nvidia RAID driver)???

I ran ATTO to get a better idea of how my system handles different file sizes and the results are quite bad!



I might try the latest SATA driver from nvidia to see if it makes any difference (I don't want to derail this thread, might have to start a new one).

I wonder if the large stripe size (128k) would account for this?

I want to put the array onto a PCI-e RAID controller eventually, I was just waiting for them to be better optimised for SSD's.

[Tiltevros]

for me 256k or 512k stripe size is better for 4xRAID in x25-ms
10 years in IT i learn something in Raid arrays. maybe its wrong u tell about it.
onboard RAIDS are most of the time (80+%) soft raids.. the real power of a h/w raid controler is in Raid5+0 6+0 .
So yes i belive that 780i is a soft raid.
here is my ATTO benchmark. And tell me what to belive... ATTO ?? CrystalDiskmark??? HDD tune PRO??? HDD Tach??? iometer???
the real test is how u feel the system even in 12xRAID 0 u can tell the difference.. maybe is like 3% more but its there. the system is smoother database or task or programs response faster.

Now for the Hardware RAID user.
take the areca, the LSI, the Promise and all the other controller that u want to bench and put SAS drives to see the power of the controller
yes maybe LSI have SSD guard but is it a raid Controller for SSD????
yes maybe the ARECA much much faster than GOD in MB/s or in IO/s but is it made for SSD????
As u can see in the market many labels made PCIEx SSD can u figer WHY? why intel or OCZ or what ever the brand is says 200.000 on PCIE ssd and not in a 2,5".
We all here love our babies (HDDs SSDs RAIDs) but its to soon to see the real potencial of H/W Raid with SSDs.

[Tiltevros]

4c. RAID-0 Performance Scaling With # of Drives
One of the things I have always hated seeing here is people who have RAID-0 arrays with 7 or 8 drives. Aside from the clear danger of disk failure, RAID-0 may scale well with a few drives, but much less so as you add more. For example, versus one disk and assuming theoretical maximums:

1 Disk = Baseline
2 Disks = 1/2 Time Decrease = 50% performance increase vs. 1 disk
3 Disks = 2/3 Time Decrease = 16% performance increase vs. 2 disks
4 Disks = 3/4 Time Decrease = 9% performance increase vs. 3 disks
5 Disks = 4/5 Time Decrease = 5% performance increase vs. 4 disks
6 Disks = 5/6 Time Decrease = 3% performance increase vs. 5 disks

And one must take into account the fact that as the number of drives increase, so too does the minimum size of the file required to be considered a "large" file. Add in the fact that overhead alone accounts for a few % of performance and you can see that past 3 disks your *theoretical maximum* increase is sitting in the low to mid single digit range.

http://www.xtremesystems.org/forums/...26&postcount=3

Many thanks to Mr Serra for his post

[Tiltevros]

OK now its time for fun
Mr Computurd hehe
can u post the everest disk benchmark with 4 hdds please
i think that i can take over ur Hdds hehe
(joking)

[Mastakilla]

LSi is the has the fastest CPU so It should help quite a bit if you plan on doing live expansion on you array, meaning adding more drives to the array.
That process is painfully slow. And I would hate to lose a drive during it.
My second choice would be Adaptec, cause I remember it for being 4x as fast at that particular task than my own 3ware 9650.

thanks for the advice!

it is not like i'll be adding new HDs every week though
so i guess it wouldnt matter very much if it took 8 hours instead of 4 hours (I have no idea if those a realistic numbers though)

but i guess this means that the "crapp" lsi software wouldnt really be a big problem, right?
I read yesterday that more specificly the controller app outside the OS (runs after bios post) is a real disaster. I guess that is the tool I need to use to setup the RAID 5 initially, right?

Also, what about that damn uberexpensive battery? Should I buy it? If for example the power falls away, then it cant write the data from the cache anymore. Can this actually corrupt my RAID 5 causing me to loose everything? Or will it only cause to loose the file it was writing at that moment maximum (and any other files it still had to write and were still in the cache)?

If it is possible to actually get a corrupted RAID 5 because of a powerloss, then will a battery be able to prevent that?

[Tiltevros]

do u know why this "crappy" controller needs the battery??? or in general do u know why H/W Raid needs the battery??
to enable write back cache in a raid controller u need the BBU. The BBU stands for the cache memory that if u have a power fail it will prevent data loss. so... if u want to make a serius Raid 5 u will buy the bbu..
Now for the crappy bios post... ok biosweb is nothing compare to ARECA but... how many times u will go in RAID bios?????? when u can do anything in MEGARAID Storage Manager that rocks...
Its really bad for a user to say wierd words for something that he even test it... Test it and then express ur self. let me tell u something.
Areca is superb... fast easy to use in bios... BUT CAPPED in 800MB/s... LSI is ok but CAPPED in 2.5GB/s .. so??? witch is better??? Areca is 3Gb/s per port LSI is 6Gb/s per port so??? and????
areca (12 series and 16 series) cost here in greece 800 euros LSI cost 412 with bbu. areca is a mature product (means that it is in the market for long time...) and LSI (9200 series) in 6month old...
now for the corrupted RAID that u said. the array will be there.. the data that was in the cache will be there for 72 hours... but the HDDs maybe wont be there :P so.. if u want to prevent data loss buy a UPS and stay in ICH or Nvidia soft RAID.

[Mastakilla]

do u know why this "crappy" controller needs the battery??? or in general do u know why H/W Raid needs the battery??
to enable write back cache in a raid controller u need the BBU. The BBU stands for the cache memory that if u have a power fail it will prevent data loss. so... if u want to make a serius Raid 5 u will buy the bbu..
Now for the crappy bios post... ok biosweb is nothing compare to ARECA but... how many times u will go in RAID bios?????? when u can do anything in MEGARAID Storage Manager that rocks...
Its really bad for a user to say wierd words for something that he even test it... Test it and then express ur self. let me tell u something.
Areca is superb... fast easy to use in bios... BUT CAPPED in 800MB/s... LSI is ok but CAPPED in 2.5GB/s .. so??? witch is better??? Areca is 3Gb/s per port LSI is 6Gb/s per port so??? and????
areca (12 series and 16 series) cost here in greece 800 euros LSI cost 412 with bbu. areca is a mature product (means that it is in the market for long time...) and LSI (9200 series) in 6month old...
now for the corrupted RAID that u said. the array will be there.. the data that was in the cache will be there for 72 hours... but the HDDs maybe wont be there :P so.. if u want to prevent data loss buy a UPS and stay in ICH or Nvidia soft RAID.

euh... you clearly misunderstood me...

When did I ever say it was a crappy controller??
I also know why the controller needs a BBU, I just don't know the exact impact / risk of not having one...
I know the differences you mentioned between the Areca and the LSI, I just don't know "how bad" the LSI software is (since I dont have experience with either the Areca, nor the LSI, I don't really know how this "bad" software will impact me...)
From the reactions here I think I can conclude that it will not impact me much...

About the last part of your message, I cant really follow... data will be in the cache for 72hours? In which case? LSI with BBU? And how do you mean that the HDs may not be there anymore?

I'm not planning to buy an UPS and I certainly dont want to use ICH or Nvidia soft RAID for a RAID 5 since I don't want those horrible write speeds... The main reason I want a RAID 5 is not because I could loose one little piece of data (the piece in the cache) in case of a powerloss. The main reason is that I don't want to have a chance to loose ALL data in case a HD gets broke or something else happens... That is also why I wonder if an LSI without BBU has the chance of loosing all data in case of a power loss...

ow yes, and where the hell did you find the LSI with BBU for 412 euro? In Benelux the LSI will only be for sale in about 10 days (i hope). The cheapest webshop will have it then for 433 euro. And the BBU costs 112 euro minimum... So thats 545 euro at least...

[XS Janus]

I think the array expanding goes at about 4MB/s on my 3ware 9650 and I believe it was about 12MB/sec on Adaptec 5405... so you do the math.
I don't thing we are talking hours here.
Remember it is HEAVY disk trashing during this

It made me invest a whole lot more, and go for 8 drives at once. Didn't want to risk it going wrong in the process.

[Tiltevros]

www.span.com
there 2 ways that the power losses.. 1 way i dont know how to say it in english... but i will try to express it in my own way... the power goes up and then dives down like eg if ur power socket ensure in stady way 220volts for miliseconds goes to 260volts then to 175Volts and then 220volts that power goes in to ur psu and after psu inside ur pc. the second way is when the power goes from 220 stady volts to 175volts and then goes up to the 260 volts. the second way for me is more more more dangerous couse the loss of the power reduse from minimum to max and that gap is very painfull to electronics.
ALL serius RAID controllers have bbu's. when u work ur data for seconds are on the cache of the controller those 512mb or more that the controller have. if the power loss hit u in that moment ur data stays in the controllers cache for 72 hours top. thats what u pay for the bbu. For a home user maybe this bbu is useless but for enteprises is gold.
Now for home users the only benefit that the bbu stands is for Write Cache Back. if u dont have the bbu u cant enable it.

[Tiltevros]

rebuilding and expading on LSI
6Gb/s SAS Performance Advantage
6Gb/s SAS is designed for backward compatibility
with 3Gb/s SAS as well as with 6Gb/s and
3Gb/s SATA hard drives. Regardless of the drive
speed, 6Gb/s MegaRAID controllers will deliver
significant performance improvements in both
read and write applications as compared to their
3Gb/s predecessors. Other new features of the
6Gb/s MegaRAID SATA+SAS controllers include
greater signal integrity using Decision Feedback
Equalization, support for more sophisticated
SAS topologies using Expander Self-discovery/
Self-configuration, and additional safeguards to
enhance data protection.

http://www.lsi.com/DistributionSyste..._PB_072209.pdf

[F.E.A.R.]

I think the array expanding goes at about 4MB/s on my 3ware 9650 and I believe it was about 12MB/sec on Adaptec 5405... so you do the math.
I don't thing we are talking hours here.
Remember it is HEAVY disk trashing during this

It made me invest a whole lot more, and go for 8 drives at once. Didn't want to risk it going wrong in the process.

What???

I expand my Raid 6 from 3TB to 4TB (netto) in only 17h (total time !!!)
The 5805 has more CPU-Power than my 1261-ML, but it´s not faster.
A consistence-check takes 3:15h (326MB/s.)


@ all

SAS-HBAs are optimized for SAS-Drives. Areca SAS-HBAs simulate the SATA-Protocoll only with software. I think all other controllers too!?


@ Tiltevros

Areca´s ARC-12xx-Series (IOP341) is old, but fast with little files.
60.000 IOps @ 512Byte and 0,02ms in HDtune - Random Access is no problem.

We need more realworld-benches.

[Tiltevros]

What???
@ Tiltevros

Areca´s ARC-12xx-Series (IOP341) is old, but fast with little files.
60.000 IOps @ 512Byte and 0,02ms in HDtune - Random Access is no problem.

We need more realworld-benches.

thats my point. its old mature and bug fixed.. let the 9200 series to mature and we will see how it perform.

[NapalmV5]

will it mature up to areca 1231 ??



^ sneak peek of whats coming

[Tiltevros]

sure why not

[SteveRo]

The 12XX controllers won't stay on top forever - nothing does.
Actually you could argue the 12XX is no longer on top even now.
Both nizzen and computurd have turned in some very high pcmark vantage numbers using sas controllers (areca 1680 and LSI 92XX).
See - http://www.futuremark.com/community/...pcmarkvantage/
Is the next king of the hill the LSI 92XX - could be.
Actually I'm rooting for a bootable pcie direct connect solution - much cleaner (if less versatile) formfactor.
Will ioxtreme get there - who knows ...

[XS Janus]

Napalm, that is your array? the one with 4 SLC drives? Which ones are they again?

[XS Janus]

What???

I expand my Raid 6 from 3TB to 4TB (netto) in only 17h (total time !!!)
The 5805 has more CPU-Power than my 1261-ML, but it´s not faster.
A consistence-check takes 3:15h (326MB/s.)

Thats how I remember it go on my card. I realised that while still experimenting with 4 drives. SO I went out and asked around the web ans saw others having similar speeds on my card.
After that I asked about 5405, since it has better cpu, and a member on some forum posted that his did ~12MB/s when he divided the time it took him.
That's my experience with it... I'm not saying it is the absolute truth.
I'll do more tests on this subject, when I will build another server.

Consistency cheks are no problem on mine, as well

[F.E.A.R.]

Intel RS2BL080 the same perf. like the LSIs?

[Nizzen]

What???

I expand my Raid 6 from 3TB to 4TB (netto) in only 17h (total time !!!)
The 5805 has more CPU-Power than my 1261-ML, but it´s not faster.
A consistence-check takes 3:15h (326MB/s.)




We need more realworld-benches.

I expanded my Raid 6 from 4TB to 6TB with my Areca 1680ix-12. 8x samsung f1 1tb. It took 9h

Not bad at all I think.


Anyone know what the new HighPoint RocketRAID 4460 24P PCI-Ex 2.0 will perform?

RAID 5/6, PCIe 2.0 x8, 1024 MB RAM, Intel IOP348 (1.2Ghz), 6x int. mini-SAS

[Mastakilla]

www.span.com
there 2 ways that the power losses.. 1 way i dont know how to say it in english... but i will try to express it in my own way... the power goes up and then dives down like eg if ur power socket ensure in stady way 220volts for miliseconds goes to 260volts then to 175Volts and then 220volts that power goes in to ur psu and after psu inside ur pc. the second way is when the power goes from 220 stady volts to 175volts and then goes up to the 260 volts. the second way for me is more more more dangerous couse the loss of the power reduse from minimum to max and that gap is very painfull to electronics.
ALL serius RAID controllers have bbu's. when u work ur data for seconds are on the cache of the controller those 512mb or more that the controller have. if the power loss hit u in that moment ur data stays in the controllers cache for 72 hours top. thats what u pay for the bbu. For a home user maybe this bbu is useless but for enteprises is gold.
Now for home users the only benefit that the bbu stands is for Write Cache Back. if u dont have the bbu u cant enable it.

was that a one time offer at span.com when you got the controller and battery for only 412 euro?
Cause now I see the LSI 9260-8I for 421,42 euro and the battery for 120,90...

I have 2 APC SurgeArrests for 8 outlets each to protect all my electronic devices, so power spikes shouldnt be a problem.
Power drops and complete power loss however can still occur.

In case of power loss I wouldnt mind very much to loose what i was working on. However the chance of loosing all data because of power loss causing a corrupt RAID 5 is unacceptable (that is the reason why I'm investing in this expensive RAID controller, I do not want the chance anymore to loose all data. And I'm tired of putting everything on DVDs :p)

So can anyone confirm that I don't need a BBU to ensure that I will not loose all data?

About the write back cache: Interesting... I didnt know you cant enable it without a BBU... Does it matter much in performance? Or where lies the difference?

Thanks!

[Mastakilla]

btw:
Xbit posted a nice (I suppose at least :p still reading) review of this card!!
http://www.xbitlabs.com/articles/sto...as9260-8i.html

[F.E.A.R.]

I expanded my Raid 6 from 4TB to 6TB with my Areca 1680ix-12. 8x samsung f1 1tb. It took 9h

Not bad at all I think.


Anyone know what the new HighPoint RocketRAID 4460 24P PCI-Ex 2.0 will perform?

RAID 5/6, PCIe 2.0 x8, 1024 MB RAM, Intel IOP348 (1.2Ghz), 6x int. mini-SAS

9 h ist very fast. I will verfy this

[Nizzen]

9 h ist very fast. I will verfy this

Good

I used the 80% "priority". I think it is the max. Used storage manager in windows

I was a bit worried when I started the online expantion because the array was full when I started