Indilinx Barefoot 3 SSD controller coming soon

OCZ’s "Indilinx Infused" Octane, Vertex 4, and Agility 4 SSDs don’t actually have Indilinx controllers. Instead, they feature a layer of proprietary firmware atop controller silicon from Marvell. OCZ has been working on a new Indilinx controller chip, though. According to CEO Ryan Petersen, this Barefoot 3 controller is now in production at TSMC. The SSD maker expects to be sampling drives based on the chip by September.

Petersen revealed the Barefoot 3’s progress during a conference call with financial analysts, and Xbit Labs has the skinny. Little was revealed about the controller, which reportedly features a 32-bit, 400MHz "Aragon" processor with an SSD-specific RISC instruction set. Interestingly, Petersen said the chip’s "primary IT blocks" are licensed from a third party. The Barefoot 3 may be unique to OCZ, but it’s not an entirely proprietary design.

Naturally, OCZ expects solid-state drives based on the new chip to outperform its current offerings. Petersen mentioned a "significant performance increase over [OCZ’s] current products," although he didn’t provide more specifics. We should have a better idea before long. The first Barefoot 3 drives are scheduled to be available before the third quarter runs out.

We haven’t been treated to a new SSD controller in quite some time, so I’m curious to see how the Barefoot 3 fares against its contemporary rivals. Most of the controllers in modern SSDs have been out for well over a year. It will also be interesting to see what sort of NAND is paired with the controller. We expect SSDs based on 20-nm NAND to be released before the end of the year, although it may be a little early to see one of those coming from OCZ.

Update: Looks like Xbit Labs misquoted OCZ CEO Ryan Petersen regarding the Barefoot 3’s use of third-party "primary IT blocks." Petersen actually said the chip’s "primary IP blocks are not licensed from a third party." You can listen to the full conference call here. Thanks to TR reader Wo for the tip.

Comments closed
    • anotherengineer
    • 7 years ago

    Winter…………….coming soon ;(

      • Firestarter
      • 7 years ago

      Brace you .. ah screw it

    • Derfer
    • 7 years ago

    And this is why they’re not price competitive. They keep forking their product line instead of focusing on lowering their prices through mass production. Crucial is winning over the market with a single line of great drives. A lot of companies could benefit from that approach.

      • chµck
      • 7 years ago

      We don’t want an Apple of SSDs.
      Or do we…

      • Bensam123
      • 7 years ago

      I think SSDs are improving so fast right now simply aiming for the bottom line and scraping every last bit of flab off a product wont yield a satisfactory result for very long.

        • Derfer
        • 7 years ago

        SSD speed hasn’t changed much since the M4 hit. OCZs recent major SSD releases have all hit similar speeds despite having successively newer controllers, and as crucial shows you don’t need to sacrifice performance to have a cost effective monolithic product line. OCZ drives may win in sequential numbers which look nice on spec sheets but the M4 drives compete with (and sometimes beat) OCZ’s premium drives in real world performance for less money.

        I’m sure when competitors consistently hit 600 read/write and over 100k IOPS crucial will release another monolithic line of drives to replace the M4s and be equally successful.

      • Corrado
      • 7 years ago

      They’re not price competitive? I just got a 256GB Vertex4 for $179 @ MicroCenter. The next cheapest 240-256GB SSD was $219. Thats a $40 difference with NO rebate involved.

    • Rza79
    • 7 years ago

    [quote<]Interestingly, Petersen said the chip's "primary IT blocks" are licensed from a third party. The Barefoot 3 may be unique to OCZ, but it's not an entirely proprietary design.[/quote<] Doesn't every SSD controller manufacturer license at least the cpu? Samsung, Indilinx, JMicron, ... use ARM cores.

      • Airmantharp
      • 7 years ago

      Everyone except Intel in their 320/etc. series, I think.

      The CPU has very little to do with the SSD though. It’s just a processor, it’s the firmware that counts.

    • cfroese
    • 7 years ago

    A new controller you say? I think I’m going to wait a while on this. As I recall the last introduction of a new controller was followed by many crashed system. For now, I’m sticking with the tested and fixed controllers.

      • Skid
      • 7 years ago

      A new OCZ controller? I’ll let someone else be the test pilot, thanks.

      • Airmantharp
      • 7 years ago

      The Samsung 830 used a new controller, and has been rock-solid.

      OCZ can take the brunt of the backlash over the Vertex 2 failure epidemic and Vertex 3 BSOD mass-suicide, they deserve that. But this isn’t really OCZ here, rather Indilinx, and Indilinx has an excellent track record.

        • Corrado
        • 7 years ago

        OCZ owns indilinx. They may be fairly hands off, but essentially OCZ = Indilinx.

          • Airmantharp
          • 7 years ago

          Yes, BUT, OCZs biggest problems came from Sandforce. They’re a better company for it, and Indilinx has a proven track record.

    • Firestarter
    • 7 years ago

    Well, colour me sceptical but with the current crop of speedy and reliable drives that routinely max out SATA3, I don’t think that new SSDs are going to be terribly exciting until consumer SSDs move to the PCIe interface.

    edit: Unless of course they manage to easily beat the current SSDs at 4k reads/writes at realistic queue depths, but I’m going to guess here that that is more of a function of the SATA interface and storage drivers than of the controllers themselves.

      • OneArmedScissor
      • 7 years ago

      Whoa there, partner! Routinely maxing out SATA 3? Not even close. Look at TR’s file copy tests:

      [url<]https://techreport.com/articles.x/22794/5[/url<] Lots of drives aren't too far off from HDDs. For the absolute fastest drives, even SATA 2's bandwidth cap is unlikely to be as much of a limitation as the particular controller itself. Some are just slow, but that would not be the case with any modern motherboard.

        • Waco
        • 7 years ago

        The copy tests read and write to the same drive, so you’re looking at a ~250 MB/s cap on performance. for that test. SATA 2 drives perform significantly worse even when they can max out the SATA 2 bus when doing only reads or only writes. At least, that’s what I get from reading the test description.

          • Bensam123
          • 7 years ago

          Each drive is tuned for each version of SATA. There are some SSDs that perform significantly worse in Sata 2 mode compared to others, even though both drives can saturate SATA 2 completely and are SATA 3 compliant.

          One such drive I’m using is the Agility 3, which saturates SATA 2 completely, but performs quite a bit worse then my Samsung 830 using SATA 2.

          It all depends on how much effort the drive makers put into compatability as well as overhead in the spec. I’m going to wager that SATA 3 doesn’t have 200MB/s of overhead though.

        • stdRaichu
        • 7 years ago

        As Waco points out, the speeds are pretty close.

        A 6Gb thoughput of the SATA3 interface (~760MB/s raw capacity under ideal conditions) typically translates to about 500-550MB/s disc throughput, in the same way that 3Gb/s SATA (384MB/s) topped out at 270-380MB/s. The underlying SATA transmission protocol uses 8b/10b encoding (in a nutshell, 2 extra bits for every 8bits you send down the pipe), so theoretical maximum throughput of actual drive data over your SATA link will be just a shade over 600MB/s. From that you’ve also got to take away ATA overhead and filesystem overhead (i.e. the stuff that goes to and fro from your drive but isn’t actually the data you requested), hence why we’re seeing various controllers currently plateauing (perfectly cromulent word) at 500MB/s-ish ; they’re [i<]very[/i<] close to maxing out the interface. Roll on SATA Express.

        • bcronce
        • 7 years ago

        [url<]https://techreport.com/articles.x/22794/3[/url<] Shows read speeds of the 830 at 3.6Gb/s, include some over-head and you're near 4Gb/s. Like Firestarter said, to gain more speed, we really need PCIe interfaces. SATA has way too much latency.

          • OneArmedScissor
          • 7 years ago

          The point was that the average person isn’t waiting around on synthetic benchmarks. :p

          I forgot that the copy test is done on the same drive, but even taking that into account, only the Samsung 830 approaches the limit, but only when copying one ginormous file.

          There’s a ways to go yet for anything else. The last two file copy tests shown are probably closer to what any of us would actually be waiting around on in [b<]real life.[/b<] New controllers could still be faster in countless ways. I'm not disputing that SATA is an issue, though, for latency and other reasons. If I had it my way, even laptops would just wire their storage right into the PCIe bus and we could start to look at throwing out the southbridge on many motherboards.

            • Firestarter
            • 7 years ago

            [quote<]New controllers could still be faster in countless ways. [/quote<] See, I think that's where you're wrong. The controllers themselves can already copy small files at enormous rates, you can see that by checking out the random write benchmarks at extreme queue depths. It's the whole system that sits between the SSD and the high level 'copy' command that is the bottleneck. That includes but is not limited to the SATA interface. The OS is at fault here too, as are the applications. I personally experience that with program/game installers, where I see them chugging away extracting one file at a time, using 1 core to decompress its contents. Considering the merits of SSDs and today's multicore CPUs, that is a wildly inefficient way of installing a program.

      • bcronce
      • 7 years ago

      Your understanding and reasoning is good, but you forgot about low to mid level SSDs. I agree that high end drives may not gain much, but low-mid drive should benefit from faster controllers. Trickle-down affect.

        • Airmantharp
        • 7 years ago

        If low-end drives are to benefit from new controllers, it will be because the controllers make better use of less expensive NAND.

    • Chrispy_
    • 7 years ago

    I’m expecting a hundred anti-OCZ comments in this thread.
    They seem to have annoyed and frustrated people more than the average vendor.

    I’m actually quite excited by this news, having never been hit by the sandforce bluescreen bug that affected all vendors, and my RMA experience with OCZ has always been fast. In two cases I received free upgrades to next-gen controllers. My work laptop is powered by my first ever SSD purchase, a 120GB Indilinx barefoot, and it has held up to modern competition really well for something so old.

      • indeego
      • 7 years ago

      Comment 1/100 can be from me. I am anti OCZ. 😉

      • BestJinjo
      • 7 years ago

      Yup. The Sandforce controller is used in many other drives but OCZ takes the hit for it since they probably sell many times over what Corsair, Mushkin, Patriot, Kingston do. OCZ’s main business now is SSDs. It also happens to be that their SSDs sell for significantly less than most competitors, only contributing to their mainstream popularity. Even if just 3-4% of their SSDs failed, that would be probably 100x as many complaints as say Plextor’s SSD based on the same controller.

      It’s good to see constant innovation in the SSD space. Unlike mechanical drives, it’s actually interesting to follow the advance in SSD space since they continue to happen at such a rapid pace, although the SATAIII connection is already bottlenecking the drives in some instances.

        • eitje
        • 7 years ago

        Yknow, a 3-4% failure rate means one in every thirty customers has a failure. That’s huge.

        • BobbinThreadbare
        • 7 years ago

        It doesn’t help when you treat your complaining customers like shit and lie and say there are no problems.

          • Airmantharp
          • 7 years ago

          We can all understand how OCZ may have felt- not to excuse their actions- when droves of customers started complaining about issues surrounding their drives. User error is still highly likely with new products, and I’d expect any company to push back when feeling that much pressure.

          Still, they violated the ‘customer is always right’ rule of sales, so let’s hope that they have learned from their mistakes. They continue to make bleeding-edge drives and sell them at very competitive prices, and I wouldn’t like to see a market player like that lost.

            • BobbinThreadbare
            • 7 years ago

            They also lied about it being a widespread problem.

            • Airmantharp
            • 7 years ago

            I have no doubt that they misrepresented the truth and I’m not advocating leniency in the slightest- just trying to get people away from demonizing the company and move on. They have other products, have serviced their customers, and are pushing the industry hard while competing against behemoths like Intel and Samsung. They deserve a little credit :).

            I had a Vertex 2 flat out die on me too, y’all, and I purchased an M4 rather than a Vertex/Agility 3 for my laptop. The Vertex 2 has since been replaced, and very quickly I might add. We’ll see how it holds up!

      • wierdo
      • 7 years ago

      Well, I’ve used many OCZ products and I have to say they remind me allot of Via for chipsets: Affordable but finicky products with mediocre customer support.

      If you wanna save a few bucks, OCZ could do that for you, but for stuff like SSDs and PSUs I’d rather pay a little more for peace of mind – the new Samsung SSDs seem to be looking great right now for example.

      • xand
      • 7 years ago

      Hmm. Unless you buy SSDs for your company or something, having two storage failures from a single supplier is… alarming :p

        • Chrispy_
        • 7 years ago

        Yeah, I do 🙂
        My sample size is still pretty insignificant but I reckon I fitted my 100th SSD at some point this year. I reckon I’ve had to RMA about a dozen of them (so the 3-4% failure rate is a little optimistic in my mind)

        OCZ, Crucial, Corsair, Patriot, Sandisk, Kingston. The only brand (that I’ve bought) which is RMA free is Intel, which is a shame because they’re typically slower than the competition and as much as 50% more expensive per GB.

        Most of the failures were 1st Gen Sandforce, and most of the RMA’s were good. OCZ probably handled the RMA process best, Corsair and Kingston were pretty good too. I’m not buying another product from Patriot or Sandisk without good reason in the future.

        I’ve said it on these forums before, that OCZ may have been economical with the truth, but find me a company in today’s market that isn’t. My issue is the fact that idiots were moaning about OCZ because of information they gleaned from the OCZ [i<]forums[/i<]. At no point have I ever gone to a forum for an RMA, I follow the RMA process. Of course the OCZ staff on the forums are going to be suggesting alternatives to an RMA, and if people want to pretend that using the forums is part of the RMA process, then that's their own stupid fault. Most SSD vendors don't even have user forums....

          • indeego
          • 7 years ago

          OCZ linked to the forums for firmware support ( which is just the crappiest design for formware updates, aesthetically and technically.) Firmware support is mandatory to fix their BSOD issues. How would you find out about how to use/fix a product if not for their or other forums? Hell I use HP’s forums to find out any caveats about various software packages for laptops, printers, servers, desktops, etc. It seems like a good way to determine what other users’ experiences are.

          As for RMA process, OCZ refused to cross ship me a SSD, it took a week without a SSD to get back (one day response, one day to prep the shipment, two to process RMA, One to ship, weekend added two days)

          And then it happened again to the RMA’d drive. Fool me once…

          Yeah so never again. I’ve only had this happen one other time, and it was a mechanic that kept on replacing a part over and over again. By the fourth time I just bailed and went to another mechanic and it worked perfectly ever since.

      • Bensam123
      • 7 years ago

      Yeah, OCZ has retroactive warranties. In other words if there is a newer version of a product available and yours is no longer sold, they upgrade you. It happened with both my power supplies. They also rewarranty the new product for the full duration.

      They’re never going to escape the BSoD bug BS though. That affected other drive makers, theirs just got hit harder (maybe due to more of them being on the market, maybe also in part due to firmware). Either way, they fixed it.

Pin It on Pinterest

Share This