Now I’ve added an Areca RAID card to the mix. This is starting to look like a worse and worse investment all the time. I wanted to try with a few different cards because I figured “Hey, now I’ve got an Areca, so it’ll just be a matter of 3 different products that function properly, but have varying levels of performance”.

WRONG!

Card #1 (ARC-1210) makes my POST & BIOS screens refuse to show. Windows boots fine, and from there I can configure RAID sets, but that’s not something I can accept as a 24/7 solution.

Card #2 (ARC-1222) boots fine in the machine, allows me to create a RAID set out of the four X25-M drives, gives great read performance, but then takes a big dump on the write performance (far worse than onboard), wavering up and down between about 225MB/s and 50MB/s (no consistency to it)

Card #3 (ARC-1680ix-8) failed to boot several times and simply hung at the “Waiting for RAID card firmware to become ready” screen the first couple attempts. I even left it overnight with that screen up and the little indicator spinning away. I decided to give it one more shot this morning to see what happened and it booted! I was thrilled, but write performance still sucks. Read/write caching are both enabled (as far as I can tell) and I’m almost at the end of my rope here…

I’m far from releasing all of my benchmarks and experiences with this new setup, but I have run into a couple of problems worth mentioning in my quest for the best possible storage performance.

First I attemped my new setup with a 3ware 9690SA-4I and I created a forum thread over at XtremeSystems.org documenting my failure with that… What a waste of time that card was… After that I decided to move on to the Onboard, but there ARE bottlenecks and compromises that have to be made doing that.

I don’t notice the same freezing/hitching or lock ups that some have claimed exist with Jmicron controllers on cheap SSDs, but I *do* notices a sluggishness at times. I don’t think it’s related to the SSDs, but rather to the southbridge. One way to replicate it reliably is the Windows Experience Index. If anyone can shed some light I’m very interested to hear why this happens during the CPU Test…

Well I decided to do some follow up performance numbers that include more drives as well as RAID-0 results with the Seagate 7200.2 drives, and the long and short of it is that I will be replacing my Samsung F1 with a pair of Seagate 7200.2s when I do my next reformat.

The performance difference is not staggering, but when combined with the reduced noise, reduced power consumption and reduced footprint (I”ll be sticking the 2.5″ drives to the back of my motherboard with industrial velcro), it’s an easy decision to make.

Here are some comparisons to some other drives that I found interesting. The mobile drive’s performance looks quite disappointing in direct comparison to the F1 750GB, but when compared to a more equivalent drive (Seagate 7200.10 160GB) it’s a closer fight.

I couldn’t help noticing that all solid state drives are coming out in 2.5″ and smaller form factors, so it got me thinking that it’s only a matter of time before a high end gaming chassis is full of 2.5″ mounts rather than 3.5″ ones.

Obviously SSD isn’t quite there yet, but even 2.5″ magnetic drives seem to compare pretty favorably to 3.5″ desktop drives. They’re quieter, smaller (duh), and with the 7200RPM spindle speeds, the assumption is that the speed is pretty close. They also dump less heat into your precious case, which means less airflow is necessary and again less noise.

I managed to get my hands on a 160GB Seagate Momentus 7200.2 and a Samsung F1 750GB in a real David & Goliath battle. I’ve only run a couple benchmarks for now because I want to get into RAID performance, which I think is really important because as they’ve discovered in servers, you can fit several times more 2.5″ drives in the same space.

Seagate Momentus 7200.2 160GB 2.5″ “Notebook” Hard Drive

Samsung F1 750GB 7200RPM 3.5″ Desktop Hard Drive

I guess my results weren’t that impressive, but it’s important to notice that as far as the access times go, the drives were actually very close. That’s important for multi-tasking, as well as for quick access to small files. It’s not that often you’re actually going to need the full read and write bandwidth of your drive in a typical desktop scenario, but that obviously doesn’t apply to I/O intensive operations like video editing or systems with limited system RAM that are using the swap file all the time.

Hoping to do a follow up shortly with RAID-0 performance for the 2.5″ competitor.