Gadgetory

given that the RX 580 is really just a 480 with an updated v bios and newer silicon we figured it'd be worth attempting to crossfire together in RS 580 and RX 480 to look at performance this seems like a realistic scenario Rx 40s will get harder to find and prices will soon skyrocket considering the end of life for the RX 40 production but our x5 ATS will live on owners may resort to then to 580 s or crossfire I want to make sure it all works as normally which may or may not include negative scaling we'll soon find out before that this coverage is brought to you by the Computex conference which runs from May 30th to June 3rd in Taipei Taiwan this year Computex is the biggest event of the year for PC hardware and technology where we preview the newest prototypes before they come to market we highly recommend attending or following this event online for industry professionals and enthusiasts learn more at the link in the description below so the first order of business is to say that yes this works you can crossfire 580 and a 480 it does not take any special hacks the question here was really just did and they imposed something that said this won't work because in terms of hardware it's the same architecture it's still Polaris the 580 in the 480 are not different other than the voltages and the frequency profiles on the 580 being a bit higher for each so that means in theory which you can do you would be able to take the 580 put it in the first slot crossfire it with a 480 and then you're really only going to be bound by the slowest card which is probably your 480 unless you were to overclock it to match 580 which is not too hard to do so that leaves us with the question of scalability and how well does crossfire do in our previous coverage of multi-gpu whether it was crossfire or sli didn't matter which one I I'm struggling to recall if there was ever a scenario where we recommended multi-gpu from either Nvidia or from Andy pretty sure we've never recommended that and it's because the scaling is kind of hit and miss so when it scales well it's really cool it's a great way to get some extra performance out of an older device or an ageing device but when it doesn't scale you're left with micro stutter and basically forced to disable the multigp technology whether that's SLI or crossfire and that means more effort involved in getting game-2 work with the system configuring profiles to enable a disabled crossfire sli based on the system and the games and it's just it's not the most trivial scenario to work with and you do generally lose some scaling and some of the games that are not really optimized for multi-gpu so it's always been a question now whether that remains true we'll see this is the first crossfire test of the year for us so we'll be testing the 580 with the 480 we use gaming X cards for each this is a 580 gaming X now this is a 480 gaming X this is a 580 gaming ax as you can see they are physically identical and we put those in crossfire enabled it in Radeon settings and we'll give it a go we've got power testing first and then game testing we've got 3dmark fire strike and pinus fight lengths in the article in the description below if you want to see the synthetics for the testing we're really clock limited by the slowest card here which is going to be the Rx for any gaming X but both of them are pretty decent for their categories the 480 gaming X was not a bad card when you look at its stock clocks and then other notes vram doesn't stack if you're not aware most you probably are but with things like dx11 there's no vram stacking so we're going to be limited to 8 gigabytes of video memory for anything that uses DirectX 11 things that use a newer api's might do some extra stuff with the memory they have the capability to it's just a matter of whether or not the developers have added that functionality to their particular game well start this test by measuring total system power draw from the wall just to get an idea of how much more consumption that is scaling with the multiple cards well 3d mark firestrike extreme the crossfire rx 580 and for 80 gaming that system is drawing 460 watts on the wall for comparison the RX for 80 game next standalone stock system to 248 watts with the 580 socks with met 281 the crossfire config is drawing about 64 percent more power than a single 580 gaming X or about 85% more power than the single R X 480 gaming acts with for honor our system power draw is that around 426 watts load that's the highest on the chart so far the profile with this game has some power management issues judging by fluctuations on a meter and by interion at 1440p and 4k there's fierce micro stutter at 4k regardless our power draw is about 37 percent more than the rx 580 gaming Xbox system which draws 310 watts and we're about 54 percent higher in power draws on the our X 480 stock GPU idle power draw isn't really affected in a meaningful way here because this crossfire doesn't really do anything when you're on desktop or windowed mode we've got a couple extra watts for the LEDs and polling but crossfire doesn't engage on the desktop so we're looking at around 80 watts idle wall draw I stated you can check the article below for 3d mark tests now let's move on to gaming we're starting with Mass Effect Andromeda for crossfire is big challenge with Mass Effect Andromeda and that seems to be overlooked when considering multi-gpu despite folks proclaiming Andromeda's multigp support there's been limited discussion around the reason why it's sort of cheating there's an option that enables by default when going multi-gpu and it lowers graphics settings to reduce micro stuttering and improve aim rates the game doesn't actually tell you which settings it's lower in live while it does it it just sort of does it and that means that benchmarking in performance mode for multi GPU does not produce comparable numbers to the single card benchmarks since those tests will be more intensive we've actually got a few screenshots of performance mode versus quality mode that we can show on the screen and that's for multi-gpu quality mode produces the same image quality as a single card so the graphics actually look the same frame by frame but there's intense micro stutter at higher resolutions and just kind of bad frame rates even at 1080p so it's not a good experience if you were to enable performance mode as you can see in the screenshots the differences between the two show that performance has a lower quality primarily in places like ambient occlusion local reflections and some changes to depth shading and refractions you can kind of look at the shiny surfaces and see that the quality of the reflections in those services is reduced and the ambient occlusion in places like door frames and ceilings is reduced as well with performance mode so again the difficulty here is that they are not linear comparisons if you're running performance mode versus a whole bunch of cards that were single GPU running the particular set up so it's kind of cheating to use performance mode but we'll put both numbers in the charts anyway just so that everyone can understand the issues and where things fall 4k doesn't look good for quality mode we're experiencing intense game breaking micro stutter that doesn't even show up in the already dismal framerate numbers we're at 19 FPS average behind both the single rx 480 and single rx 580 this is the first time we've seen negative scaling in a game and it's really pronounced here you'd be way better off to say when crossfire in this case but let's move on to 1440p ensure both performance and quality mode numbers quality mode produces the same at quality frames as every other card and puts us at 38 FPS average the worst on the chart behind even the RX 470 there is intense negative scaling with crossfire in this title at least until we switch on performance mode which again cheats the graphics to be a bit lower the game at now peg is a 580 and 480 crossfire configuration at 54 FPS average tied with an overclocked rx 580 single card configuration we're ahead of the RX 580 gaming X single GPU stock by about 1 to 2 FPS even with this mode you're far better off disabling the particular cross fire configuration we're running and using a single card instead because one your graphics quality will be better and to the framerate will be mostly equivalent at best we're seeing three percent scaling an average FPS 1080p looks about the same for quality mode plant in the crossfire config below the RX 470 and average frame rates and far below it in frame time consistency scaling is negative here and it's negative in a big way 29% switching to performance mode puts us at 100 FPS average with one prism lose at 69 and 0.1% at 56 that's finally showing some real scaling over single 580 gaming X which operated a seventy eight point seven FPS average but again the single card was under heavier workload so it's debatable how much you can really count that as a true scaling result if the image quality is not identical between configs for honor is another game that has some interesting challenges crossfire performs better for us at lower resolutions of this game like 1440 p.m. 1080p and did a bit worse in 4k with 4k we saw fierce micro stuttering that would necessitate disabling crossfire or lowering resolution as it was utterly unplayable despite not showing up as much in benchmark numbers we'll skip the 4k results for that reason moving instead to a more playable 1440p we're getting seventy-nine FPS average with 71 and 68 fps low is on the mixed crossfire configuration showing positive scaling over the single rx 580 gaming X the performance gain is about 32 percent with an insignificant amount of teryan as a side effect not bad scaling it over a single rx 480 is about 39 percent this puts the configuration between the 980ti and the 1070 reference card at 1080p scaling is again positive with a 120 FPS average throughput and lows at or above 100 FPS the arts 580 gaming X single card performs around 93 FPS average showing that the mixed crossfire card scale over a single 580 gaming X by about 29% we're still below a 1070 reference card here Doom is up next on the list using Vulcan as the API of choice Vulcan just like OpenGL has never really delivered for multi-gpu users in Doom we're not seeing any scaling here at 4k as we can highlight on the screen the 580 and crossfire configuration are basically the same that carries over to 1440p where there's slight overhead without any scaling to speak of at least in a positive way crossfire devices even if you were to use the exact same device just don't scale in this game 1080p is no different as will briefly show here there's slight overhead in the worst case or just equality in the best case Ghost Recon wildlands is a bit worse than lack of scaling we just couldn't even get the benchmark to start with this one the game hard crashed at all resolutions with our very high test settings and Ghost Recon wildlands running it earns the crossfire rx 580 and 480 a solid DNF they did not finish any of the tests with this game Sniper Elite offers some redemption though had 4k with the x12 and async compute enabled we're seeing the crossfire rx 580 and 40 gaming X cards placed at 79 FPS average with lows at 63 and 65 80 gaming a single card operated at about 40 FPS average with lows it linearly lower than the crossfire config at scaling of nearly 2x we almost never see that in games Sniper Elite is one of the very few titles that can make a case for multi GPU but it is surrounded by a games market that is otherwise mixed in multigp readiness the 79 FPS average of our mixed crossfire configuration plants the cards ahead of a reference 1080 and about 8% behind the 1080 TI gaming acts and that's impressive performance to say the least it's rare but impressive in this instance of sniper leaves for unfortunately it's not just a low-level API thing we didn't see any scaling with Vulcan so that's out nash's of the singularity posted about a 46.5% scaling as you can see in this 4k high test chart so nothing special to speak of there so the short of this is you could crossfire a 580 with the 480 it's not can you probably don't even have to flash the BIOS on the 480 to get it up to a 580 it just will work so that's nice if you wanted to do it you could do it now as far as recommending it this is still not something we recommend we've never recommended crossfire we've never recommended SLI at least not for gaming it doesn't make a whole lot of sense Sniper Elite is a fantastic fringe case if you wanted to hold something high and say look at how great multi GPU is you would choose that game to push the narrative because it looks amazing 100 percent scaling is unheard of for the most part in our testing at least so yes it's happened before it's happening it's like relief or this isn't a common thing as you can see what the other games here if you were to play the games on this bench as you're kind of monthly suite of games that you play around with you'd be basically configuring profiles for each one where Ghost Recon you turn it off for honor you turn it off if you're at 4k maybe leave it on if you have 1080 or something like that sniper leaf or absolutely leave it on but then you look at some other games the x11 titles particularly which there's still a lot of those scaling is mixed it's either maybe equal sometimes there's negative scaling the division showed negative scaling last time we tested crossfire and SLI so this is not something we necessarily recommend is the point but it works and when it works well it's actually really cool so if you're actually considering that's what you need to look at more is what games am I playing and then if we haven't tested it go find a crossfire benchmark for that game and see if it actually scales because if it doesn't and you only play a few games it's a huge waste of money you'd be better off using a single card avoiding the headaches and waiting for either Vega or just getting something like a 1070 although that's kind of a weird upgrade but you get the idea basically you're better off with a single GPU that's more powerful as has pretty much always been the case for gaming at least in the last few years then going with multi-gpu even if you already own one of the cards unless you have a specific game that you play that scales well otherwise look toward Vega or look towards something like a 1080 or something similar to that that's going to be more powerful than a single one of these cards so that's all for this one as always you can subscribe for more go to patreon.com/scishow sells out directly store doc gamers Nexus dot net to grab some of the shirts that we wear on the videos thanks for watching and I'll see you all next time you