A week after I previewed OCZ’s IBIS and the HSDL interface, SandForce revealed the specs of its next-generation enterprise SSD controller. The specs for the SF-2000 series call for up to 500MB/s sequential reads and writes, nearly saturating the newly introduced 6Gbps SATA bus. It should be no surprise that OCZ is very interested in moving away from SATA.

We met OCZ’s first PCIe based SSD two years ago with the Indilinx based zDrive. Take four Indilinx Barefoot controllers, RAID them together on a PCIe card and you’ve got a zDrive. The first SandForce based PCIe solution actually took a step backwards: the OCZ RevoDrive only used two SandForce SF-1200 controllers.

Performance was of course much better than the old zDrive. SandForce has all but put Indilinx out of our minds (and systems). But the recently announced IBIS and the suspicious unpopulated connector on the original RevoDrive made it clear that there was room for more performance.

Meet the RevoDrive x2:

Identical to the original RevoDrive in every way, the x2 adds a second PCB complete with two more SF-1200 controllers. With a total of four SF-1200s on board, running in RAID-0, you should get IBIS-like performance without the HSDL interface.

The architecture remains unchanged. To save on costs OCZ uses a PCI-X based RAID controller: the Silicon Image 3124. The 3124 has four independent SATA ports, each one connects to a SF-1200 controller.

Between the Sil3124 and the PCIe x4 interface is a Pericom PCI-X to PCIe bridge. It converts the parallel PCI-X signaling into serial PCIe. The Sil3124 can deliver 1GB/s of bandwidth to the Pericom bridge, as can the 4 PCIe lanes (1GB/s in each direction) so there are no interface bottlenecks here. A quartet of SF-1200 controllers can’t realistically push more than 1GB/s of data.

As with all RAID enabled solutions, there’s no TRIM support but you do get idle garbage collection. 

There’s no performance advantage over you running four of your own SF-1200 based SSDs in RAID-0. The RevoDrive x2 is pretty much a four drive SF-1200 SSD on a stick for those who want simplicity. 

The previous RevoDrive was supposed to be cost competitive with a two drive RAID array. Today, looking at pricing, a 240GB RevoDrive sells for $519 while a pair of 120GB Vertex 2s will set you back $480. You pay a premium for the simplicity but performance is identical to rolling your own SSD RAID setup. 

The Test

Note our AnandTech Storage Bench doesn't always play well with RAIDed drives and thus we weren't able to run it on the RevoDrive x2.

PCMark Vantage & SYSMark 2007 Performance

With a four controller SF-1200 setup the real performance advantage won't be seen on most desktop workloads.  The RevoDrive x2 is only a few percent faster overall than the original RevoDrive and about 10% faster than a single OCZ Vertex 2. 

In the HDD suite the performance advantage is much greater. The RevoDrive x2 is also a bit quicker than the IBIS, perhaps because of the lack of overhead posed by the HSDL interface. There's a 31% performance improvement over the original RevoDrive. 

SYSMark shows that most CPU bound desktop workloads won't be impacted by the RevoDrive x2. If your applications are IO bound, expect to see an impact but otherwise there's not much to gain here.

Random Read/Write Performance

Our random tests use Iometer to sprinkle random reads/writes across an 8GB space of the entire drive for 3 minutes, somewhat approximating the random workload a high end desktop/workstation would see. 

We present our default results at a queue depth of 3, as well as more stressful results at a queue depth of 32. The latter is necessary to really stress a four-way RAID 0 of SF-1200s, and also quite unrealistic for a desktop (more of a workstation/server workload at this point). 

We also use Iometer's standard pseudo random data for each request as well as fully random data to show the min and max performance for SandForce based drives. The type of performance you'll see really depends on the nature of the data you're writing. 

At best a single RevoDrive x2 (or four SF-1200 drives in RAID-0) can achieve over 500MB/s of 4KB random reads/writes. At worst? 286MB/s of random writes. 

Sequential Read/Write Performance

Our standard sequential tests write 128KB blocks to the drive, with a queue depth of 1, for 60 seconds straight. As was the case above, we present default as well as results with a 32 deep I/O queue. Pseudo random as well as fully random data is used to give us an idea of min and max performance.

The RevoDrive x2, like the IBIS, can read at up to 800MB/s. Write speed is an impressive 677MB/s. That's peak performance - worst case performance is down at 196MB/s for light workloads and 280MB/s for heavy ones. With SandForce so much of your performance is dependent on the type of data you're moving around. Highly compressible data won't find a better drive to live on, but data that's already in reduced form won't move around anywhere near as quickly.

Garbage Collection & The TRIM Bug

Like any RAID array of SSDs, there's currently no way to pass TRIM along to member drives. Once data is written to the RAID array it remains tracked and accounted for until the addresses are overwritten. This is the performance over time degradation issue we talked about at the beginning of our SSD coverage a couple of years ago.

SSD controller manufacturers have avoided the TRIM issue by equipping drives with idle garbage collection. The principle behind idle garbage collection is simple. When the SSD controller detects a period of no activity, it can query the file system for available addresses, and then internally mark those addresses for cleaning.

The RevoDrive x2 features idle garbage collection, which for most users should be sufficient to keep the drive running at full speed. There is a problem however. Current SandForce drives have a bug that impacts TRIM (and idle garbage collection). If you fill all available NAND on a SandForce drive by writing incompressible data to all user accessible LBAs and all spare area, the drive will no longer be able to restore itself to full performance via TRIM or any idle garbage collection. This impacts the RevoDrive x2 as well as standard SandForce SSDs.

I'm still in the early stages of measuring the real world impact of this bug and I made SandForce aware of it two weeks ago. At this point I'd simply caution against operating a SandForce drive near capacity filled with primarily incompressible data (e.g. compressed videos, photos). The problem is even more sensitive on a card like the RevoDrive x2 since there's no way to secure erase the drive to fully recover from a reduced performance state.

Expect to hear more about this in the next major SSD article on AnandTech.

Final Words

The RevoDrive x2, like its predecessor, continues to be a very niche product. Both sequential and random performance are measured in the multiple hundreds of megabytes per second. Like all SandForce drives, the RevoDrive x2's performance is highly dependent on the type of data you're moving. Highly compressible data will see speeds as high as 800MB/s, while incompressible data may behave more like it would on a single SSD from Crucial. The beauty of SandForce's technology is you get great performance on traditional desktop workloads, but as always you need to be mindful of what you're storing on the drive when you decide to go with SF.

The reality for OCZ however is there’s no performance benefit to purchasing a RevoDrive x2 over four Vertex 2s and creating your own RAID array. We’re talking about a simple RAID-0 here, something any performance mainstream motherboard supports right out of the box. To make for an even tougher sell, the RevoDrive x2 will likely cost more than four Vertex 2s. The original Revo was supposed to be slightly more affordable than an array of Vertex 2s, but a quick look at Newegg shows us that you pay a $40 premium for the Revo.

What OCZ (and other companies) ultimately need to do is introduce a SSD controller with a native PCI Express interface (or something else other than SATA). SandForce’s recent SF-2000 announcement showed us that SATA is an interface that simply can’t keep up with SSD controller evolution. At peak read/write speed of 500MB/s, even 6Gbps SATA is barely enough. It took us years to get to 6Gbps SATA, yet in about one year SandForce will have gone from maxing out 3Gbps SATA on sequential reads to nearing the limits of 6Gbps SATA. 

On the surface, OCZ is offering us what we want - a high performance PCIe SSD. However the RevoDrive needs to offer something more than ease of installation. 

All of the major controller players are hard at work on PCIe based SSDs, but I haven’t been made aware of immediate plans to release any of them. The focus continues to be winning 2.5” SATA SSD market share at this point.