OCZ Vertex 3 Pro Preview: The First SF-2500 SSD
by Anand Lal Shimpi on February 17, 2011 3:01 AM ESTFinal Words
I didn't really believe that SandForce could pull it off when I first heard how fast the SF-2000 line would be. Even after CES, I didn't really believe the drives would be this good in real world use cases. Consider me pleasantly surprised.
When connected to a good 6Gbps controller, the Vertex 3 Pro is significantly faster than anything else on the market today. Obviously the V3P itself is an unreleased drive so things could change as its competitors show up as well, but the bar has been set very high. The Vertex 3 Pro is the first SSD to really put 6Gbps SATA to good use. In fact I'd say its the first drive that really needs a 6Gbps interface. Whenever you Sandy Bridge owners get replacement motherboards, this may be the SSD you'll want to pair with them.
Even writing incompressible data the Vertex 3 Pro is faster than current SandForce drives running full tilt. The performance gains we see here are generational, not a simple evolutionary improvement. SandForce has also successfully addressed the limited shortcomings of the original SF-1200 controller with regards to writing incompressible data.
Clearly performance isn't going to be a problem with this generation. The real unknowns are how well will the Vertex 3 (non-Pro) perform and how reliable will these drives be? Intel is still king of the hill when it comes to drive reliability, however OCZ has been investing heavily in improving its manufacturing. I suspect that this next SSD war will be fought both along performance and reliability lines. Unfortunately for us, the latter is very difficult to quantify without a significant sample of drives.
With new controllers from SandForce, Intel and Marvell due out this year we're going to see SSD performance go through the roof and SSD prices to continue to fall. We're still a couple months away from knowing exactly what to buy, but if you've been putting off that move to an SSD - 2011 may be the year to finally pull the trigger.
Facebook
Twitter
RSS
144 Comments
View All Comments
eva2000 - Thursday, February 17, 2011 - link
Thanks for the review, curious why the 256GB Crucial RealSSD C300 only included in some benchmarks and not others ?Anand Lal Shimpi - Thursday, February 17, 2011 - link
That's simply a data entry issue, I'll get it fixed :)Take care,
Anand
therealnickdanger - Thursday, February 17, 2011 - link
Also, I'm curious as to why none of the REVO benchmarks are included? On the one hand, I understand how niche PCIe drives may be, but on the other, they can offer significant performance over both SATA 3Gbps and 6Gbps drives. It would be nice to see the new Heavy and Light benchmarks applied to those as well.ludikraut - Thursday, February 17, 2011 - link
+1 on including REVO drive results in the drive comparisons.markjx1 - Thursday, February 17, 2011 - link
because REVO drives are garbage. they're nothing more than a bunch of vertex2's slapped onto a board with a cheap Silicone Image controller that has a PCI-X to PCIe bridge. Resale value on REVO is crap and if one of the drives dies, you have a nice paperweight.bji - Thursday, February 17, 2011 - link
Your own personal opinion on the drives are not a reason to disclude them from the review.therealnickdanger - Thursday, February 17, 2011 - link
Garbage that can pull down 800MB/s...? Go away, troll. For users without SATA 6Gbps, it's a very practical solution to achieve huge speed. In fact, even people WITH SATA 6Gbps would get a boost. The cost/GB is even under $2 for some of the REVOs, making them a reasonable alternative.Anand Lal Shimpi - Thursday, February 17, 2011 - link
All graphs should now have the 256GB C300 in them :)Breogan - Thursday, February 17, 2011 - link
Is the data in the Incompressible Write Speed Test mixed for the SF2500 drive or does it perform actually worse after trim?. It seems weird to me that a dirty drive performs as if it was stock and the trim actually worsens its perfomance.Anand Lal Shimpi - Thursday, February 17, 2011 - link
It's not TRIM that is making the drive slower, it's the next write pass that's actually pushing the drive into a lower performance state as there's more garbage collection/cleaning that's going on. If I hadn't TRIMed and just ran another pass of the AS-SSD test you'd see the same number.Take care,
Anand