OCZ acquired controller maker Indilinx more than a year ago. Since then, it's been tight-lipped about the Indilinx Everest controllers that anchor its Octane and Vertex 4 SSDs. Perhaps now we know why. AnandTech has learned that the controller chips bearing Indilinx's name are, in fact, Marvell hardware. The original Everest controller is a Marvell 88SS9174, which purportedly runs at higher clock speeds than versions of the chip employed by other SSDs. The Everest 2 controller associated with the Vertex 4 is believed to be Marvell's next-gen 88SS9187. OCZ has confirmed that its Octane, Vertex 4, and Petrol SSDs all use Marvell controller silicon.
There's nothing wrong with building an SSD using off-the-shelf controller hardware, and OCZ points out that its proprietary firmware was developed entirely in-house. We were under the impression that Everest and Everest 2 were OCZ hardware, though, and we were wrong. Based on other coverage of Everest around the web, it seems like everyone else in the tech press had a sip of the same Kool-Aid.
OCZ's press releases and briefing materials are carefully worded to refer to Everest as a "controller platform" rather than proprietary silicon. Although there are numerous mentions of Everest as a controller, OCZ tells us those are just short-form references to the overall solution. Apparently, it decided to go with a short-form title for the very press release announcing Everest's arrival, which reads: OCZ Technology unveils Indilinx Everest series solid-state drive controller. My personal favorite is a slide from the Octane's introductory briefing document, which compares the OCZ drive to other SSDs based on the same Marvell silicon.
To be fair, OCZ refers to the controller as a platform earlier in the same document: "Octane is based on the Indilinx Everest controller platform, the follow-up to the renowned Barefoot controller."
OCZ says it didn't see the need to announce the Everest controller's origins in its press releases. Crucial and Intel didn't mention Marvell when they announced their own drives based on the company's controllers, either. However, both were candid about using Marvell controller silicon, a fact that OCZ has certainly obfuscated. When asked whether it considered firmware to be a part of the controller, OCZ told us it views the firmware as "the key component to the total package of what a controller offers." Marvell didn't have a problem with OCZ putting Indilinx's name on the chips, it added.
None of this changes our opinions of the Octane and Vertex 4 SSDs. We'll have to be more careful in parsing OCZ's press materials and more pointed when questioning the company, though. OCZ has confirmed that it's developing proprietary controller hardware, but it's not ready to divulge specifics yet.
|Some 840 EVOs still vulnerable to read speed slowdowns||58|
|Details leak out on AMD's first Zen-based desktop CPUs||78|
|Nvidia: the GeForce GTX 970 works exactly as intended||99|
|Report: 4GB of RAM coming to GTX 960 in March||106|
|Early deal of the week: A 27" G-Sync monitor for $480||39|
|Gearbox's Homeworld remake due February 25||46|
|Nvidia admits, explains GeForce GTX 970 memory allocation issue||242|
|Here's my guest appearance on tonight's Alt+Tab Show||12|