Gadgetory

welcome back to our own box today we are looking into how much RAM you need to play the latest and greatest gaming tiles released in 2018 now it is about this time each year that I set up an a memory capacity quest and last year's expedition led me to conclude that for gamers 4 gigabytes is out 8 gigabytes is the bare minimum and 16 gigabytes of the sweet spot while 32 gigabytes is overkill that being the case for the 2018 version I will be dropping the full gear beta configuration goes without saying while they're getting rid of that one now and we will be focusing on 8 16 and 32 gigabyte capacities but before we get too far into this memory capacity comparison today's video sponsor is Tim group and their incredibly awesome award-winning T Force Nighthawk RGB ddr4 memory originally available in either black or white they now offer the legend series featuring a striking gold heat spreader complete with the same trademark RGB lighting they're available in 16 and 32 gigabyte kits supporting a range of frequencies and timings so please check the link in the video description for more information now if you're after a short and to the point answer this video won't be for you we generally don't tell our viewers what's what just expect them to take our word for it and we try and provide as many facts and fuze as we can making it more of a learning experience opposed to just hey do this or use this also as the title suggests the emphasis is on gaming for applications well it really depends on the application there are a lot of applications and even then what you're doing with said application a Premiere Pro for example using 4k footage that likes lots of RAM I think 64 gigabytes plus so it's best to try and research the requirements for the particular application you intend on using when it comes to games testing the impact of memory or RAM capacity is no easy task and there are many factors at play here as I did last year before we get into the results I just like to quickly discuss a few of the challenges faced when testing system memory capacity so challenge number one is picking the right Hardware the graphics card you use can influence how much memory you'll need for smooth gameplay and in worst case scenarios the speed of your storage can also impact performance for example last year I found the GTX 1063 gigabyte would often result in higher system memory usage when compared to the six gigabyte model and this is because at times you'll run out of erm forcing the operating system to tap into system memory then if you run at a system memory some game s that's a move to local storage and at this point the performance it will be so significant the game will almost certainly become unplayable so basically what this means is testing with an Archie x20 a TTI for example and it's 11 gigabyte vram buffer along with a super snappy nvme SSD probably isn't the best way to determine how much system memory your average gamer requires I'd also assume that if you have a $1000 plus graphics card buying 16 gigabytes of RAM or more really isn't an issue there's still more factors to consider as well the quality settings used can really have an impact for example if you have a GTX 1063 gigabyte graphics card and you're happy turning things down like textures to a medium type quality setting then that will reduce how much data is offloaded to the system memory taking a list into account i've tested a number of different hardware configurations and i'll be showing the impact this has on performance in a moment during my testing last year i discovered another issue when trying to show the difference between the various memory capacities so typically our benchmark passes run for 60 seconds and then that run is done multiple times usually we report an average of three runs and this means the system has a chance to cache the pass so while the results from the first run might see a shockingly low 0.1% or 1% low frame time figure this can be improved quite dramatically on the second run and then again on the third so giving you guys a three run average can be quite misleading when talking about the impact of various memory capacities have on performance the best solution I came up with at the time and I haven't found a better solution yet a year later was to run each benchmark pass once then reset the entire system load up the game and then run the next pass and then I would do that three times and then I would take the average result there so I would have an average of three runs but it would be an average of all these first passes the benchmark pass is actually 90 seconds in total but I'm only appalling performance for the last 60 seconds as the performance here more accurately reflects what you'd see when gaming is you're not always loading into a game for the initial 5 to 10 seconds at least to the game I can still be loading in assets and this will cause some issues with frame drops and things like that and this is even present on computers with sufficient system memory typically the solution to this caching problem is to just run the benchmark pass once usually throw away that result or at least average it against two or three more runs and then you get a more reasonable result there something that's more true to what the system can do and this is what we do when testing CPUs and GPUs but as I said this method isn't particularly useful when measuring the impacts of memory capacity finally benchmarks aside a good indicator for working at how much memory you'll need to play the latest games is just to monitor memory allocation but of course to do this you will need more memory than the game actually require so it's good to do this with sort of an overkill amount of memory whether that's 32 gigabytes or whatever having said that it's not foolproof but it does give us a pretty good idea of how much system memory a game requires to avoid any slowdowns again the choice of Hardware will also impact the amount of system memory used so having said all that I'm first going to show you memory allocation and a few popular modern titles of course released of this year and all this gameplay testing was conducted on our core I 999 wrote a test system with 32-year bytes of ddr4 34-hundred memory and a geforce gtx 28 ETI graphics card the test system has a number of applications running in the background so it's not just some completely fresh install of Windows with absolutely nothing else running we have things like steam running origin you play battlenet epic games launcher there's discord on there running not that it's doing anything about chatting or anyone but it's they're open in a chat session we have chrome with a few tabs open there's 10 tabs in total and then we have MSI Afterburner rivatuner collecting some on-screen statistics for us and then fraps I should also note that all testing has been conducted at 4k use in the highest possible quality preset again with an RT X xx atti so starting with Assassin's Creed Odyssey we see around 29% of our 32 gigabytes allocated throughout our testing this RAM usage hovered between nine and a half to nine point three gigabytes with me RAM usage around 8.1 to 8.2 gigabytes this suggests that a system with just 8 gigabytes of RAM could see performance issues in this game moving on we have battlefield 5 multiplayer using the 64 player mode here RAM usage gets up to around 33% hitting 10.8 gigabytes and basically never dropping below 10 point 4 gigabytes vram allocation was also very high holding steady at nine point seven gigabytes on the Narvik map so again we have another tile that will likely run into stuttering issues with 8 gigabytes of memory next up we have call of duty black ops 4 and this title is a serious memory Pig at least based on the small amount of testing they've done with this title Ram allocation held pretty steady at 12 gigabytes while VRAM usage was also very high at 10.5 gigabytes making this the most demanding title that we've tested yet f1 2018 is a title that you'll probably get away just fine using 8 gigabytes of memory as we saw system memory usage peak at 8.5 gigabytes that's right on the edge for sure but you'll probably get away with that at the 4k resolution VRAM usage was quite high at seven and a half gigabytes but any GPU that can handle this extreme resolution will have at least 8 gigabytes of vram anyway next up we have Far Cry 5 and here we saw RAM usage hit 9 point 8 gigabytes though VRAM usage was still reasonably low at five point eight gigabytes still that almost 10 gigabytes of system memory you're starting to get a bit too far beyond what an eight gigabyte RAM capacity will let you get away with hit man to push total system memory allocation just over 10 gigabytes for the entirety of our test and we saw at peak at ten point three gigabytes VRAM usage was also similar to what we saw in Far Cry 5 at around 5.6 gigabytes just cause 4 was very light on memory though avalanches apex game engine is seriously showing its age in 2018 less than 7 gigabytes of RAM was allocated in total and less than 5 gigabytes of vram was consumed monster hunter world is a title that should work just fine with a gigabytes of RAM here we saw allocation peak at 8.3 gigabytes and VRAM usage was also quite low under five gigabytes shut off the Tomb Raider was certainly one of the best looking games released this year and while it uses a reasonable amount of system memory it's certainly not extreme at around nine gigabytes VRAM usage though was quite high consistently using just over eight gigabytes Star Wars Battlefront - as a memory intensive title and here we were consistently seeing RAM allocation pushed above 10 gigabytes generally - around ten and a half gigabytes VRAM usage was also quite high at around 8.30 bytes second last game we're going to look at a strange Brigade and this is a well optimized title that runs very well on a wide range of hardware therefore it wasn't surprising to find that the memory requirements aren't that high just seven and a half gigabytes of RAM was allocated though what the 4k resolution VRAM usage did still hit six and a half gigabytes certainly nothing extreme there but higher than some of the titles we've tested so far last up we have vermintide - and here system memory allocation hit nine gigabytes while VRAM usage was relatively low at five and a half gigabytes this is a title that you probably could play with eight gigabytes of RAM but you would be right on the edge okay so now that we've had a look at memory allocation using a high-end hardware and a dozen tiles released this year let's take a closer look at the performance numbers when testing battlefield 5 hitman 2 and shut off the term roundup' I'll be doing this in two parts firstly with high end GPUs at 1440p and 4k and then with mid-range GPUs at 1080p and 1440p please note for all this testing I'm still using the kora 999 or okay test system the only hardware changes other than the graphics cards is obviously the memory for all the memory testing I'm using team groups T force dark for the 8 gigabyte configuration either kit with 2 4 gigabyte sticks using self 15 timings and then basically we've used the same type of memory but with 8 gigabyte modules for a 16 gigabyte capacity then for the 32 gigabyte testing I'm not going to go with just two sticks I'm going to actually use a 4 of the 8 gigabyte modules so - 16 gigabyte kits giving us a 32 gigabyte capacity team group does offer a 32 gigabyte kit with just two 6 thinking about modules but they use looser timing so I opted for using the 8 gigabyte sticks instead so that way we have the same timings across all tests figurations okay so first up we once again have the r-tx 2082 i but this time on hand for testing we also have vega 64 liquid and the gtx 1070 starting with battlefield 5 at 1440p we see virtually no difference in performance between these 16 and 32 gigabyte configurations and given what we just saw when looking at the on-screen statistics this shouldn't surprise anyone however as we reduce the system memory capacity to 8 gigabytes we start to see a reduction in performance at least we're looking at the frame time performance that is the average frame rates are at most 2 FPS change but if we look at the 1% low and in particular the 0.1% low results we see that the frame rates have become a less consistent for the r-tx 28 ETA we see up to a 7 percent drop in frame time performance 8% for Vega 64 and 7% for the gtx 1070 moving to the 4k resolution and here the r-tx 20 ADC is frame time results suffers by 5% but more interestingly than that is the noticeable reduction in performance when going from a 32-year bytes to 16 gigabytes which was very unexpected furthermore the 0.1% low result is reduced by 10% we're going from 32 gigabytes down to 8 gigabytes however for the much more heavily GPU bound configurations using Vega 64 in the GTX 1070 we see that the memory capacities have little impact on performance at the 4k resolution the hitman 2 results are interesting for a few reasons firstly the r-tx 2082 I consistently saw a performance reduction as we lowered the amount of available system memory and it wasn't just frame time performance that suffered we saw previously this game was quite memory hungry but 16 gigabytes of RAM really should be more than enough to cover it so it's interesting to see 32 gigabytes offering a 4 to 5 percent performance boost that said the performance game with the 32 gigabyte configuration wasn't seen when using either Vega 64 or the gtx 1070 in fact vega 64 delivered the same performance across the board whether the system had 8 or 32 gigabytes of memory but with the gtx 1070 while we do see these same performance with 16 and 32 gigabytes of ram the frame time performance starts to really suffer with just a gigabytes here we see an 11 percent drop in the 0.1 percent low performance moving to the 4k resolution and here we again see better performance for the r-tx 2082 I went paired with 32 gigabytes of system memory this time frame time performance was boosted by as much as seven percent again Vegas 64 saw no change in performance even with 8 gigabytes of RAM and in the gtx 1070 was much the same though again we did see a small drop off with just a gear bytes of ram shut off the Tomb Raider has some very interesting results for us for this title the r-tx 28 ET i delivered the same results whether it was using either 16 or 32 gigabytes of RAM was all much the same however with just 8 gigabytes the performance hit was massive though you wouldn't know it if you just focused on the average frame rate the hit to frame time performance though was substantial and at times the stuttering in-game was very noticeable here we saw up to a 32 percent drop for the 0.1% low result that said once again the performance it for Vega 64 is in that extreme at most we saw an 11% reduction in performance while the hit for the gtx 1070 was much more severe at 23% then moving to 4k we see at the ihx 28 ET i doesn't suffer nearly as badly with just 8 gigabytes of memory though we do see a reasonable drop-off for Vega 64 in the gtx 1070 we're looking at frame time performance so bit of a reversal there okay so previously we've found that it's the lower end graphics cards featuring less of VRAM that suffer the most when running with limited system memory so I went back and retested all over again with the radiant rx 588 gigabyte RX 574 gigabyte and GTX 1063 gigabyte at 1080p and 1440p so let's see what kind of impact the various Ram capacities had on these configurations first that we have battlefield 5 and as expected little to no difference between 16 and 32 gigabytes however with just 8 gigabytes we see a noticeable performance hit to the RX 570 and gtx 1060 so much so that even the average frame rates dipped we saw this in the last year's test that is the 3 gigabyte gtx 1060 really struggling with frame time performance and modern tiles when using less than 16 gigabytes of system memory naturally the situation isn't improved at 1440p though it's really interesting just how much the RX 570 suffers with just a gigabytes of RAM in this title they hit for the 3 gigabyte GTX 1060 is pretty sniffing as well but damn does that radio and GPU get hit really hard basically you don't want to be playing battlefield 5 multiplayer with just 8 gigabytes of around especially if you have a lower-end GP with 4 gigabytes or less vram that's said for titles such as hitman 2 the performance is basically non-existent we're not seeing much outside of the margin of error here the margins do open up a little bit at 1440p for the 3 gigabyte gtx 1060 when paired with just 8 gigabytes of ram but overall nothing too extreme as seen previously with high end GPUs shut off the Tomb Raider does run into a few issues with just 8 gigabytes of RAM and we're certainly seeing that here with these three mid-range graphics cards they hit 2 frame time performance in particularly 0.1% Lowe's was quite extreme for all three configurations naturally this was also seen at 1440p and here the 3 gigabyte gtx 1060 really tanked with frequent and very noticeable frames stuttering being a big problem the 1440p resolution also proved to be a bit much for the RX 570 ok so at this point I've probably hit even the most dedicated Harbor on box viewers with enough benchmark numbers time to start wrapping this one up it seems pretty clear at this point if you want to play the latest and greatest the most demanding titles you'll want 16 gigabytes of RAM I found this time last year that 8 gigabytes really was the bare minimum and now that's truer than ever of course if you play less demanding tales such as fortnight overwatch a rocket League for example or old but still very popular titles such as dota 2 and csgo then a gigabytes of system memory will be plenty but if you want to get amongst it in battlefield 5 or enjoy the breathtaking visual scene and shut off the Tomb Raider without the constant reminder that your system isn't quite up to par then I'd recommend getting at least 16 gigabytes that said for mid-range to lower end rigs where - stuttering can at times be caused by a lack of erm a slow storage device not quite enough CPU cores and so on being on the edge for memory capacity probably isn't going to make that much difference for example a core i3 a t3 50k system with a basic TLC SSD a graphics car with 2 to 3 gigabytes of vram you're gonna see some stuttering and titles such as battlefield 5 so upgrading from a gigabyte 16 gigabytes of RAM will certainly help it's not going to completely solve the stuttering issues alternatively those rockin high-end hardware you know something like an RT X 2082 IO all this talk about memory capacity is kinda redundant anyway just get 32 gigabytes and forget about it after all if you're spending at least $1300 u.s. on a graphics card and you honestly going to think twice about spending three to $400 u.s. on ddr4 memory probably not for most of you 16 gigabytes is going to be the sweet spot and this is what I recommend so in a nutshell for today's latest games eight gigabytes really does seem to be out now sixteen gigabytes is the sweet spot as I said and 32 gigabytes is still overkill and that is going to do it for this one if you did enjoy the testing be sure to the like button subscribe for more content and if you appreciate the rotor power box then consider supporting us on patreon thanks for watching I'm your host Steve and I will see you next time unless I lose all my memory and forget how to do this