We've known for some time that there would be two versions of SandForce's upcoming SSD controller: the SF-1200 and the SF-1500. Only the latter is compatible with the SLC memory typically found in enterprise-grade SSDs. As one might expect, the SF-1500 is also a little faster. The chip is rated for 30,000 4KB random-write IOPS, while the SF-1200 tops out at only 10,000. Other performance specifications, including random writes and sequential peaks, are identical between the two controllers.
Interestingly, AnandTech has learned that the SF-1200 and SF-1500 share the very same silicon. The SF-1200's random-write performance isn't capped by the chip itself, but by the SSD's firmware—or, rather, by one specific version of the SF-1200's firmware: the 3.0.5 release intended for mass production.
According to AnandTech, OCZ's close working relationship with SandForce has earned the company exclusive access to a faster firmware revision for the SF-1200 that offers all of the random-write performance of the SF-1500. Fair enough. However, SandForce distributed a 3.0.1 release-candidate firmware for the SF-1200, which doesn't have the 10k random-write IOPS limitation in place, to all of its partners. One of those partners, Corsair, has begun selling SF-1200-based Force solid-state drives using the 3.0.1 firmware.
Anand explains a series of potential concerns with the firmware imbroglio: early reviews that show higher performance from SF-1200-based SSDs than drives based on production firmware will deliver, a possible reliability issue with the 3.0.1 firmware, and the potential for end users to see performance drop after an update.
This episode reinforces the fact that SSD firmware can have a dramatic impact on performance. We publish SSD firmware version numbers in all our reviews for a reason. Prospective SSD buyers would do well to stay informed about which firmware revisions are classified as ready for mass production and which offer better—if unauthorized—performance.