Dragon's Dogma 2 is a Mess: GPU & CPU Benchmarks, Bottlenecks, & Crashes

[Music] Dragon's Dogma 2 has been an absolute nightmare to work on check this out that's what happens when the game tries to render I.E play also we had uh multiple accounts get locked out because you get five Hardware changes in a 24-hour period and that's it cuts you off it eventually Cycles through and that is thanks to what's known as DRM we had a solution though we bought four copies of this game there's other complaints too like has a singular save slot which makes things annoying to Benchmark but also just is really terrible from a play in the game perspective uh we struggled to transplant save files between steam accounts despite it being basically a single player game and ultimately what we did was buy multiple copies play through to the same point on all those copies so some of the team got paid to speedrun Dragons Dogma 2 to the first city uh the past couple days and it was a full team effort this is It's ways it's archaic but we're not here to critique the systems of dragma dragma dragma dog and 2 uh we are instead here to talk about the performance and there's a lot to talk about with this the game is somewhat of a nightmare in that aspect too fortunately for us that gives us a great opportunity to show this which is the freshly updated GPU busy plus the new GPU weight and the CPU busy and CPU weight metric that allow us to pinpoint bottlenecks precisely even in times of uncertainty uh we'll look at all that today plus some research we have ultimately a CPU Benchmark with a list of about that was 20 or so CPUs on it and we would have done more but because the five device activation limit that's what we get and they need time to update it anyway uh but at least they got 280 bucks out of us for this video so you can go over to store. Camas access.net if you want to help fund our purchasing of games that have activation limits so go grab a GN 3D metal emblem pint glass on the GN store and pour one out for our nearly $300 while supporting our content for this video though I'm personally the most excited about playing around with the new GPU busy and GPU weight metrics before that this video is brought to you by thermal take and the tower 300 the tower 300 is a full-on showcase PC case built to present the computer straight on with its angled tempered glass windows or on a unique mounting stand to show off the build in new ways the tower 300 has a layout that positions the GPU fans against the mesh panel with ven ation on the opposite side for liquid coolers and CPUs there's also an included 2 140 mm fans up top the panels use a quick access tooless system to be quickly popped in and out for maintenance and you can learn more at the link in the description below this is doing a lot of stuff in the background on the CPU uh with the NPCs this is something that the developers have actually pointed out in their recent pseudo apology update post because they're getting destroyed on Steam Reviews right now now a few of the issues we encountered with Dragon's Dogma 2 included unpredictable game crashes but then more predictable ones as well on Lower vram cards like the 4060 we also noticed micro stutter and simulation time error wherein the frames would appear to be relatively fluid even if you look at frame times but the animation within the frame was improperly timed other issues include wildly variable performance depending on the area of the game so unlike most games it'll switch from a hard CPU bind say in the city to a hard GPU bind in Wilderness areas most games are more balanced in one area and might be hard Bound in another or they have a soft bind in one area and a hard bind in another and in this instance hard versus soft bind refers to sort of what percentage of the total frame render time each component is waiting for the other the game is already ranked overwhelmingly negative as a result of performance issues game crashing and DLC and the performance issues alone leave us in the usual stage of absolute and utter bewilderment at why these game companies keep doing this to themselves surely they knew about this before launch but here we are doing it again uh there may be some optimization they can do here they seem to suggest that that's in the cards but anyway we'll just get straight into it we'll start with the research get through some of the charts and the CPU benchmarks then we're going to look at those new metrics as an overlay on top of gameplay this section will show you some of the research that went into finding a test area this is conducted by by actually playing the game and taking frame time samples during the course of play we take these manually and we assign names to each sample here's the result this was done with an overclocked i7 CPU and an RTX 490 at 4K in every single game we test excepts balder's gate 4K is enough to shift the load entirely to the GPU and avoid a CPU bind that was true in many instances for this game also but in the city even at 4K it was shockingly not enough to overcome CPU limitations even with our overclock we'll talk about the 7800 x2d and the 1400k results next overall though this is what we saw across 90 minutes of gameplay the range in this game is huge with a Max minus Min result of about 91 FPS here that's massive that range is the size of a whole different GPU on its own and it's a huge swing across different areas of the game in the wilderness areas we were in the range of 120 to 170 FPS average and we were GPU Bound in those scenarios which will prove momentarily with some captured gameplay combat did not have a heavy impact on performance unless High counts of NPCs were involved the towns were the heaviest though down in the 70s to 80s for our verw worth passes and in the 98 range for the smaller town M the first Camp was also around 98 FPS average the common factor was the presence of NPCs after all this work and still set on doing a GPU Benchmark we settled on the highest load area for the the game as we prefer to represent something closer to a worst case scenario to give people Headroom for figuring out what they need unfortunately at the time we didn't yet appreciate just how CPU bound this scenario was here's the first round of GPU Benchmark results as soon as we observed the 490 yielded the same performance as a 4070 TI super we knew that GP benchmarking would be pointless at least it would be in this area and this by the way was on the 14900 k at 5.9 GHz so we had all already switched CPUs to something higher end even switching to the 7800 x2d which was a major timeing commitment to set up a whole new test bench yielded the same result as the 14900 K and this was all at 4K so we gave it as heavy a bind as we could in this set of tests the 4070 TI and the 490 performed the same 78 x3d didn't change that uh but it's not because of a GPU bind even though we're at 4K here and we're seeing similar performance the 4060 dipped but this just isn't enough range to produce meaningful GPU comparison benchmarks without test in a field instead of a city which we could do but it'd be less demanding as an area of the game and we want to investigate this Behavior after this we moved on and decided to make this video a CPU Benchmark instead so let's do those charts next and then we'll come back and look at some really interesting information with GPU busy GPU weight CPU busy and CPU weight and we'll talk about some of the simulation time error problems we observed so here's the results of the 4090 when everything is stock the 78 x3d leads the chart at 86 FPS average but the game still feels terrible to play at times with simulation time error clearly appearing in animation this is a metric that we're still working on we been an interview with Tom Peterson talking about this some more the I9 1400k is about the same level here running at 83 to 84 FPS average when CPU bound that has the 7800 x3d as much as 3% ahead of the 4900 K it's not enough to matter the 1200f has historically been one of the best Ultra budget gaming CPUs but it occasionally shows hard limitations this is one of those times the frame times are actually a lot worse than shown here as averaging multiple test passes is hiding the horrible stutter in the first pass we'll come back to that look at a moment we would consider the 1200f unplayable for this game there's really nothing you can do we tried tweaking the settings we dropped everything to the lowest possible doesn't matter all the settings are basically Graphics intensive so it didn't help here the 5600x exhibited similar Behavior averaged over multiple passes it does okay but the first time through an area brought its 0.1% lows down to about 6 FPS for that metric indicating large frame time spikes other points of interest include the 5800 x2d versus the 5800 X for we see a 16.5% benefit from the 3D cache the 750x was unable to leverage its additional cores in a meaningful way in this one there's a point of diminishing returns and it's past that point higher frequency would be more helpful here than going to the extreme core count of 750x let's move on to the frame time here it is this is the insanely blown out frame time plot from our first run on the 1200f where we first hit an area of the city that must have begun some heavier background calculations to the extent that we thought the game actually crashed which it also has done a lot when this happened so what this means is that the threads of the CPU are fully occupied processing in the background and none of those threads are handling the actual rendering and presentation of the frames to the screen instead they're doing other things for perspective that looks like this what you're seeing right now if we Zoom the scale in to see the rest we still had to raise the roof to 800 milliseconds to accommodate the next two largest spikes for reference at 60 FPS one frames about 16.7 milliseconds to present an Excursion of 8 to 12 milliseconds is noticeable in here we have excursions of about 750 milliseconds in one instance the game freezes up hard and it rests at times on the 1200f even the newer and better 5600x occasionally hits these problems in this plot you can see we have a frame time Spike nearing 270 milliseconds with regular spikes to 60 milliseconds in fact every time we pivoted the camera left and right which is part of our Benchmark course we saw this sort of micro stutter and animation of the main character at best or just actual playback stutters of the entire frame at worst now we're moving on to the really fun part that I'm excited about this is game capture while using GP busy CP busy GPU and CPU weight to illustrate bottlenecks the bottleneck is a moving Target in this game which makes this perfect for this new testing and it shifts between the CPU and the GPU depending on the area tested this is expanded from the past GPU busy benchmarks we did because now we have that weight metric and that's all explained in a separate uh Deep dive engineering piece that will link below you should definitely check it out if you don't understand what it is won't go into it here but it's really cool and it gives us some new insights in the city tested at 4K with a 490 and a 14900 K we observed CPU busy functionally at parody with frame time which means the GPU is waiting instructions and can't be fully leveraged we see this also in our GPU and CPU weight metrics where GPU weight is often in the 4 to 6 millisecond range while CPU weight is basically zero we're completely limited by the CPU even at 4K and high settings you can also see this in the GPU power metrics histogram where the GPU is under its power Target click to photons about 8 milliseconds here which is acceptable CPU utilization is not useful in this plot but it's here because it's a great learning Lon for the many who rely on this single number for this measurement because if you hear us talking about being CPU bound this whole time but you're staring at that 30% number going no you're not that is uh not really a useful number so here's why if you were to check Hardware info or task manager even which also isn't great the latter uh you'd see that at least eight cores and sometimes up to 12 are 100% loaded by the game the remainder of the cores go largely underutilized which contributes to the total utilization percentage being lower because it's averaging across the entire CPU and also not all load is the same as calculated here so we are in fact completely CPU bound even though that number is not showing 100% that's why that number is basically useless in most instances daytime is similarly constrained to nighttime we're close to parity on the CPU and the frame time total metric which means that we are CPU limited moving to other areas of the game upon exiting the city you see the Lo shift somewhat rapidly to the GPU as soon as we're outside of bounds of where it's processing all the AI for the city you can see the frame rate climb overall which means the frame time is reducing and now the CPU and GPU busy metrics are closer to parity with frame time to actually balancing out here and the further away from the city we go the more the load shifts to an entirely GP bound situation power consumption has also climbs to more expected numbers in combat load moves around and occasionally triggers a CPU bind but generally it's GPU constraint click to Photon Trends closer to 12 milliseconds at time still fine and you can also see a CPU load climb disproportionately to the GPU load as the Griffin makes its way into the water which we were appreciative of we can't be sure if something else happened in the background maybe this is from the pathf finding but it could be from other NPC calculations you can see another big spike in CPU load as we get a level up and the camera pans back toward the village in the town on return passing the ox cart and re-entering the town confirms what we knew the GPU is now reducing its load and it is no longer fully loaded it wants to be but the CPU can't keep up it's not able to send enough tasks to the GPU the CPU is busy getting hammered by the AI processing for the NPCs the takeaway is that for testing in this game uh our options are these if we wanted to test a worst case scenario period we'd go to the city if we wanted to test a CPU scenario it'd be the city and if we or other reviewers wanted to evaluate GPU performance in as isolated a way as possible and make the most sense to do so outside of the city ideally in a field with a lot of grass and increased grass quality as that had a hard impact on performance we noticed some animation micro stutter with high grass quality as well and as our earlier research chart showed it didn't really matter whether we were in combat or not you just find one of the heavier Outdoors areas with a lot of polygons and geometry and hopefully basically zero NPCs and that becomes the GPU load but this was all at 4K and so that scaling would be reduced at lower resolutions so this game is definitely CPU heavy and it's actually kind of interesting for us because in a lot of situations with games they just don't load that many threads that heavily and in this one we're seeing significant like 100% load on multiple threads it's not a single threaded title like you see with some of valves for example like in the past Dota 2 I haven't tested it recently uh csgo before it was phased out things like that this is on multiple threads and we see that again manifest like with the 1200f but it is uh CP Bound in such a way as to bring the game to a crawl it just seems like there has to be something more that can be done there uh especially since they've kind of already hinted at that being something they're working on the developers that is the load is heaviest in towns outposts basically Villages and the city itself the city is the heaviest of of those however a town or a village or an outpost whatever you want the the first Outpost is very similar in performance to some of the city performance later on in the game so uh in the planes and Wilderness and most combat scenarios if they're outside of a city then uh the performance goes up significantly it starts to become a GPU bind at least with something like a 4090 uh at 4K but if you're not at a higher resolution uh and you're on a higher end card then you might still be kind of in the middle more balanced between CPU and GPU load the good news is if you want to improve the performance in the city there is a way to do it you can kill everyone kind of like in the Elder Scrolls that's going to get me on a list it is wild though to see the performance varies so much because you'll get as much as a 30% bump just by walking out the gate of the city into the immediate outline area it can double in other areas too so uh depending on what device combination you have there is a wide range of performance this means that for the most part the game can remain playable on higher frame rate in situations where higher frame rate matters the most like combat in the wild so that's a good thing it also means that if you are GPU heavy mean your GPU is maybe your highest invested part in the build like is the case for most people and you find yourself being largely CPU constrainted here you can kind of blast all those settings they're not really going to hurt the performance in those City and outpost areas you'll start seeing it maybe in uh Wilderness and combat but the frame rate also boost there so probably get away with it at least we could with the test setup we were using for most of this then for things like the 12100 F the 5600x we suspect that being more limited in core count like four on the 1200f means that there aren't enough threads available to task everything so between the scheduling of the engine and what cores are available we think it's maybe dropping the render thread in favor of whatever else NPC tracking pathfinding physics game State whatever's going on in the background AI uh and once an area is loaded it's much more reliable but loading those areas can freeze the game and put it on the verge of crashing or just straight up crash it the 5600x wasn't nearly as bad but we still ran into occasional 250 to 300 millisecond spikes in those scenarios I will say it was cool seeing the game load 6 to8 threads pretty reliably at 100% when they were available going up to 12 uh in some instances the vram can also be heavy but for the most part unless you're manually maxing out some of those settings beyond what the default high-profile does uh you're more likely to be CPU bound first than anything that's all with brand new drivers by the way so all that Nvidia testing in this video was done with the newest driver that does support officially Dragon Dogma 2 we did some RT testing too it was really lightweight uh so just for the test I mean so we saw about a 10 FPS win from about 72 to about 80 something 82 84 when we did that on off test in the city didn't go hunting for a bunch of R tracing comparisons and things like that cuz Focus again became CPU is but we did collect some of that data that's the answer for you makes it a little more GPU heavy but it wasn't crazy this might be a game though where overclocking actually poses some value uh especially if you're on an older device where there was more readily available headro for boosting the performance so that's it for this one it was It was kind of fun to Benchmark because we got to use all these new tools see cool things in the results but the Steam Reviews talking about performance they're grounded in reality uh objectively so definitely one of the heaviest games we've benchmarked and sadly one of the worst we've tested in the last year with regard to things like instability uh crashing and just kind of generally unpredictable performance at times so um hopefully they fix it but this seems to for whatever reason be the the default way that gamees shipped now uh couple exemptions like balers Gate 3 so that's it for this one thanks for watching support us as always on store. Gamers access.net or patreon.com Gamers access to throw a few bucks our way and cover that $280 so I don't know what I'm going to do with these four accounts now um anyway subscribe for more we'll see you all next time

Channel: Gamers Nexus

Views: 423,771

Keywords: gamersnexus, gamers nexus, computer hardware, dragon's dogma 2 benchmarks, dragon's dogma 2 gpu benchmarks, dragon's dogma 2 cpu benchmarks, dragon's dogma 2 cpu comparison, dragon's dogma 2 gpu comparison, dragon's dogma 2 best cpus, dragon's dogma 2 best gpus, dragon's dogma 2 performance optimization, dragon's dogma 2 graphics settings, dragon's dogma 2 max settings

Id: twEERkUyAXE

Length: 19min 44sec (1184 seconds)

Published: Mon Mar 25 2024