Patriot Memory’s latest Wildfire Sata 3 SSDs come with the promise of read and write speeds in excess of 500MB/s along with read and write IOPS in the 60K region. All of this blazing speed is available in 120GB (£240), 240GB (£420) and 480GB capacities – a 60GB version is in production, but not currently available in the UK. …
"So I'm inclined to think my rig might not be particularly SandForce-friendly as, given the right hardware, reports suggest this drive does indeed deliver on its promises."
I'm more inclined to think Patriot are claiming performance figures that their products cannot deliver - it certainly wouldn't be the first time. The only worthwhile performance metric for SSDs that is worth looking at is random-write IOPS using 4KB blocks. Most of the rest is purely for making the marketing brochures look impressive.
80%? A reg review hasn't got that in a while....
SSD Testing Hint
Don't rely on your motherboard's SATA controller. Get a quality (And relatively cheap) LSI SASII (compatible with SATAIII / 6Gbit) HBA (Not a RAID controller, there's too much black box magic on those.) The LSI SAS 9211-4i is a lovely 4 port card that will eliminate any questions as to whether your motherboard SATA controller disagrees with a given drive.
It's also been reasonable well documented that the top-end performance of a 6gbit drive on a 3gbit controller has little relation to what you'll see on a 6gbit controller. And by 'documented' I mean either I read it somewhere on the Internet, or I did it myself... probably the former, since this sounds like a lot of work that I'd remember complaining about. Anyway it's down to latency and queue depth and things that all get better on the newer interfaces..
Also (haven't had my coffee yet, so apologies if I missed this) it doesn't seem that you specify what kind of data you're writing to the drive. Sandforce controllers have lots of magic going on, so if you use repeating test patterns, or compressible test patterns, you'll see vastly different performance than if you have a utility which uses non-compressible, unique test patterns. I strongly suspect that many Sandforce vendors quote the best possible performance numbers found using the most unrealistic data possible.
Also, this drive shouldn't have any power fail protection, as the Sandforce 22xx line isn't supposed to support that feature. Can't see myself buying a drive that doesn't have a bank of capacitors on it these days, and I wouldn't recommend anyone else does either.
Will they resolve the Sandforce controller issues?
Since OCZ, Corsair, Intel, Crucial and others have had Sandforce controller and other SSD issues, I'm not ready to jump in and lose my data. It would be nice if they could sort out the compatibility and operational issues with these SSDs as they could be a nice performance improvement if they were reliable.
Gbps not the only useful measure
It all depends on what you are doing with your drive, as to whether the maximum transfer rate is a significant issue. If you are doing lots of small transfers then the IOPS measure is the significant one, and the amount of data being transferred won't flatten even a 1.5Gbps SATA.
Booting an OS is quite likely to be an example of the latter. Can't we have time to boot an OS as a real-world benchmark number? (Obviously it has to be the exact same O/S image every time).
Re: Gbps not the only useful measure
>>Can't we have time to boot an OS as a real-world benchmark number? (Obviously it has to be the exact same O/S image every time).
Even the same image each time may speed up with drives such as the Seagate Momentus Hybryid.
So far, the fastet drives I've found are the pcie based ones, no controller mismatch issues, obviously no good for 2.5" applications, but then not all 2.5" SSDs work in all situations like some disk bays (the SSD profile tends to be square but disks have a recess).
Why so much money? A 120GB Crucial M4 is only £180 and looks to be very nearly as quick.
the received wisdom on the hardware sites is that Marvell controllers don't do terribly well with the new-gen Sandforce drives; the Intel P67 onboard 6Gb controllers are generally approved of. Might want to re-test with one of those.
You mention the Vertex 3 running the same controller, yet don't compare them
I will not buy this ...
... on account of it having a stupid name. "Patriot Wildfire"? Calm down FFS, cut back on the red meat. It's a storage device not some Tom Clancy wet dream.
...is the company name.
- Analysis BlackBerry Messenger unleashed: Look out Twitter and Facebook
- Comment Mobile tech destroys the case for the HS2 £multi-beellion train set
- Nine-year-old Opportunity Mars rover sets NASA distance record
- Things that cost the same as coffee with Tim Cook - and are WAY more fun
- IT bloke publishes comprehensive maps of CALL CENTRE menu HELL