Gadgetory

hey everyone this is steve from gamers nexus dotnet and today we're benchmarking kingston's v300 against itself there's been a bit of a stir in the tech community lately after reports a few months ago that users were seen significantly lower benchmark scores with their ssdnow v300 than the ones reviewed by the press and and some of the initial models of course it was largely understood that the performance difference is a result of Kingston silently switching to microns 20 nanometer asynchronous NAND from their original models 19 nanometer toggle mode synchronous mend from Toshiba the important part being the synchronous bit vs. asynchronous so just to be clear the current model is asynchronous and the model that is no longer shipping that was originally tested was using synchronous NAND I wasn't satisfied with the user test out there since they primarily rely on incompressible benchmarks like a SSD and and so I built a trace based real-world test to accompany a suite of synthetic benchmarks to test the claims that the performance dropped fol disclaimer this video is going to be a bit shorter than normally because I have a flight in a couple of hours for GDC and gtc so my apologies hit the link in the description below for the full article that goes in full detail on what all this synchronous where it's asynchronous stuff means the objective of this experiment was to investigate the performance differences between asynchronous and synchronous and Nan as it pertains to Kingston's v300 SSD the switch happened entirely for cost and supply reasons and is the sole reason that the v300 is currently available for about seventy to eighty dollars these days for a 120 gigabyte model anyway let's briefly talk about the two types of NAND used the difference between synchronous and asynchronous an end largely stems from the way the timing circuit prompts data transmission in this graphic by Samsung we can see that the timing circuit used for this particular asynchronous nanded design is prompting a read event that's re signal on the fall of each signal wave the synchronous NAND ever sends its signal on the rise and fall of each signal wave so about two times as often and produces almost a 2x difference on paper so with SSDs on paper is very rarely translated to real world because the way controllers work and operating systems work and all the overhead in between but that's what we're testing this is sort of like the difference between an asynchronous stopwatch that you can stop with a button and asynchronous or continuous wall clock in theory asynchronous NAND should see the biggest drawbacks and sequential and incompressible actions you can read about the methodology below but we're just going to dive straight into benchmarks because I want borrowed time for that GDC flight so the devices were preconditioned for each type of task before testing initiated first of all starting with my in-house test suite we can see that the windows 7 boot time differences we're totally unnoticeable and within margin of error you can ignore them the winrar extraction and compression differences were pretty note as well fifteen percent in the case of compression and that's going to be the very few users will do a lot of compression or extraction on a daily basis if you do you know who you are great for the handbrake transcoding process we saw absolutely zero difference primarily because this test is more bound by the cpu and ram than storage for this test i used largely incompressible data that was ten point 3 gigabytes in total size the data consisted of several hundred songs and a couple of large video files all in mp3 mp4 and avi formats because this data is already heavily compressed by the nature of the formats the SSD considers it to be largely incompressible data and will thus perform slower by the nature of the sandforce controller we see a massive 51.7 percent difference between the old v300 that's no longer available and the current v300 so that's pretty that's pretty big you're going to notice that if you're moving a lot of files around but as for how many users move enough files to notice on a regular basis I will talk about that in a moment for premier encoding there's about a six percent delta which is sizable for anyone spending 20 hours a day rendering but outside of that it's really not noticeable and moving on to a forty six percent incompressible benchmark or rather forty-six percent compressed data benchmark using anvils storage utilities we see that asynchronous the asynchronous be 300 that I'm calling to be 300 a suffers hard in sequential operations and operations of a high-q depth but seems to do fine and lo q depth operations so that's sort of expected now as a user you're going to be mostly using 4k random operations and even 1k random operations almost entirely at a queue depth of one or two for most actions and you'll occasionally peak at four depending on what you're doing most games that I've done trace testing in our 1k queue depth one so that's kind of what you're interested in if you're gaming but keep in mind the performance is already so high on both of these drives for things like gaming you're really not going to notice a difference in your load time or anything like that it's Italy fractions of a second for most gains so do keep that in mind finally we have some iometer tests I have these full tests on the website of course and it's really just showing more of the same for anvil it's showing our one KQ depth one bench that I talked about and again we see huge differences with the higher queue depth but for for the lower stuff q1 q2 at four it's somewhat unnoticeable for the average user so the conclusion here or the question rather for the conclusion is is the new v300 actually worth buying with all of these changes is the asynchronous land of so detrimental that you should no longer be buying it the short answer as always is it depends if you are a power user if you're doing compression and file transfers and rendering and stuff that I do for the website like all these videos you should absolutely care because those that fifty percent difference in a file transfer for large media files when you're dealing with a hundred gigabytes of media files that's a big deal and the same goes for our rendering a six percent Delta for one video a day is not going to be noticeable when you're rendering 20 hours a day that can be huge that can be an extra video you can put out that day so that's sort of where you want to go towards Kingston's own a hyperx actually which is within spitting distance of the v300 it's ten five or ten dollars more and it uses synchronous and and it's a bit bit faster and by a bit I mean a lot faster so that's that's sort of where I would push most power users and users who can identify with my own usage patterns because for ten dollars you get something that will make you a bit happier when you're producing content for the average user who's trying to find an SSD to stick in their laptop to boot a little faster and the load applications a little faster you're probably going to be fine with the p300 the other option right now I guess is the m500 is about the same price I haven't tested that one so I can't recommend it but at the seventy or eighty dollar price point of a 120 gigabyte we hunted it's perfectly fine for everyday actions you're not going to notice those differences all that much and for the most part you're going to be doing compressible file transactions because that's what your OS does that's what all your log files do that's what your browser does when it's caching data that's what a lot of saves and other document based applications will do gaming does deal with a lot of incompressible stuff and media is a lot of incompressible stuff but you're still sitting in the forty-six percent compressible data range so my conclusion I guess then is that if you are a main stream user who's sticking this in a laptop or a desktop that's doing office applications you shouldn't get too hung up on this if you are a power user you should probably care and do more research if you're just gaming will say quote just a gamer and you're not doing a lot of content creation then once again not a huge impact you shouldn't get too hung up on this but for ten dollars more I would push you all day long toward the HyperX except if you're on an ultra budget if you're building one of our cheap bastard systems you might as well just save the ten bucks I guess and get the v300 if that so pleases you if you can't fit a ten dollar more drive in your budget which is totally reasonable so that's where I stand on it please check the article for full details and i will see you all next time peace