Good morning all,
General Observations –
Controllers: Both the LSI 9260-8i and the Areca 1231ML-2G take about the same time to start up – seems like ~ 20-25 sec.
Both seem very reliable, I did not encounter any errors or dropped arrays, which is saying something because I have significantly stressed both controllers over the last two days.
Both controllers use the same cable connectors – I used both Highpoint and Areca (fan out) cables with no problem.
The size of the 9260 is an advantage in tight places and it seemed that the 9260 ran cooler than the 1231.
The 9260 has a couple advantages over the 1231 – pcie 2.0 bandwidth to the motherboard and price.
The 1231 transfers data to the motherboard at only pcie 1.0 and is $200-300. more expensive than the 9260.
The 1231 has a couple advantages over the 9260 – 2GB vs. 512MB onboard memory cache and the 1231 offers native SATA support.
The 9260 is both a SAS and SATA controller in one, perhaps SAS via h/w and SATA via s/w? Could be.
Drives - I use what SSDs I have on hand.
The SSDs I have are perhaps not representative of the average SSDs that most people are now using.
Obviously, your results will vary based on the drives you use.
The Acard ANS-9010 is a DRAM based SSD and the Mtron Mobi 3000 is an early generation SSD based on SLC memory.
Both are SATA (not SAS) drives.
From a random access speed perspective the acard and the mobi are at somewhat opposite ends of the performance spectrum.
The acard 9010 is extremely fast at random IO while the mtron mobi is extremely slow (4KB writes) - by today’s standards (about the same as a good platter drive).
In sequential read writes, both are on the slow side – the acards come in at ~ 175/150MBps read/write per drive while the mobi’s read/write at ~ 100/80MBps per
drive.
Test set up – Test system is i7 965 on a Gigabyte Extreme mobo, 3x1GB G. Skill 2000/C9, 2xGTX295 Quad SLI.
Boot drive via Mtron Mobi 16GB on the ich10r controller.
All arrays are standalone on the controller (only the array to be tested is attached).
Clocks - cpu @ 4.48 (35x128) ht off, memory @ 1792 (8-8-8-30), pcie @101
LSI 9260-8i set up – I tested several variations and found that Biker’s set up was the best – see – http://www.xtremesystems.org/forums/...&postcount=112
All 9260 arrays are set to 64-stripe size (optimum for the arrays used in this test).
Areca 1231ML-2G set up – After testing several variations I found that the default set up is the best.
All 1231ML arrays are set to 64-stripe size (again optimum for the arrays used in this test).
Acard array set up - Four acard 9010s, all set (via jumper) to raid 0.
So 8xR0, each drive is 4GB (2 sticks of 2GB each DDR2).
Mtron Mobi 3000 array set up – 4 16GB SSDs all in R0
Summary of performance – Well, for the SSDs that I can bring to the test, I don’t think there is a clear winner here; both controllers have their strengths and weaknesses.
The LSI 9260-8i has the best sequential read speed with the acard array and the best workstation IOps with the mobi array.
The Areca 1231ML-2G has the best 4KB random speed for both arrays, the best sequential speed for the mobi array and the best workstation IOps for the acard array.
See the iometer summary table below –
Several benchmarks follow for each array on each controller - interesting that HDTach doesn't seem to work on the 9260 for reads.
Benchmarks of Acard ANS-9010 8xR0 on LSI 9260-8i -
Benchmarks of Acard ANS-9010 8xR0 on Areca 1231ML-2G -
Benchmarks of Mtron Mobi 3000 16GB 4xR0 on LSI 9260-8i -
Benchmarks of Mtron Mobi 3000 16GB 4xR0 on Areca 1231ML-2G -
Reply With Quote
Originally Posted by NeedMoMegaHurtZ
Bookmarks