AMD AFMF TESTED - AMD Fluid Motion Frames is WILD… (LATENCY + FPS)

you might have heard about amd's AFM feature recently that being AMD fluid motion frames which Yes means say AMD ammf is like saying rip in peace ammf is heralded as the oneclick way to double your frame rays in basically any game but like what actually is or what are fluid motion frames well in short it's frame generation the same idea as nvidia's dlss frame generation but it's done in the driver that means that much like radio and super resolution the game doesn't need to support ammf or frame generation at all in fact the game and even performance monitors like frame view or oat won't know anything about the generation happening now it does have a few limitations it only works on RX 6000 or 7,000 series cards and 700m graphics on apus I'm testing with an RX 7600 here and something I only found out after starting to test it only supports DirectX 11 and 12 games no Vulcan support which means Rainbow Six Siege does not work strangely even when I launched Siege in DirectX 11 mode it still said it wasn't supported perhaps AMD just blacklisted the cgx who knows but for a plethora of reasons that I'll explain later I switched to Starfield instead oh and it's worth noting that you're meant to have both HDR and vsync disabled for this too right let's look at some data first as a baseline I tested the system with just radon antiag enabled it's a feature I would recommend enabling regardless of any other settings assuming you have an AMD GPU anyway so I'd consider that a fairly Baseline setup the next test was just turning ammf on so the only two AMD features enabled were anti-ag and ammf lastly I also enabled the new hyper RX mode which turns on every tool in the toolbox including things like radon super resolution as for recording data for latency as always I'm using my open-source latency testing tool that I developed and sell over on osrtc by the way Link in the description and for FPS as mentioned tools like present modon and frame view which is just present modon with an Nvidia rapper won't detect the total performance here so I'm using amd's own driver based performance monitor one extra advantage of that is that it'll also report how much latency the frame generation is adding so I've included that data here too so what's the results well if we look at the performance data first you can see how effective ammf really is we go from 104 FPS average to a whopping 166 with just AF FMF enabled and enabling hyper RX and all of those other features like rsr on top gives you another 10 FPS or so and actually gave a significant boost in the 1% low results too a full 30 FPS higher here clearly enabling FMF is a no-brainer then right well if you notice that extra number at the bottom there the frame gen latency average that's pretty important obviously with no frame generation enabled we're adding zero latency to our gaming experience with ammf enabled though you're looking at 20 milliseconds of added latency if you've watched any of my you know latency guide videos you'll know that 20 milliseconds is an awfully long time to be adding on top of the game itself let's look at the latency latency results in fact as expected for a game that isn't running overly fast even on a 165 HZ display 100 FPS means new frames every 10 milliseconds or so and so it looks like Starfield on all low settings with FSR 2 enabled takes about three and a bit frames to render an input that's pretty slow as is but when you add ammf on top of that you end up with a frankly appalling 59 milliseconds of average delay between giving and inputs to that input being shown on screen for a game like star field it's a genuine toss up whether the improved smoothness of the artificially increased frame rates is worth it over a doubling of the input delay the big problem here is that the higher the latency the harder it is to aim with any accuracy if you spend most of your time in the dialogue options and running away from fights ammf is the tool for you if you want the competitive advantage in a gunfight you might want to switch ammf off now I was going to leave it there and move on but in the interest of thoroughness I tested two more games to see how they handle it the first is the other super intensive single player game everyone loves to hate cyberpunk 2077 on the medium preset which uses FSR 2.1 by default anyway unsurprisingly ammf and the hyper RX mode offer sign significantly better performance than without like this is a 60% Improvement on average impressively this RX 7600 even at 1440p still offered 121 FPS on average but it's very obvious that you get a smoother experience with ammf on interestingly likely thanks to the frame rates being pretty similar the frame gen latency is about the same this time around 18 milliseconds on average that's confirmed when we look at the latency chart too where the stock results are a reasonable if kind of slow 28 milliseconds on average versus both FMF results at an astonishingly sluggish 54 milliseconds on average this one at least on the 7600 on the medium preset anyway is a little harder to justify cyber Punk relies a lot heavier on Gunplay and action and how even just trying to drive with double the latency just isn't going to be a great experience if you care more about visual Fidelity and are so more likely to be playing on a higher setting and therefore lower frame rates then I think that tips the scales in favor of using afms lastly I also tested CS2 I was almost certain that this would be a wildly different set of results and conclusions to Starfield and cyberpunk but honestly I couldn't have predicted the results I got starting with the FPS data it looks like ammf made the gaming experience considerably slower like we lost over 100 FPS on average here by turning ammf on we did gain a bit of stability bringing the 1% lows right up to the average rather than around half the average with FMF off but losing performance isn't a great sight I have to assume that part of that has to do with the frame gen latency figure 11 milliseconds is way way too high for a 2 millisecond frame time there was clearly a decent bit of time taken to analyze and render the fake frame and then display it and all of that work is for notd because not only do you lose performance your inputs take nearly three times longer to register we go from an average endtoend latency of 9.7 milliseconds a fantastic result to 25.9 Mill seconds which for a game like CS2 is a death sentence these two graphs should be all the evidence you need to not use ammf or hyper RX with CS2 don't do it just just don't it's worse in every single way okay got it good now something I want to take a second to explain is why these features whether they come from amd's open-source arms or nvidia's proprietary Clause will always add latency it's a fundamental function of how they work basically to be able to generate a somewhat useful intermediary frame it needs to be able ble to look at the current frame and the next one let me visualize that here so we've got the current frame that's on its way to the display that's all ready to go now any type of frame generation feature ammf dlss frame gen it doesn't matter it has to wait around for the GPU to draw the next frame then instead of just pushing that new frame straight out to the display which is is as low as latency as possible the frame gen tool then compares the two frames and renders an in between frame that will take some time to do as we've seen from the data and then it will send the fake frame out to the display and then after all of this time waiting the next frame actually get shot the fact that these features have to see what the next frame looks like to interpolate between them to create an extra one in the middle means that these features will always add latency and the only optimization here is how quickly they can do that interpolation that's why even nvidia's in engine version takes just as long interestingly though there is one key design Choice I'm yet to talk about here which is the AMD D has chosen to disable ammf during fast movement specifically this only seems to apply to Mouse movement testing in Starfield just walking around but not moving the mouse it shows the full 150 FPS or so but the second I start moving the mouse the FPS drops down to the more native 90 100 instead in theory this should give you the best of both worlds right I mean for just walking around you get all of that extra performance and smoothness but when you're in intense combat it disables itself for that quicker reaction time right well as far as I can tell no I mean the first problem here is that if it's tied to Mouse movement I can't think of a time when playing say Starfield I wouldn't be moving the mouse even just to navigate the world that seems to defeat the purpose of this feature to me second the latency test I was doing was a mouse move it should have been disabling itself for every test result but as far as I can see it stayed active the whole time or at least the latency stuck around you know whether this is because the you know OSL basically instantly shifts the mouse position giving ammf no time to turn off first or because the latency part is just there the whole time even if the actual frame generation doesn't end up happening I can't really answer that one but it doesn't seem all that appealing to me with that said the whole point of fluid motion frames is to provide well fluid motion and that's something that is worth viewing subjectively as well as objectively so let's have a play in Starfield to see what it actually feels like to play with so we have ammf on here so as I mentioned with the uh the whole movement thing if I move the mouse we see we're at like 70 80 FPS if I stop moving the mouse we're at 173 uh so it's a lot smoother now obviously I mean we've almost double the FPS uh but the second that I start moving the mouse to you know fight this terramorph um that drops away quite quickly although it does seem to be the the faster the movement the before it cuts off which I think might also explain why the whole um latency Figures were so different yeah see like trying to dodge this it drops way down to 80 rather than the like 10 whatever 150 170 it would be normally oh I don't have enough uh Mana or whatever called star power I haven't noticed any visual Jank yet occasionally it feels like it's stuttering as it kind of swaps between [Applause] modes that's by far my favorite uh star power by the way the uh the plasma much better it's definitely an unusual experience um it it goes from being you know a decent frame rate at like you know 8090 to being a an exceptional frame rate at like 170 but there is a noticeable difference between the two and how it decides to cut in and out also can be a bit jarring cuz like right now even though I'm jiggling the mouse it's at 170 or so but you know bigger movements are like adsing it's a lot smaller but then also it seems if I stay ads even if I move it goes back up to6 so yeah it's a bit weird I think it's time to wrap this up with some closing thoughts frame generation Tech is pretty cool although it certainly has some drawback on the one hand it's literally free performance but on the other it's a bassac solution to increasing your frame rates ammf itself is a pretty clever implementation the fact that it's baked into the driver means that it works with a bunch of games without the game developers having to do literally anything to support it that means that regardless of if you want to play Starfield CS2 or train Sim World 4 you can boost your FPS with a FMF the catch is that it only or the catch that it only works in DirectX uh games means that games like Rainbow Six Siege are off the table although much the same as CS2 you wouldn't want to use it there anyway strangely both with CS2 and with Siege I had a few bugs where it just claimed I was using an unsupported API when I wasn't and just wouldn't enable I'm not sure what's going on there at least with that one but anyway the onoff control behavior is actually pretty clever but it can also lead to stuttering and a bit of a jangy experience lastly on the latency front it's a pretty big ask double your latency for like 50% more performance for some games like Starfield or training Sim World 4 I can see that being a trade-off that ends up being worth it but for anything even remotely competitive there's no chance I'd be leaving this on of course with that said those are my thoughts but I'd love to hear yours in the comments down below what do you think about ammf and frame generation in general let me know in those comments also if you do want to be able to test stuff like this yourself you can pick up an open source latency testing tool that I developed and make literally at my desk downstairs uh available at o.com /lt otherwise if you want to see more videos like this one you can hit the subscribe button check out plenty of other videos in the end cards including those latency guides if this one was interesting for you and uh yeah you can also support the Channel with a load of links in the description I'll leave you to take a look but yeah otherwise thanks for watching hope you enjoyed it we'll see you on the next video

Channel: TechteamGB

Views: 12,483

Keywords: techteamgb, Tech teamGB, GB, tech, Tech team GB, amd fluid motion frames, frame generation, amd fluid motion, amd afmf, amd frame generation, frame generation amd, fluid motion frames, amd fluid motion frames test, amd fluid motion frames 6000 series, amd fluid motion frames cyberpunk 2077, amd fluid motion frames cyberpunk, amd fluid motion frames fortnite, amd fluid motion frames on vs off, frame generation on vs off

Id: hfjPXUl3Vw4

Length: 17min 45sec (1065 seconds)

Published: Mon Feb 05 2024