Gadgetory

welcome back to how we run box today I'm gonna try something new the plan is to review feedback given by you guys the viewers on a recent video and as a guinea pig episode 2 what could potentially become a semi-regular series I've chosen the how much RAM to gamers knee video from a little over a week ago so the idea isn't to call people out or anything like that although I'm sure there will be the odd funny comment here and there that we can address but anyway the idea is to follow up on your feedback was there something that I missed something that I got wrong or just something that you would have liked to have seen me cover in a bit more detail now since this is a guinea pig episode please let me know if you found it enjoyable interesting yeah just do that in the comment section below and of course if you thought that it sucked and you don't want to see any more replying to comments type videos or episodes then please also let me know about that in the comment section below today's video sponsor is Skillshare the online learning community where you can learn real-world skills I've always found it easiest to learn by watching others and with skill share you can do exactly that as professionals correct project-based classes covering topics such as design business technology film photography and much more it's a great resource for those looking to improve their skill set a sure passion or advance their careers and the best bit is for the first 500 viewers to sign up you will get the first two months of skill share premium for free so it doesn't get much better than that check the link in the video description ok so let's get into it Johnny X Y Z pointed out something interesting that I only touched on briefly in the video not exactly the same thing that he's talking about here but kind of similar anyway what he said was the really interesting thing is that you're able to buy 578 gear but instead of the 4 gigabyte version for the money of a caballo ddr4 so I think he's trying to say the cost of 8 gigabytes of ddr4 memory they're basically the 8 gigabyte version of the rx 570 costs $30 more than the 4 gigabyte version while 16 gigabytes of ddr4 costs $40 more than 8 gigabytes so it's not quite the total cost of 8 gigabytes worth of RAM it's only about half as much still if you had to compromise on one it would actually be better to do so with the RAM so instead of getting 16 gigabytes of RAM just get 8 gigabytes and that is because the full gigabyte RX 570 really does need 16 gigabytes of RAM to avoid stuttering and a lot of modern titles so if you can get the 8 gigabyte version it's probably going to give you a better gaming experience that is assuring me that you using ultra quality settings at well at least 1080p maybe 1440p it's probably stretching it a little bit there either way that on that scenario there is just a $10 difference depending on which compromise you make of course with 8 gigabytes of ddr4 3000 memory price that's just $60 and 16 gigabytes and $100 I suggest spending the extra $40 to ensure all games run as smoothly as possible same goes for the graphics card if that gear by version was $50 more say I'd probably recommend the full gigabyte model but for just a $30 premium it makes sense to get the 8 gigabytes of vram especially if you're not on an extreme budget for those of you who are on an extreme budget for gigabytes of vram and 8 gigabytes of RAM will do and particularly if you're playing less demanding titles such as fortnight for example just lastly on this one a much better or much worse example depending on how you look at it would be the GeForce GTX 1063 gigabyte for those with just 8 gigabytes of system memory the 6 year back version of the GTX 1060 provides significantly better performance and that means saving the 50 something dollars when they're priced at the MSRP for the 3 gigabyte model that really isn't worth it as you need to spend $40 more on memory to avoid the noticeable stuttering in a lot of the modern titles even at 1080p rumor has it Nvidia is releasing a 3 gigabyte version of the r-tx 2060 in 2019 if that is the case I would urge reviewers to test it with 8 gigabytes of RAM at least for a couple of tests just to see how much different it is compared to what I believe will be the fully fledged 6 gigabyte version and yet I'll be doing a few benchmarks like that in my day 1 coverage ok so next up we have ng Russ ingress ingress commented I don't know all just sorry if I don't get all right anyway the comment was something along the lines of they can no longer live without gigabytes of RAM not because they need it for gaming because they want to keep 30 browser tabs open when playing games I saw quite a few comments like this and it's fair enough that said if you're jumping into battlefield 5 and you're not constantly out tapping out of the game then it doesn't really matter this is because the browser tab data is relocated from the fast access memory to local storage as you load the game so in other words it's moved to the page file so as long as the required game assets can be stored in this system memory you won't suffer any slowdowns or frame time issues if you all tap out of a game between rounds - I don't know watch a youtube video check your email or whatever reason you might have out of a game between rounds or you have a second screen which is probably a worst case scenario with a YouTube video active while you're playing I know that sounds kind of odd but a lot of you guys in the comments section claimed that you watch YouTube videos while playing games so that's impressive I'm usually pretty flat out just concentrating on the game but whatever the case may be this is where more memory will really have an impact because you're running a lot more high priority processors so if you don't have enough memory when you alt-tab so forget about the second monitor for a moment when you alt tab out of the game the system will be very sluggish and somewhat unresponsive while it reshuffles things in and out of the memory so we alt tab it'll prioritize your browser again because you're using that and then when you jump back in the game it'll be much slower to do that and in extreme cases the game might crash but usually it's just quite slow so you'll be presented with a black screen for a longer period of time than you would be otherwise with enough memory and then you can get on with the gaming and even then the gaming will probably be a bit choppy for the first 30 to 60 seconds a user z4 v3k came at the internet browser thing from a bit of a different angle claiming that the conditions that I tested with were unrealistic as gamers close all that stuff before gaming and I suppose again well fair enough but also again it really makes no difference the results would have been the same if I test it with a skeleton crew of processes on a fresh install of windows or if I test it the way I did with a few applications open such as steam origin you play Chrome and so on and that 3 gigabytes of RAM that's claimed to be sucked up by these unnecessary applications it really isn't once you execute your game again all that stuff is moved to the slower local storage into your page file and that makes way for the higher priority data in this case the game that you intend on playing actually saw a memory capacity video done by Linus tech tips earlier in the year and they found a zero difference between a clean install and then one with loads of applications running which included chrome tabs and all that sort of stuff when they were gaming so well I get where you're coming at and I agree with some of the points you made I do strongly disagree that the video was misleading or unrealistic in any way because of the reasons that I just described if I had to disabled the windows page file then yeah you might have a point did you turn off windows page file else all those background apps will just go into page file right you are sir right you are on that note you might think why the hell did I bother running all those applications if it makes no difference at all as I've been saying well the reason is you guys a few years ago I did a memory capacity test it wasn't the first one I did but it was one of the one of the originals and I had a whole lot of people complaining that the test was unrealistic and that US benchmark because they're out of touch with the common gamer and yadda yadda yadda because I was running just the game and well the bare essentials no chrome tabs or anything else in the background at the time again I knew it didn't make that much difference I knew that stuff would get moved to the page file and systems without enough memory we had that upset a lot more people than this method did last year that I used so I'm sticking with it but that's the reason why yeah it doesn't really make much difference either way but I feel like more people probably aren't that keen to shut down everything they've got open and then run their game it'll be just more convenient if he was Randy game and then go back to whatever it was that you were doing before you played the game ok so despite the testing done in the previous video the video we're discussing quite a few viewers agree with what mr. freaked out to say a gigabytes of memory will I should do fine and 16 gigabytes is more than enough mr. Frick did say that it's important not to run memory in a single-channel mode and well I couldn't agree more with that as for the comments on memory capacity again most titles I will work just fine with a gigabytes of RAM and we discussed what those tiles would be in the in the video in question but there are a number of modern triple-a titles where you will see frames stuttering if you only have 8 gigabytes of memory and we saw a 16 gigabytes all that sort of studdering I did go away so it's in my opinion if you're playing modern games 16 gigabytes really should be the minimal configuration that you want to use there and that's very true for titles such as battlefield 5 or shadow of the Tomb Raider for example so nothing really wrong with this comment just don't assume that a gigabyte should do the job for modern tiles because for some of them it's simply not enough I noticed that quite a few people requested that I look into single versus dual channel memory performance for games and a lot of the requests wanted me to look at both AMD and Intel CPUs I really like this idea so that's something I'll make happen probably in January and I'll cover as many hardware configurations as well humanly possible the outcome will definitely be always running dual channel so there'll be no big surprises there but it would be nice to see where dual channel has a real advantage over a single channel so where it makes the biggest difference what sort of games what sort of settings and that kind of stuff so so yeah well I'll do that one soon here we go ah finally a serious comment from none other than jerod of Jarrod's Tech and he hit us with a classic one-liner this was the most up voted comment on my video that took me weeks of work the human eye can't see more than 9.3 gigabytes of memory classic Jared a few guys aren't familiar with Jared and while he's channel Jarrod's tech be sure to check that out I'll put a link in the video description Jared I'll probably comment on this video as well and it'll probably be the most uploaded comment for reasons unknown but you guys check out his channel he does a great job of checking out a whole lot of tech he mostly focuses on gaming laptops some really good top coverage there so if you're interested in gaming laptops then yeah you definitely want to be subscribed to Jarrod and if I haven't settled already here's a fellow as a youtuber okay so I'm not going to read out any individual comments here because there was a whole barrage of them so yeah I'm just going to get into it with this one twelve gigabytes of RAM now last year's memory capacity video was littered with comments asking why I didn't test with the best bang for your buck memory configuration twelve gigabytes now I didn't address the twelve gigabyte capacity thing in the 2018 version I foolishly just assumed that most had cottoned on by now but yeah I was wrong on that one so yeah once again the most heavily commented subject was twelve gigabytes of RAM why didn't I test with it it's the best 16 gigabytes as overkill get twelve so on and so forth in short you can't or well I shouldn't say can't in short you really shouldn't use twelve gigabytes of memory on a platform using ddr4 memory firstly you can't buy and have never been out to buy two gigabyte ddr4 memory modules they simply don't exist so you can't have two four gigabyte modules with two two gigabyte modules for a dual channel twelve gigabyte configuration that was certainly possible with ddr3 but not with ddr4 I should also note that triple-channel platforms using three four gigabyte modules also commonly had a twelve gigabyte capacity but that's not possible on dual channel systems unless you want to cripple memory performance but what about a 4 gigabyte module with an 8 gigabyte module that's certainly 12 gigabytes well you can do that but again I strongly suggest you don't for Rison owners this will seriously cripple performance but even those using an Intel platform are you will be throwing away quite a bit of performance Intel offers a flex mode that allows a for an accurate module for example to still run a dual channel basically you get 4 gigabytes from each module running a dual channel with the odd 4 gigabytes left over running in the single channel mode now admittedly this is better than single channel operation but it's nowhere near as good as true 4 gigabyte or 2 egg you bought modules in dual channel in other words 2 evenly-matched sticks for example a single 8 gigabyte ddr4 3,200 module paired with the 99 rekha has a read throughput of around 24 gigabytes per second then pairing a 4 gigabyte module with an 8 gigabyte module boost that to around 34 gigabytes per second so that's not too bad but to evenly-matched modules so to 4 gigabyte or to a gigabyte modules will allow a read throughput a 46 gigabytes per second that's an additional 35 percent memory bandwidth and that doesn't go unnoticed actually depending on the software you use the bandwidth discrepancy can be much larger than that I'd a 64 for example measures the peak bandwidth and that was the software I used to quote all the numbers that I just did however I usually prefer to use size soft sandra's sustained memory bandwidth test and here the matched modules offered 50% more bandwidth in short performance and battlefield 5 was much worse with 12 gigabytes of memory using the 4 plus 8 configuration compared to 2 full gigabyte modules for an 8 gigabyte capacity the 0.1% low frame time performance improved by 13% with the matched 4 gigabyte modules so for the most part gamers are better off sacrificing memory capacity for bandwidth when it comes to comparing the 8 and 12 gigabyte configurations it's probably a bit more to it why the performance is a bit lopsided there with the 12 gigabyte configuration but yeah I did see similar things and shut off the Tomb Raider and hitman 2 and not as Extreme as what was seen in battlefield 5 but still the 8 gigabyte configuration did allow for slightly better slightly more consistent frame rates ok cube 1 V X asked a pretty good question here about benchmarking or testing load of screen times a few people asked this few different people and no other questions or comments with that heavily uploaded by I suppose of the rest of you but here there was enough people asking why I didn't test this so yeah I thought I would answer that one so basically the question is why in my memory capacity benchmark video I didn't measure the time it takes to load games and I suppose load levels with the different memory capacities and memory capacity will certainly speed up things here especially for those with applications open in the background that they don't want to close when they play games again this is because all that stuff will be moved from the system memory to your local storage so the lower priority stuff will be simply moved to the page file and this means as you load into games there's a bit of shuffling going on and yet slows down the process the reason I haven't done this testing is because while memory does have an impact are things such as local storage and even the processor impact load times far more still this is something that I can look into in future versions especially if a lot more of you want to see this happen so you start commenting you know benchmark load times and low-level stuff with memory capacity see what impact that has then yeah that's something I could do in next year's video and I think that would be something I'd really like to look into if 2019 is another year where 16 gigabytes of RAM is well it's the sweet spot 32 gigabytes is overkill and well it's the same as what we found last year and this year so if that's the case then you're testing something different might shed some more light on the subject okay so last comment or at least the last one that I'm going to address there was well over a thousand comments but I picked the more interesting ones all the ones that were more heavily upvoted like Jared's the I can't see 9.3 gigabytes of RAM whatever that hilarious comment was anyway moving on we have one from soya I can you test Ram speed for render applications like blender yeah in short yes I can and I actually am at the moment so I suppose that's going to make soya quite happy I am only doing it for the 9900 Kay so no AMD testing at this point we've done a bit of AMD memory stuff when the second gen Rison processors came out anyway and yeah it's a lot of work doing memory scaling benchmarking testing at least I think it's at least eight different memory frequencies and I'm testing quite a few applications and a few games I'm also doing some game resolution testing a show where under realistic conditions memory frequency may not impact performance as much as what we see in a lot of the benchmarks focusing on 1080 or lower but anyway that should be a very interesting video and one worth keeping an eye out for well that is just about going to do it for this first episode of replying to comments again please let me know if you enjoyed this video and if you have any suggestions for this well I suppose this potentially new series yeah just feel free to drop them below anyway I hope this was an interesting recap of the memory capacity video and it wasn't really any negative comments everyone was quite nice on that video and fortunately I didn't make any mistakes at least any that will point it out to me apart from a few people questioning why did certain things but I think I stand by those so yeah anyway it was probably not the most interesting video to start with because there wasn't any sort of little dramas surrounding it but yeah let me know what you guys think as always if you did enjoy the video feel free to like button for us we always much appreciate that I subscribe for more content if you haven't already and if you appreciate with your hair unbox and you want to gain access to a few little perks then you can start to our patreon we have private discord channel for the the dedicated harbor box members let's say and we also do a monthly live stream or two monthly live streams so anyway that's enough about that thank you for watching I'm your host Steve see you next time