Gadgetory

our recent coverage of Intel's DCP 4800 X data center obtain Drive it served as a primer to 3d crosspoint technology the architecture jointly created by Intel and micron intel's deployment of this architectures house under the octane name while microns will be quantex in an article we published a few days ago we revealed that Intel has plans for octane as a consumer targeted device for release on April 24th of this year that's a big change from the $1500 data center drive we previously looked at we're back to discuss the potential challenges that octane memory faces its target market and messaging that needs to get across on Intel's part before we get into that this coverage is brought to you by EVGA 1080p IFT w3 video card with the new icx thermistor technology and if you are curious about that card you can learn more at the link in the description below the basics of 3d crosspoint don't change your intel octane memory is still using cross points and the same architecture underlined as a DCP 4100 X the difference is in the implementation with octane memory Intel's looking at more of a consumer focus whereas the P 4800 X was data center focused although you could technically put it in a consumer system if you wanted to the thing here is that Intel is still attempting to fill its unique gap between memory and storage solutions and is trying to do so by offering a low latency high endurance type of memory that effectively serves as they drive cache and you would use something like rst to enable it as a caching device this is a story that we've all heard before many of you may remember the earlier SSDs that were too small to really use as a proper boot drive we're being used as cache drives that idea while interesting did largely fail and was replaced by proper larger sized SSDs because it just made a lot more sense in terms of complexity price usability all that stuff as they became more popular that's what changed the price of SSDs fell and mostly eliminated the need for any type of caching solution today Intel is trying to bring back a type of cache though it is a lot lower latency it is theoretically higher endurance and there are all kinds of claims they've made that make it look pretty good it's just a matter of is there a use case for this Intel obtain memory technology so it's a story we've heard and for most of our audience it seems that solid-state drives plus hard drives are sort of the primary use case where you're buying likely one of each and using the SSD of the primary drive with your favorite applications on it the hard drive as archival storage or game storage or whatever that seems to be about the main use case Intel is looking at targeting more of a hard drive only use case for the octane memory implantation something we'll talk about here and in some regard this is reminiscent of those earlier cache drives and of solid-state hard drives or SSH DS where they had a bit of SSD on there and a bit of hard drive on there neither has been particularly widely accepted by the market so Intel's got a long road ahead for octane memory the question is where this technology has used cases and if it fills a market gap that is already unserviceable 'ti octane memory is fitted on an m2 form factors stick and although several generations of Intel platforms host m2 sockets only KB Lake will officially support octane memory and the rst update Intel notes that this is for validation reasons though the technology is mostly software driven and just needs an m2 slot on the hardware side there's nothing special in KB Lake that skylight does not also possess with regard to opt in memory support on the hardware level aside from additional testing and validation done on KB Lake that probably for time reasons wasn't done on skylake or older platforms the m2 memory stick acts as something of a front loader for the system drives Intel markets this as a $44 stick as targeted at hard drive users rather than SSD users and thus only presented significant data in a hard drive versus opt in cast hard drive use cases and we have plans to perform SSD cache use case testing as well to determine if octane can offer use for SSD plus hard drive configurations as most of our viewers and readers are likely running either way when pre buffering the hard drives the performance gains are tremendous so it's a technology worth looking at opt in memory can be configured through rst as a 16 gigabyte or 32 front buffer to primary storage and Intel says that the device will quote learn user behavior to determine which files get cached on the stick windows files will make up a lot of the file composition on the stick with frequently accessed applications art files or games taking the rest in some brief testing we performed at the press event we saw blender and photo editing applications like again reduce their launch times from roughly eight and a half seconds to roughly four seconds after multiple test passes were averaged the first test passed with a hard drive made it look a bit worse but it did improve over time though nowhere near the extent that octane improved as it learns the application will not be publishing any hard data until we can further validate in our review as this was a limited run test in Intel's facility but it has some promise despite being called memory Intel octane memory is actually a non-volatile solution so this is good for a few reasons one as we discussed in the data center coverage the idea of octane and 3d crosspoint is to provide a higher capacity lower cost somewhat partner to ram as opposed to an SSD which is still significantly worse in Layton sees compared to RAM but filling somewhere of a middle market so that means that middle devices always have a weird challenge just like we saw with SSH DS and cache drives where they might not see market adoption in some segments this one octane storage looks like it's got a pretty serious future potentially in data center enterprise things like that octane memory is the consumer targeted alternative and that's the front buffer and then there's also octane storage eventually for consumer that we don't have any details on it yet now as far as what Intel is advertising because we haven't tested ourselves yet so we're going off of their numbers for now and then we'll test and check back for those results Intel claims a 2x increase in overall responsiveness that is their word for basically meaning how Windows fields so if for Intel slides and for the purposes of this discussion when they say responsiveness what they mean is sort of what the user experiences when they're launching things like applications or Windows Explorer things like that and the point here that they were making is that for consumers who are not enthusiast audiences like us or like you all for the consumers who are less informed responsiveness on Windows with a hard drive doesn't feel great and there is merit to that that's the discussion we had when solid-state drives became a thing so there's merit to that discussion now as far as measuring it we'll do that in our own testing if it's even possible but other claims that were made were 4.1 times faster media project launches so that would be things like large blender files or large project files and CAD stuff like that and in some respect you can think of the Radeon SSD which has that really large solid-state cache on the video card it's kind of comparable to that in ways where if you're accessing a huge CAD file regularly it might get stored in this opt in memory because it's non-volatile it'll stay there and then every time you access it hopefully it should be faster up to an extent where it's not possible to continue accelerating it beyond that point so all the tests that they showed us and the claims they made forthwith with one exception I believe were versus a hard drive a single Drive solution spinning disk not versus solid-state drives so obviously the reason you do that is because if you're Intel and you say well octane front-loading an SSD or whatever shows some improvement but nothing like 4.1 X which is a pretty large number so we're not going to show we're going to show the less favorable or less interesting or less sensational scenario that you would get with a hard drive that's how I read the situation so unfortunately we don't know how or if opt in memory helps in a significant way with the solid-state drive plus hard drive solution that most of us are likely running but front-loading hard drives it looks pretty good they also showed some other tests with different applications Windows boot time stuff like that power draw should be about the same as an SSD of similar deployment and spec so power draw is not a big deal there and shouldn't impact laptop users too much as for support immediately until notes that 130 plus motherboards will officially support octane at launch including asu's EVGA gigabyte MSI as rock and then other vendors will support following launch and that would include system integrators power saira power all the usual suspects in that department messaging is Intel's biggest challenge with octane memory for enthusiasts we've been told before that front-loading your drive will speed it up but it's never really been quite worth the effort obtain memory completely react acts at a hardware level though is not too different to the end-user in terms of what it's actually doing you keep in mind too that octane memory is just one of several implementations there will again likely be AIC octane storage devices at some point this year or next that tends to what happens with these things those might be more interesting but we'll see anyway for messaging challenges Intel informed us that octane memory is it more to help users who feel Windows is too slow and it's normally because the hard drive is the weak link not the OS itself so we then asked Intel the obvious question this seems like a problem that already has a solution those would be SSDs so why do you feel obtain is different what makes it seem like this technology is going to succeed in a different capacity or better capacity than solid-state drives have Intel's main argument here was very non enthusiast focused so they were looking at audiences like the person who walks into Best Buy or other local retail chain and purchases a complete system for example the $500 Dell box that's sitting at Best Buy next to a $500 HP box these two devices have one thing in common they probably have hard drives in them and the consumer is likely buying it because it is $500 and at the store if they're not informed and they sit there and they use one then they use the other then they use a $2,000 system from I buy or cyber or whoever they all feel the same in terms of Windows usability there is no applications no bloat loaded on these things they're in the best light so they probably feel the same the user who is uninformed therefore buys a $500 box now that argument is valid in some ways anyway and Intel is basically saying look we're going to take those boxes put off a memory in it it'll be a $44 retail stick we have no idea what the market will be for the sis probably not a whole lot so $44 for stick and it will perform a whole lot better okay so that's the argue is but SSDs exist so what else what else can you give me if I can still spend 80 bucks and put it pretty damn good SSD in there without much markup from these at sighs why is this that much better when it's not that much cheaper their answer was well SSDs are hard to use so that might be laughable to us I think actually probably is a bit laughable but the thing is again with that non enthusiasts audience they're basically looking at as this user doesn't know what an SSD is doesn't know where to put their files doesn't know that putting stuff on the SSD with limited capacity should really be reserved for special types of files that are accessed frequently are really critical to operation whatever so they're saying opt in memory solves this because you only see one device the hard drive and then op same memory and rst work out the rest on its own they figure out what's what the user needs okay so that's that's the argument now the thing is even if we follow that argument and ignore the whole enthusiasts market just look at the consumer market the first thing that comes to mind is the difference between Microsoft Windows and Apple's arises is that Microsoft Windows runs on an open ecosystem Dell HP CyberPower ibuypower main gear all of them they all have their own initiatives to build systems to aspect that they like that they think competes they're not restricted by a company like Apple where Apple says you know what we want an SSD in this entire product line you don't get anything else so that means what happens of course is there's a race to the bottom Dell and HP might have two boxes on the shelf HP's $544 got Intel obtain memory Dells $500 the end user is still going to buy the $500 box if the argument to begin with was they buy the cheapest box on the shelf so until it's got a problem there now MDF can certainly help solve that and pushing the manufacturers and the S eyes and integrators to use off paying for those low-end devices that can do a whole lot of good depending on how much they're willing to spend on marketing and MDF for these eyes so there's potential there now ultimately we're seeing a race to the bottom no matter what that doesn't mean I obtained useless obtain certainly looks very promising based on the testing we've done versus hard drives at the Intel event lot of caveats there and requirements but it looks promising it's just a matter of messaging that is Intel salute biggest challenge it's not the technology which seems good it's the how do we explain this to people who want to know what it is people like us and how do we make sure the people who are buying in bulk are still not just going for the cheapest thing on the shelf so there's a lot of testing to do here the biggest one is going to be if we use opt in with a traditional SSD how much does it help us because the latencies are still a lot lower than traditional SSDs so that's promising depending on what you're doing and we'll have to test that other than that it's it's really just a matter of seeing what happens we should be getting samples to test will be releasing hopefully reviews by launch I think that was April 24th so launching April 24th pre-orders are up now don't preorder it just wait and see what people say about it because it may or may not make you sense for your use case as an enthusiast we'll have to see if you're using a hard drive it sounds good but again if you know how to use a solid-state drive it's kind of like should you just buy a solid-state drive instead for 80 to 90 dollars and just use that so we'll find out for you we'll let you know if front buffering front loading whatever helps out with those use cases or not for now links in the description below for more information we have an article on this from a few days ago we have an article on the DCP 4800 X along with a video that explains the architecture a bit more then we have an architecture piece on 3d crosspoint which is Intel micron co-developed so you can learn more in all of those links thank you for watching subscribe for more I'll see you all next time you