***31/12/2007 UPDATE
Installed the Highpoint 3520 Controller last night on my 2nd PCIe Slot on a GA-EP45-DS3.
The Highpoint 3520 is a 8x PCIe Controller witch fits also in a 16x Slot. Thanks God that the Gigabye Board supports the full 8xPCIe Speed.
I've read the MacPro's only supporting a lower Bandwith on their 2nd PCIe Slot. Can anyone confirm this?
See the Teste here at www.dv.com - It says: "...these features and more are found in the RocketRAID 3522 ($560), an 8x PCIe 1.0 card that works in any PC or Mac with PCIe slots. Note that in a Mac Pro (early 2008), it will work at 4x speeds in either of the 4x slots but will only work a 1x speed in the 16x slot (which only supports PCIe 2.0 cards at full speed). I tested the card in a Mac Pro (early 2008), dual 3.2 with 14GB RAM running OS X 10.5."
1st Try - So far so good. Did a clone of my existing OS X 10.5.6 Retail, booted up from the Clone. Installed the RAID Driver Restart.
Hang! No way to boot up! No Idea what the probleme is here. Will figure that out in another Test.
2nd Try - Cloned a Backup of an existing OS X 10.5.5 Retail. Installed the Driver and reboot! Conected 3 Samsung F1 320GB SATA HDDs to the Highpoint.
Made an Array via Highpoint RAID Webinterface (easy to use - very fast Array setup!) Went to DiskUtility - formatted the RAID Array into one Partition.
Voila - the RAID comes up (Unfortunatly with a Folder Icon, but with the Function of a Partition!).
Next Step - Cloned the Exitsting OS X i was on to the new RAID Array.
After that I used Xbench on the Raid Array. Amazing Performance - but I think Xbench is not the right Tool to Measure that Kind of Setup.
Had Write Speeds arround 750MB/s and Read Speads arround 500MB/s. Overall Points: +-750!
Restart - Next Step Booting from the RAID Array.
Quick check with CTRL+H in controller setup choosing the RAID Array as Boot Device.
Restart and WOW. System comes up much more faster than with a single Disk (20 sec.) even than with a RAID0 Softraid i had with the 3 HDDs before.
So far so good. Don't want to waste the Time in testing the HDDs. I wait for the SSDs coming up an concentrate for that test.
Conclusion so far:
+ The Highpoint 3520 works on a OSx86 Hackintosh and seems to have a great Performance
+ The Controller was only handwarm (have to Test a Raid5 with the SSDs!)
- Had Problems with driver install and reboot on 10.5.6 but I think this should be solved quickly, maybe someting was wrong with the DSLT Patch...
***29/12/2007 UPDATE
Cause Supertalent ws not on stock I decided to switch form the Super Talent SSD to the new OCZ Vertex SSD Drive
I hope this one will be shipped early January...
Highpoint 3520 arrived today - will test this one with 3x F1 Samsung 320GB HDDs tonight
Fantec Housing also arrived, unfortunatly it was defect. 2 Slots did not close in the aluminum cage... Cheap Plastic!
More Infos coming up...
***27/12/2008
Hey guys,
here i am preparing a little report of a SSD-RAID-Setup I ordered one day before X-Mas.
8x Super Talent 32GB SSDs (approx. 150MB/s read, 100 MB/s write) 76,- EUR each
1x Highpoint 3520 SATA Raid Controler for OS X 445,- EUR
2x FANTEC MR-SA1042-1 Backplane for the 8 SSDs 68,- EUR each
What I like to test:
1st - Highpoint 3520 Raid Controller under OSX86!
2nd - Speed of 8 SSDs in a Raid-5 Array on the Highpoint 3520 Raid Controller
The Highpoint Contoller should be able to max out the speed and performance even of cheap MLC SSDisks, cause of his 256MB Cache.
I would be very happy to achieve read-rates from 120-130 MB/s and write-rates round about 60-70 MB/s.
In a Raid-5 this should be something about +-800MB/S read and 420MB/s write. The Raid 5 should give me overall aprox. +-200GB of Diskspace
I will give you all the details in setting up the stuff, installing OSX etc. in details when the hardware arrives.
If i can get that the speed out of the SSDs it think it is worth the price I paid! And I use this workstation for work not for private purpose!
By the Way, here are some links who gave me the inspiration to try this:
Thanks to drmanic - Video on youtube
Battleship Mtron
...to be continued
