M14x R2 with 2 Mushkin 120gb in Raid 0 (msata/sata) Tests from ATTO/CDM/AS

So I decided to buy 2 ssd's to test out raid 0 in my m14x.

Mushkin Chronos DX 120gb sata3 - max read 560mb/s - max write 515mb/s
Mushkin Atlas 120gb sata3 (msata) - max read 555mb/s - max write 525mb/s

Both use Toggle NAND (MLC). I tried to match up the best drive I could with the Atlas. There are less expensive 120gb Chronos drives but they do not use Toggle which the Atlas does.

I'm not an expert and this isn't an extensive review. I just wanted to show numbers of before, after, etc. so others can see. I was curious myself about raid 0 performance and couldn't find any solid answers.
Hopefully this will answer some questions for people.

Equipment:
Alienware M14x R2 (900p screen)
i7 3610QM
8gb DDR3 1600
2GB GT650M
ODD still inside
500gb Seagate 7200rpm (stock)
120gb Mushkin Atlas DX mSATA SSD
Newegg.com - Mushkin Enhanced Atlas Series MKNSSDAT120GB-DX mSATA 120GB SATA III MLC Internal Solid State Drive (SSD)
120gb Mushkin Chronos DX SATA SSD
Newegg.com - Mushkin Enhanced Chronos Deluxe MKNSSDCR120GB-DX 2.5" 120GB SATA III MLC Internal Solid State Drive (SSD)

Base:
BIOS is set to RAID for all tests.
Intel RST 11.6 installed prior to all tests.
Both drives are running latest 5.0.4 firmware.
No other drivers installed, except for the first stock test.
Two CDM tests done. Default (Random) and All 0x00 (0 FIll)


First test (500gb Seagate 7200rpm drive that came from Dell ST9500423AS)


120gb Mushkin Atlas mSATA installed while keeping 500gn in bay. Windows installed on Atlas SSD.


120gb Mushkin Chronos DX installed while keeping Atlas mSATA in bay. Windows installed on Chronos SSD.(I forgot to run AS SSD.)


Both drives in Raid 0. WIndows originally installed on the Chronos drive. Intel RST used to create Raid 0 array. (Notice Intel RST 11.6 supports Raid 0 trim.)


My take on all these numbers.
According to ATTO, both of the drives are operating correctly (Mushkin uses ATTO.) ATTO shows higher numbers overall.
Both drives are operating very close to each other in terms of R/W speed.
I'm unsure of which tests are more real life performance? Even just in CDM. Going from Random to (0x00 Fill), there's almost a 600mb/s write difference.

Overall:
Seeing the numbers are great. I didn't really notice any decrease in windows boot time going from either drive separately to the raid. It would be nice to eliminate the Raid Configuration Utility and the Alienware post screen at boot to load even faster. I spent $250 for these two drives for 240gb (223gb usable). I'm happy with my purchase and hope these drives serve me well for a long time to come.

If I missed something or there are unanswered questions, I will answer them to the best of my ability. If there are anymore tests to be done, let me know and I will run them.

 

That highlighted point is the key takeaway for me of your well documented test of raid 0 on an R2. I suspect however, that there should be other operations that can better display the value of a raid storage speed increase over a single SSD drive.

Perhaps somebody more knowledgeable can chime in on what to try other than a benchmark.

+1 rep to you.

 

Since it is sandforce, you get better score when dealing with compressible data(ATTOS/Fill), real life will be somewhere middle of compressible/incompressible.

so msata can raid with 2.5 afterall? Gonna try that when I get a msata on my m6600 ~~

edit: windows boot up tend to be random read/write (thus real slow on platters), so raid may not show improvement.

 

you should turn on the write back cache in the RST utility. it makes a massive improvement in small file size write speeds. probably make that .5 10,000 go to 70,000. i have a very similar setup (120G Force GT/ 120G Atlas) but must have an issue with my msata in that my read on the msata wont break 350 by itself. I am running RST 11.2 and in raid I get the following.

 

So I enabled write cache. I couldn't do it through RST. I kept defaulting to disable upon every reboot.

I had to enable it by going through device manager.

Here are the results:



Some improvement I must agree.

It seems like Windows is taking longer to load than when first installed. Granted I have stuff loaded on it now, but I figured it wouldn't take as long as it does.

I was roughly at 7-8 seconds after post. Now I'm 12-13. Going to have to dig to find out what is going on.


*EDIT* Would it have something to do with me disabling page file? I doubt it but I'm somewhat confused on what it is.

 

My experience is that leaving a small pagefile behind seems to work better than disabling it. Some programs won't even run without a pagefile. The worst example I can think of right now is Dawn of War II that requires a minimum pagefile of 1.5GB to run. Found that out the hard way when I disabled the pagefile on my desktop computer, which at that time had 8GB ram.

Personally I always create a fixed 2GB pagefile on my system partition and leave it at that.

 

@thauch:

Thanks for sharing your numbers -- they look really great.

This is entirely a taste question, but do you feel any performance increases (or decreases) while gaming?