What is faster for SSDs, ICH10R or Areca 1231ML-2G?

OK so I guess Napalm has already addressed this but I wanted to post my results anyway.
This is a rather unscientific test but perhaps a useful compare none the less.
In the first set of tests I copied a single 5.13GB image file from one Mtron array to another Mtron array.
In the 2nd set of tests I copied a 4.84GB folder of 6967 items from one Mtron array to another Mtron array.
In the 3rd set of tests I copied the same single 5.13GB image file from one Acard 9010 array to another Acard 9010 array.
In the 4rth set of tests I copied the same 4.84GB folder of 6967 items from one Acard array to another Acard 9010 array.
Each array is set up exactly the same: 2xR0, primary, NTFS, "Turn off Windows write-cache buffer flushing ..." set to on.
All ICH10R arrays are set to 128 stripe size.
All 1231ML arrays are set to 64 stripe size.

In these tests it appears the 1231ML arrays were nominally twice as fast as the ICH10R arrays.
Interesting that moving from the slower Mtron Mobi’s to the faster Acard 9010’s the increase in performance of the 1231ML over the ICH10R arrays remains substantial.

Also interesting to note the large difference in the ICH10R tests.
Take a look at the table.
In the initial copy test - the time was always lowest - then with a "delete and recopy back" to the original drive it takes almost twice as long - curious.
You can see this in both the single and multiple file copy test, both for the Mtron’s and for the Acard’s.

Looking at the 4.84GB folder copies on the1231ML, this might actually work the other way here, the first copy took the longest time and the second and third seemed to take less time.

Table of test results is as follows –

http://img178.imageshack.us/img178/833/copycompare.jpghttp://img178.imageshack.us/img178/c...jpg/1/w654.png

Various benchmarks of each set of arrays –

2xMtron Mobi 3000 R0 ICH10R –

http://img245.imageshack.us/img245/5...alibration.jpghttp://img245.imageshack.us/img245/i...pg/1/w1523.png

2xMtron Mobi 3000 R0 1231ML-2G –

http://img188.imageshack.us/img188/7...alibration.jpghttp://img188.imageshack.us/img188/1...pg/1/w1523.png

2xAcard 9010 R0 ICH10R –

http://img198.imageshack.us/img198/4...librationb.jpghttp://img198.imageshack.us/img198/i...pg/1/w1520.png

2xAcard 9010 R0 1231ML-2G –

http://img6.imageshack.us/img6/3474/...alibration.jpghttp://img6.imageshack.us/img6/1231m...pg/1/w1522.png

Great tests!
Thanks for all the trouble.

Could you please run R/W test on just one array, as well. I would like to see how each controller handles simultaneous read and write tasks on a single array. :)
DAMN areca is fast! I know its cache plays a good role in it, but hey , that's what its there for... :D

Dont forget to save the folders that you made, so you can use them in the future and make fair comparisons of other suetups without retesting old ones. ;)

Nice work Steve, that sorts out the wheat from the chaff quite nicely.

Do the Mtron's stutter on ich10?

Do you confound the Copy-Benchmarks Mobi-ICH10R and Mobi-Areca?
I can´t believe that...

Please try "FC-Test"...
This app write real files to the SSDs.

Look @ my old thread (MTRON-Review)
There´s a big different between Onboard or Areca (ARC-1210)

http://www.xtremesystems.org/forums/...d.php?t=172012 (please look the FC-Test-Results)


And why are your Copy-Benchmark-Results so bad (2x Acard @ Areca)? :confused:

My results

http://www.xtremesystems.org/forums/...4&postcount=39
http://www.xtremesystems.org/forums/...70&postcount=7

Could you try some actual loading times of apps/games? Copy paste tests are not too useful...

We must find a test for everyone. I mean a self-made-test by (e.g.) XS

I give NapalmV5 the idea to create a *bat.
His bat includes by this time 106 apps.
It´s to time-consuming for me to create a bat with more than 100 apps,
because i´m working with my sys. every day (the long time for big benches is over)

And i must use exactly the same apps. Otherwise it´s all for nothing.

Great test SteveRo, thanks for sharing.

Quote:

---

Originally Posted by XS Janus

Great tests!
Thanks for all the trouble.

Could you please run R/W test on just one array, as well. I would like to see how each controller handles simultaneous read and write tasks on a single array. :)
DAMN areca is fast! I know its cache plays a good role in it, but hey , that's what its there for... :D

Dont forget to save the folders that you made, so you can use them in the future and make fair comparisons of other suetups without retesting old ones. ;)

---

"Run R/W on a single array"? - what am I writing to and reading from?
Read and write at the same time from one acard 2xR0 to another acard 2xR0?
Yes - I saved both the image file and small file folder for future tests.

Quote:

---

Originally Posted by F.E.A.R.

Do you confound the Copy-Benchmarks Mobi-ICH10R and Mobi-Areca?
I can´t believe that...

Please try "FC-Test"...
This app write real files to the SSDs.

Look @ my old thread (MTRON-Review)
There´s a big different between Onboard or Areca (ARC-1210)

http://www.xtremesystems.org/forums/...d.php?t=172012 (please look the FC-Test-Results)


And why are your Copy-Benchmark-Results so bad (2x Acard @ Areca)? :confused:

My results

http://www.xtremesystems.org/forums/...4&postcount=39
http://www.xtremesystems.org/forums/...70&postcount=7

---

I have not run FC test before, it is on my to do list.
"And why are your Copy-Benchmark-Results so bad?" - The performance is about what I expected.
Maybe I did not explain the set up very well - this is copying from one acard "box" to another acard "box".
Each "box" is set up as 2xR0 via the back panel jumper.

I just finished reading through your 2xR0 hamma on areca 1210 post - excellent work.
I think my results are consistant with yours.

Quote:

---

Originally Posted by One_Hertz

Could you try some actual loading times of apps/games? Copy paste tests are not too useful...

---

I have COD4 or would you reco something else?

Quote:

---

Originally Posted by audienceofone

Nice work Steve, that sorts out the wheat from the chaff quite nicely.

Do the Mtron's stutter on ich10?

---

I have never noticed the mobi's stutter - not once.
That said - the mobi small file random writes are really bad aren't they - even worse when compared next to the acards I suppose.
I never realized how bad they were prior to having AS SSD runs on them.
The mobi's are SLC based, that is probably what saves them from the dreaded stutter problem.

wow, really nice test steve :) i seriously need to stop browsing the storage section... or else i wont be able to pay for my next semester =(

Quote:

---

Originally Posted by GunterFalstaff

wow, really nice test steve :) i seriously need to stop browsing the storage section... or else i wont be able to pay for my next semester =(

---

Gunter - wait a bit - ioxtreme and a selection of the next gen h/w raid controllers should be available soon.

Thanks for sharing these results SteveRo

Awesome setup(s)!

Got any links with any more info on those next gen h/w raid controllers SteveRo?

E.g. these controllers http://www.xtremesystems.org/forums/...d.php?t=231054

Quote:

---

Originally Posted by F.E.A.R.

We must find a test for everyone. I mean a self-made-test by (e.g.) XS

I give NapalmV5 the idea to create a *bat.
His bat includes by this time 106 apps.
It´s to time-consuming for me to create a bat with more than 100 apps,
because i´m working with my sys. every day (the long time for big benches is over)

And i must use exactly the same apps. Otherwise it´s all for nothing.

---

chicken!! :rolleyes: :)

reason why you wont do it/try.. maybe you have and your acards are still half the performance of my slc ssds :rofl:


whats the big deal? you got ati.. run the ati app/s.. you got diff sound card.. run you sound cards app/s.. etc

thats the 100app.bat

but for the 50app.bat my 106app.bat alone contains enough apps.. all the apps ive used are free/attainable but you guys dont have to go with any of that

whatever you guys want to compile the 50app.bat of.. no problem @ 50 app load i can do ~7 apps/second ;)

Ah those, yeah saw the first few post in the thread, gonna be interesting to see more tests of these cards.

@ Napalm

Who give´s you the idea for a *.bat, chicken? :ROTF:
We both know that DRAM-based SSDs are faster than Flash-based-SSDs. ;)
I´ve no time for games...

OK - so I spent some quality time learning FC Test today.
Much thanks to the folks at X-Bit Labs for a very useful tool –

http://www.xbitlabs.com/articles/sto...y/fc-test.html

Ok – so I ran FC Test against –

1. 2xR0 Acard on ICH10R, both the windows (boot?) pattern and my same home grown large folder of 6967 items totaling 4.84GB.
2. Ran the same FC Tests against the same 2xR0 Acard array but this time on Areca 1231ML-2G.

Results – The 1231ML still wins but the delta is much less than the initial "stopwatch" tests (1st post of this thread).

FC Test results table –

http://img36.imageshack.us/img36/990...aresummary.jpghttp://img36.imageshack.us/img36/fct...pg/1/w1021.png

FC Test screen captures –
Windows pattern – 1st ICH10R then 1231ML –

http://img41.imageshack.us/img41/126...indowspatt.jpghttp://img41.imageshack.us/img41/fct...jpg/1/w778.png http://img36.imageshack.us/img36/590...indowspatt.jpghttp://img36.imageshack.us/img36/fct...jpg/1/w722.png

Home grown folder pattern 1st ICH10R then 1231ML –

http://img208.imageshack.us/img208/6...omegrown69.jpghttp://img208.imageshack.us/img208/f...jpg/1/w768.png http://img32.imageshack.us/img32/699...omegrown69.jpghttp://img32.imageshack.us/img32/fct...jpg/1/w836.png

Quote:

---

Originally Posted by F.E.A.R.

@ Napalm

Who give´s you the idea for a *.bat, chicken? :ROTF:
We both know that DRAM-based SSDs are faster than Flash-based-SSDs. ;)
I´ve no time for games...

---

games? you guys just chicken out thats the game :shrug:

ive seen what your dram based ssd can do @ real apps.. unfortunately far from impressive :(

you think youve seen the best of my system?? faster ssds soon.. next goal is to reach ~10 apps/second @ high load ;)

heres my fctests:

http://img527.imageshack.us/img527/5838/90577133.png
http://img526.imageshack.us/img526/6514/95593394.png
http://img521.imageshack.us/img521/7769/91324190.png
http://img411.imageshack.us/img411/8056/24745490.png
http://img341.imageshack.us/img341/1170/68672894.png

Quote:

---

Originally Posted by NapalmV5

games? you guys just chicken out thats the game :shrug:

ive seen what your dram based ssd can do @ real apps.. unfortunately far from impressive :(

you think youve seen the best of my system?? faster ssds soon.. next goal is to reach ~10 apps/second @ high load ;)

---

At last check I thought you were running 4xJmicron SLC drives?
What make/model? Array stripe size @ 64?
Anything unusual in the setup that you would like to share?

Quote:

---

Originally Posted by SteveRo

At last check I thought you were running 4xJmicron SLC drives?
What make/model? Array stripe size @ 64?
Anything unusual in the setup that you would like to share?

---

excuse me ???

heres the link once again: http://www.youtube.com/watch?v=MvBKB2NndbI

the same freakin 64k stripe ive tout about around here

the whole setup its in the vid! nothing unusual


the fact of the matter is.. all your acards are good for is benchmarks but when it comes to real apps.. 10x :down:

sorry.. please show me otherwise

Quote:

---

Originally Posted by NapalmV5

heres my fctests:

http://img527.imageshack.us/img527/5838/90577133.png
http://img526.imageshack.us/img526/6514/95593394.png
http://img521.imageshack.us/img521/7769/91324190.png
http://img411.imageshack.us/img411/8056/24745490.png
http://img341.imageshack.us/img341/1170/68672894.png

---

I will run a full FCTest set later today - against 6xR0 Acard 9010 on the 1231.
We can compare.