Compare Crucial C300 64GB versus Acard 9010 under Areca 1880 4GB Cache

[SteveRo]

Good morning all,
Compare of Crucial C300 64GB versus Acard 9010 under Areca 1880 4GB
All comparisons are 5xC300/64 vs. 5xAcards vs. 11xAcards
C300’s are looking good – SSDs are closing the gap but Acards still look to be a little faster, (at least for the settings I have used here)

The following AS SSD, CDM and pcmv HDD tests are all with 980 at 4.5 speed, pcie 100 -






The pcmv suite scores that follow are with 980 at stock (no oc), pcie 106 –



Settings for all arrays are below but with C300s at 16k stripe and acards at 4k stripe, is there a better sweet spot for the C300s? –

[alfaunits]

Try changing the volume read ahead cache to Conservative and Disabled, I'm wondering what the 4K will look like then.

[RADCOM]

11 x Acards hmmmm affordable storage

[mbreslin]

This isn't affordablesystems.com.

[SteveRo]

Try changing the volume read ahead cache to Conservative and Disabled, I'm wondering what the 4K will look like then.

Think I tried that already but will try again and post.

11 x Acards hmmmm affordable storage

I agree completely but look at the 5x performance, both C300 and acard - not far behind.

[Anvil]

Steve

Strip size was 4K, a bit of over provisioning (60GB allocated per drive, -> 4GB extra op)

Your controller/VD settings looks identical, PCIe 107, can't think of any other difference than mine was the only volume and used as the system drive.

[SteveRo]

I tried scsiport no real difference, both as c drive and not - no real difference
I tried again changing the volue read ahead - aggressive, normal, conservative and disabled - all seem to make minor changes - no significant change.
I can't do pcie 107 but usually 106.
I will try using smaller partition size but seems like I tried that already also.
Anything else anyone can think of?

[randman76]

SteveRo: What motherboard are you using? I am trying to figure out why my results are so bad. I have the ARC-1880-ix-12 with 4GB of cache connected to 5 x C300 128GB in a RAID 0 config. I even used the same raid config you used. When I run AS SSD I get sequential reads of 1200MB read and 1400MB of write (sorry at work right now and dont have any screen shots) This is the second card I tried, I originally thought there was an issue with the first card but it appears to be clear there is something with my config that is causing a bottleneck of some sort. The odd thing is when I use my ARC-1231ml with 4GB of cache my read speeds are higher. I receive read speeds of nearly 1500MB of read and write speeds of 1400MB. Those speeds are probably normal because of the limitation of the 1231ml. I have also removed all cards from my system (except video and RAID) and I am not getting any improvements.
For testing I have also enabled JBOD and tested all 5 drives. They all bench out around 350mb so seperately they appear to be OK.
Any thoughts / suggestions you may have would be appreciated.

Thanks,

Randman76


EVGA X58 Classified 4-way SLI
Intel 980X (OC to 4.2 GHZ)
12GB DDR3 Corsair RAM (8-8-8-24)
(2) EVGA 480 GTX SLI
ASUS Xonar DX2
ARC-1880-ix-12
Antec 1200 PSU
5 x C300 Crucial (128GB)

[Nizzen]

SteveRo: What motherboard are you using? I am trying to figure out why my results are so bad. I have the ARC-1880-ix-12 with 4GB of cache connected to 5 x C300 128GB in a RAID 0 config. I even used the same raid config you used. When I run AS SSD I get sequential reads of 1200MB read and 1400MB of write (sorry at work right now and dont have any screen shots) This is the second card I tried, I originally thought there was an issue with the first card but it appears to be clear there is something with my config that is causing a bottleneck of some sort. The odd thing is when I use my ARC-1231ml with 4GB of cache my read speeds are higher. I receive read speeds of nearly 1500MB of read and write speeds of 1400MB. Those speeds are probably normal because of the limitation of the 1231ml. I have also removed all cards from my system (except video and RAID) and I am not getting any improvements.
For testing I have also enabled JBOD and tested all 5 drives. They all bench out around 350mb so seperately they appear to be OK.
Any thoughts / suggestions you may have would be appreciated.

Thanks,

Randman76


EVGA X58 Classified 4-way SLI
Intel 980X (OC to 4.2 GHZ)
12GB DDR3 Corsair RAM (8-8-8-24)
(2) EVGA 480 GTX SLI
ASUS Xonar DX2
ARC-1880-ix-12
Antec 1200 PSU
5 x C300 Crucial (128GB)

With mutch cache on the controller 1gb+ the sequential tests in AS SSD benchmark is testing the cache and not the disks/ssds. It is just a tip :p

Also test one of the betafirmwares from Areca.

[Rubycon]

Evga Classified boards have poor(er) IO performance than other X58 boards. Have no idea as to why.

My REXIII has the best of the non server boards. The P6T7 is also very close. Both 759 and 762 Classified boards are considerably slower in IOPS.

EDIT: At least you can actually use an 1880 in the 762. Mine refuses to be recognized unless it's in the uppermost PCI-E slot.

SteveRo: What motherboard are you using? I am trying to figure out why my results are so bad. I have the ARC-1880-ix-12 with 4GB of cache connected to 5 x C300 128GB in a RAID 0 config. I even used the same raid config you used. When I run AS SSD I get sequential reads of 1200MB read and 1400MB of write (sorry at work right now and dont have any screen shots) This is the second card I tried, I originally thought there was an issue with the first card but it appears to be clear there is something with my config that is causing a bottleneck of some sort. The odd thing is when I use my ARC-1231ml with 4GB of cache my read speeds are higher. I receive read speeds of nearly 1500MB of read and write speeds of 1400MB. Those speeds are probably normal because of the limitation of the 1231ml. I have also removed all cards from my system (except video and RAID) and I am not getting any improvements.
For testing I have also enabled JBOD and tested all 5 drives. They all bench out around 350mb so seperately they appear to be OK.
Any thoughts / suggestions you may have would be appreciated.

Thanks,

Randman76


EVGA X58 Classified 4-way SLI
Intel 980X (OC to 4.2 GHZ)
12GB DDR3 Corsair RAM (8-8-8-24)
(2) EVGA 480 GTX SLI
ASUS Xonar DX2
ARC-1880-ix-12
Antec 1200 PSU
5 x C300 Crucial (128GB)

[Rubycon]

That's OK though - 4GB cache is massive and when your work runs mostly in the cache it sings.

With mutch cache on the controller 1gb+ the sequential tests in AS SSD benchmark is testing the cache and not the disks/ssds. It is just a tip :p

[SteveRo]

SteveRo: What motherboard are you using? I am trying to figure out why my results are so bad. I have the ARC-1880-ix-12 with 4GB of cache connected to 5 x C300 128GB in a RAID 0 config. I even used the same raid config you used. When I run AS SSD I get sequential reads of 1200MB read and 1400MB of write (sorry at work right now and dont have any screen shots) This is the second card I tried, I originally thought there was an issue with the first card but it appears to be clear there is something with my config that is causing a bottleneck of some sort. The odd thing is when I use my ARC-1231ml with 4GB of cache my read speeds are higher. I receive read speeds of nearly 1500MB of read and write speeds of 1400MB. Those speeds are probably normal because of the limitation of the 1231ml. I have also removed all cards from my system (except video and RAID) and I am not getting any improvements.
For testing I have also enabled JBOD and tested all 5 drives. They all bench out around 350mb so seperately they appear to be OK.
Any thoughts / suggestions you may have would be appreciated.

Thanks,

Randman76


EVGA X58 Classified 4-way SLI
Intel 980X (OC to 4.2 GHZ)
12GB DDR3 Corsair RAM (8-8-8-24)
(2) EVGA 480 GTX SLI
ASUS Xonar DX2
ARC-1880-ix-12
Antec 1200 PSU
5 x C300 Crucial (128GB)

good evening, my testing was with gigabyte x58 UD7 (1st edition).
I have also seen evga folks report that they have been unhappy with io performance of x58 classified pcie raid arrays.
It may just be the models with the NF200 on them?

correction - NF200

[Rubycon]

Not sure if there's a relation with NF200 as the P6T7 has two of them.
I have not tried the 760 without the NF200, however.

The problem sounds classified to me. I know horrible pun.

good evening, my testing was with gigabyte x58 UD7 (1st edition).
I have also seen evga folks report that they have been unhappy with io performance of x58 classified pcie raid arrays.
It may just be the models with the GF200 on them?

[Computurd]

i think the implementation of the NF200 on classified boards is terrible. The difference between the e759 (with NF200) and the e760 (no NF200) is night and day. i have heard others who say NF200 is no big deal on other boards, so i chalk it up to EVGA having shoddy NF200 implementation.

[randman76]

I agree. I've read that the NF200 chip caused extra latency, I just wasn't sure how much, should have gone with an Asus board. Anyway, I might head on over to a local Fry's Electronics and see what they have in terms of motherboards. I would order from NewEgg but I believe they dont allow returns on mobo's, only exchanges. (just in case a non NF200 board does not improve). I'll keep you guys posted..

Randman76

[randman76]

Well, I got an Asus Extreme III board and just my luck I somehow zapped the ARC-1880. I am not exactly sure how but when I installed it on the board and powered up it hung at the "Waiting for F/W to become ready..." screen for about 2 minutes then I got an error stating that the firmware timed out and the system rebooted. Keeps running through that cycle over and over now. Just to make sure it was not the Asus board causing it I put it back in the Classified however it did the same thing. So I think I may just return it for a refund and wait a while to try again. Darn thing is expensive anyway, need cash for xmas.

-Randman76

[Computurd]

do you have a BBU on the card? if so, disconnect it and reboot. have you changed the ram recently?

[Metroid]

Amazing comparison. How did I miss this thread

[randman76]

Well, I ended up shipping the card back. It had no BBU and I also disconnected all the drives. Per Areca tech support I tried everything I could try, just appears to be a bad card. I think I am going to try one more time though with a new card, hopefully my luck will change.

I also noticed Crucial just released new firmware for their C300's (Rev 0006)
I wonder if "corner-case performance" includes RAID?
I'll probably order a new 1880, flash the drives, and see if that improves performance.


Release Date: 11/16/2010
Change Log:

Improved power management
Improved power-up behavior after uncontrolled power-down
Improved full drive and corner-case performance
Improved TRIM behavior
Note: This is firmware from Revision 0002 to Revision 0006. It will not work for updating between any other versions. If your drive has Revision 0001 firmware, then you MUST first perform the intermediate step of updating to Revision 0002.

[boomboommusic]

Hi Steve,

Your result give me the GO to offer an Areca 1880ix-24 for my Acard ANS 9010.
I am building my new computer now, and have a major issue concerning the Acard. I know that they work with Areca 1280ML too, so I'm a little confuse after all my test...
Below is the mail I just sent to Acard concerning my issue, can I ask you some of your time to give your thought about it ?

Thanks a lot.
BBM

Problem description:

Hello,
I have difficulties to make work Areca 1880ix-24 with Acard ANS 9010 in Raid Configuration.

To start, I have ever used Acard in raid on the SATA controller of my previous motherboard (EVGA 680SLi), and it worked.
Now, on my new configuration, I try to use them on a new Areca 1880ix-24 controllers.
The areca card let me create the raid Set with my Acard, I am able to create volume Set, then I go to disk manager (on Windows 7), format the drive, sometime it take time (5 minutes during all disks are not accessible), sometime it failed (something like disk manager time out issue).
When the disk is available for windows, the performance are bad, when I copy a 20Mb file, it may hang the controller during 5 minutes (then the file is readable on the acard disk).
I have try to test the utility Crystal Disk Mark, but it still on "Preparing.." status during controller hang.
Sometime, after a too long hang, I reboot the computer hardly (with power button), after rebooting and looking at the Areca webpage, I can read lot of "2011-04-18 20:36:04 Enc#2 SLOT 06 Reading Error".
I have not make many test but the Areca plug on this slot doesn't create error when it is plug alone...
On this same controller, I have two Raid 5 array of 4 disk and they work perfectly.
Now this is the test I have made:

• Upgrade of motherboard Bios
• Upgrade of all Areca Firmwares
• Upgrade of ANS 9010 Firmware to 2.05 (and trying with other firmware)
• Testing "Medium" with SSD utility on my 3 Acard on both port for each (6 tests), results were "No Bad Sector Found".
• Trying to make RAID 0 / 5 Volume Set
• Trying to plug one areca port 2/3/4/5/6
• Trying different PCIExpress port on my motherboard (7/6/5/1)
• Trying different SAS port on my Areca

Model Name* ANS-9010 Firmware Version* 2.059
Serial Number* ex: 5629866B_1
System Configuration
Motherboard/System model: EVGA SR-2
SATA host controller model: Areca 1880ix-24
SATA host BIOS version: v1.49
Other I/O card
Operating System: Win7 x64
DDR2 DIMM brand & model Aeneon
CF card brand & model: N/A
Utility Version: SSD Version 1.6

[SteveRo]

Mr. Boom - Did you try at sata150? Jumper on the back of the 9010.
Areca says it is an acard problem, I have not asked acard if they plan to try to fix (maybe f/w?).
Dropping from sata300 to 150 probably does not hurt the acard that much anyway.

[boomboommusic]

Hi Steve,

Thanks to take of your time to answer me.
The same answer had been sent by Acard.

Dear Sir,

Regarding the Areca 1880ix-24 with ANS-9010, we don't have chance to test with that, but we do have experience with Areca 1680ix-12, for this case, we must set ANS-9010 SATA transfer rate on 1.5GB (default is 3GB). So please set ANS-9010 on 1.5G transfer rate (check quick manual for jumper setting) and try again.

I need to find some jumper now
I keep you updated of the situation.

Thanks again.

[mosie]

boomboommusic va voir tes MP sur HFR.


I have the same problem with my new 1880ix 16 with 4X acard 9010.

I have test during 3 day all possibility = NO solution.


areca 1880 can't work correctly with acard's whene is in SATA II (300) mod.

Its work whene acard's is in mono sata in sata I (150), but bandwidth SUXXXX.

1x ACARD 2x SATA II = 2X 175MB/s = 350MB/s but this mod not work with 1880ix ( HD tune bench,random,check error, work; but when you try copy data (file bench or data copy) its not work)

1X ACARD 1X SATA I = 1X 110MB/s ITS SUUXXXX and its full work with 1880ix



I have already areca 1231ML rev 2 with 4G

I have take a 1880ix principaly for UP the cumulate bandwidth 1231ML limitation: 4X acard = 880MB/s

theorical 4Xacard its 1.3GB/s 1880ix can make bandwidth.


I have contact ARECA for tel this problem but no answer.

My 8X iram not worck with 1880ix (bug in post screen areca, infinity check).

My 7X WD20EARS full work with 1880ix ( expand and migration data WORK fine).


Very need areca suport find a solution for this acard SATA II compatibility problem.


SteveRo

I think you can presise the SATA mod you have use for test.

Your result is NOT demonstrate the ACARD POWER bandwidth, its not significatif.

5X sata I its not the same that 10X sata II .



Just for fun :

4x acard in RAID0 in areca 1231ML rev2 4G ACTIF




More info for my config: http://forum.overclocking-tv.com/index.php?topic=351.0

sry for my bad english

[SteveRo]

^^ I sent an acard 9010 to areca sometime ago - they say it must be fixed on the acard side.
Still acard at sata150 on 1880/4GB works very well - see -
http://www.xtremesystems.org/forums/...2&postcount=49
and
https://docs.google.com/leaf?id=0B2A...uthkey=CIyYugY

[mosie]

^^ I sent an acard 9010 to areca sometime ago - they say it must be fixed on the acard side.
Still acard at sata150 on 1880/4GB works very well - see -
http://www.xtremesystems.org/forums/...2&postcount=49
and
https://docs.google.com/leaf?id=0B2A...uthkey=CIyYugY

GG for sending acard to areca

THX for your fast reply, dude.

Your good bandwidth its principaly due to 4G cach.

Disable this and try bench again hihihihihi ( i m not like AS SSD because its more impacted but 4G cach).

Its very bad use acard just in mono sata I mod.


You have a personnal contact in areca ?

You think areca can find a solution for this bug ?

16x acard you killing me ...

You have this items personnaly or a sample ? You sell it ?