AMD Radeon RX 7900 XTX Review & GPU Benchmarks: Gaming, Thermals, Power, & Noise

[Music] today we're reviewing amd's RX 7900 XTX the 1 000 video card rdna3 that's launching while we're on the other side of the world but we're still going to review it and test it in time for this embargo lift because this thing's pretty interesting and this has impressive performance and rasterization and at a thousand dollars to MSRP Nvidia risks becoming seriously displaced by AMD if it continues to ignore the more affordable price class of gpus as it's done with the 4080 which is 1200 bucks the 4090 at least it has the technical best category but at 1600 it's not cheap either now a thousand dollars for amds still isn't cheap clearly there's my 100 7900 XT we're also reviewing today but there's still a dearth of options in the Sub 900 a mid-range category right now and that is where we think the battle will move next in early next year the overcast cloudy sky and the rain is sort of symbolic of the despair that Gamers have felt trying to find affordable gpus for about a generation now actually it's just raining and we didn't have anywhere else to film and I already shouted into a mattress in the Airbnb for the voiceover starting with Rainbow Six Siege test but it's still kind of symbolic let's get started with the 7900 XTX review and see how it does versus nvidia's 4080 especially before that this video is brought to you by Arctic and the liquid freezer 2 liquid coolers now including an argb model in the lineup the liquid freezer Series has been a top performer in our Benchmark for years now and Arctic has continually fine-tuned its products even post launch with things like tits for ryzen argb fans for new flare and new radiator sizes the company also has its brand new MX6 thermal paste on the market now learn more at the links in the description below so quick recap for you first of all the 700 XTX and the 700 XT are the ones that are launching tomorrow the reviews go up a day in advance thousand dollars and 900 respectively for those we already covered the specs and the architecture for rdna 3 as a separate video those are on the channel if you want to check them out we'll link them in the description below rather than rehashing them here if you need to get up to speed on it now the 40 80 12 gigabyte is expected to get a Rebrand and a price drop as well in January that'll likely be around CES timing but for now for today it's not out and we don't know what it's going to cost so it's relevant to the discussion because only the 4080 and 40 90 are out right now in the 4D series our benchmarks Today Show that the XTX is a seriously competitive and rasterization and it's in the maybe close enough territory for some Ray tracing but it gets blown out of the water and others depending on what the testing is like cyberpunk for example so Ray tracing is hit and miss and that's going to be the key Battleground where AMD needs to gain a foothold in the future and as for mechanical design it's something that we're looking at in the tear down specifically but we'll also have pressure tests and flatness testing on this one so let's get straight into the benchmarks we'll start with power testing for the benchmarks we ran transients and we ran nominal power consumption alike transients allow us to observe 100 to 200 microsecond bursts up to 300 microseconds where the power consumption really spikes hard and it's enough to potentially knock power supplies offline if the GPU spikes too high above its nominal draw so we'll be looking at that but first we need to look at the average power consumption to get an idea for what it pulls when playing games or working and averaging over the transients we'll start with the easier one average power consumption in a power virus workload sets us off the RX 7900 XTX ran at 351 Watts during load which is right around its 355 watt TDP that hasn't consuming more power than the RTX 4080 stock card by about 40 Watts or an 11 increase in power draw the 49d pulls 41 more power than the 700 XTX and you'll see in the benchmarks that these numbers often actually align with the performance differences as well generationally the 6900 XT pulled almost the same 350 Watts since a standalone so AMD has stayed approximately within its power envelope over the years now for transient testing in this Benchmark we use an oscilloscope and we clamp the cables to monitor actively for Peak power spikes this allows us to see if there are any poor power design characteristics in the video card it's at the PCB level and the architecture level alike and it helps us see if your power supply might shut down or if you might need to accommodate a higher power Peak draw when building your system inspecting it out in transit testing we entered some serious coil one or as Stone from our team wrote holy coil wine Batman when running the high FPS scenarios we noticed that none of the transients reached one millisecond in duration with most at 500 microseconds or less this is different from the 40 90 where the transients weren't as spiky in the vertical Direction but we noticed that horizontally they were longer in duration this time the worst Peak for the XTX was in Rainbow Six Siege at 1440p where we measured a transient Spike at 92.6 percent higher than nominal power we also observed the 100 microsecond Spike to 725 watts and the loading screen for do maternal although we had difficulty replicating this particular behavior and we don't have a scope shot of it happening that unpredictable nature is worse than predictable reproduction though because it means that it could strike at any time and depending on the power supply quality potentially knock it offline we saw power consumption in the range of 680 to 690 Watts for transient spikes and high FPS gaming scenarios additionally we noticed that the reference 700 xdx has some voltage regulation issues it dropped as low as 11 volts under Peak current loads in some scenarios down from a clean 12.0 and that most likely comes down to the reference board AMD probably won't be out doing its partners for power design delivery and thermals unlike nvidia's internally competitive Fe this is an area where AMD needs to do some power delivery cleanup work now for pressure testing in this one we use a special chemically reactive paper we reassemble the GPU with that paper in between the GPU and the cold plate for the cooler we get all the screws back in it so that we can see if amd's pressure and mounting solution for the cooler to the GPU core has improved over the last couple Generations where they've had some trouble here here's the pressure map overlaid on a photo of the XTX and we'll zoom in closer the bottom mcds are lacking in pressure which makes sense given that the screws are primarily located at the left the right flanks and above there aren't many below the GPU the leftmost edge of the gcd is also lacking in pressure the pressure over all is good there's coverage of the majority of the dies but lacking in the bottom area shows that some different screw locations or a different design for the base plate might resolve this issue this is not a GPU specific problem it's a board plus cooler problem so it's something that the board Partners can work on lack of pressure though doesn't mean lack of contact it just means reduced contact pressure at those points and AMD has improved generationally in massive ways with its reference designs over the last two generations including this one so it's still a move in the right direction here's the flatness testing results the reference 7900 XTX has a median point-to-point depth and microns of 13 with the maximum at 22 as an outlier the median Point depth is overall good on the reference card this is a scenario where it's not the best we've seen but it's also not bad in any way and or is it particularly relevant as difference overall for what we're seeing here time to get into gaming AMD talked a lot about 8K so let's start with that since it's apparently so important to AMD that it was worth spending in a good amount of its performance discussion on front's presentation total Warhammer 3 at 8K is first as you can see we've ensured the scale is appropriate to fit the massive performance we're definitely going to see the 7900 XTX ran at 24 FPS average at 8K allowing the RTX 490 Elite of 38 at 32 FPS average the RTX 4080s 28 FPS average allowed at a lead of 18 so great that's that's exciting and we're glad that Nvidia and AMD have both decided that this point is worth talking about with Nvidia doing it for the 3091st at AMD doing it now and making the same mistakes cyberpunk with FSR and dlss at so-called eight so-called K so-called gaming had the 3090 TI at 13 and the 4080 at 7.4 FPS average it did worse than the 3090 TI when we did validation testing we realized that's because we were Bound by GPU memory and the 309 DTI has more than the 4080 so it helped it here the RX 7900 XTX ran out and impressive show stopping 11.6 FPS average show stopping because you it's one-fifth of a playable frame rate and I think we can all agree that this was a good use of time just like it was an amd's presentation so let's move on to something less useful like 4K and 1440p gaming in Total War 3 at 4K the RX 7900 XTX ran at 91 FPS average with low space similarly to other gpus at 70 and 65. that allows the RTX 4090 a lead of 37 percent at 60 percent higher MSRP or in other words it allows the 700 xdx to achieve at 73 of the performance of the 4090 at 63 percent of the price compared to the RTX 4080 the 7900 XTX is functionally equal in average FPS and is proportional in loads but the XTX has the benefit of being cheaper the RTX 4080 still has one major advantage though it's the most vocal on the nose overbearing cash grab we've seen from Nvidia since they launched sorry unlaunched the 40 80 12 gigabyte for other cards the 700xdx out does the 6800x team that 50 more expensive than the current price or compared to the original MSRP of a thousand dollars at launch the same price the 5700 XT and Vega 64 cards are also plotted for owners of previous flagships but we'll just highlight them and then move forward we're moving on to Total War at 1440p last generation AMD gains disproportionately High Ground at 1440p and 1080p so we'll see if it can repeat that behavior here the RTX 4090 ran at 236 FPS average year leading the 180 FPS average of the 700 XTX by 31 percent reduced from 37 percent at 4K so AMD is still gaining ground at least in this one when below 4K compared to Nvidia that is the RX 7900 XTX has a technical lead of about two percent over the RTX 4080 at a significant cost reduction MSRP to MSRP compared to last gen Parts the 700 XTX leads the 6950 XT by 45 or the 6900 XD by 58 realistic upgrades might include the RX 5700 XT to the 7900 XTX which jumps from 56 FPS average to 180 or Vega 64 at its original 45 FPS average so with either of these you'd see a massive increase to a 7000 series GPU of stay in AMD at 1080p the 700 XTX encroaches on the 4090 in part thanks to approaching the CPU limitations there's only a 22 lead for NVIDIA here that widen with a higher end CPU so it's not a pure comparison but the 7800 XTX versus the 4080 is as the 49d shows us the ceiling is still distant for either of these two cards the 700 XTX is functionally the same as the 4080 but with a technical Victory as for the 6950 XT or the 6900 XT it's about 40 to 45 percent ahead generationally lows haven't changed in scale either between the 4080 and the 700 xdx the XTX still makes more sense in pure rasterization based on at least this game but we need to look at some more shadow of the Tomb Raider at 4K is next this game behaves differently from Total Warhammer and it's useful for seeing another side of performance the 700 XTX ran at 143 FPS average in this one and it didn't have any issues with the lows like most other cards recently tested that has the 4080 at 3.5 percent ahead of the 700 XTX for 20 more MSRP the 4090 is a more substantial 41 percent ahead so at least that one is still kind of defensible as being the literal best there's some value there whereas the 4080 loses the mandatory value argument because it's below the flagship it has to have value and it just doesn't and it never really did versus some previous flagships here's how it lines up the 7900 XTX leads the 6900 XT by 38 the 5700 XT by 170 percent Vega 64 by 249 and the RX 580 because why not by 450 percent it's a high enough percentage that it becomes truly stupid to even compare with percentages and instead just becomes the point of anything as an upgrade so you shouldn't think about it too hard because you're going to notice the difference at 1440p the rx700 XTX ran at 232 FPS average allowing the 4088 technical but irrelevant lead the 4090 is bouncing off the CPU ceiling here so it isn't a valid comparison versus last gen we're looking at a 15 uplift over the 6950 XT this resolution is limiting us with this game so let's just move on to another one in Final Fantasy at 4K the 7900 XTX ran at 148 FPS average and was again within functional equivalence of the RTX 4080 for rasterization and videos and rough shape for its only other current gen offering the 4090 at least keeps its 45 lead that we've seen in other tests allowing Nvidia to retain the technical best Crown until AMD can contest it at some point with something higher at but the real battle honestly is going to be in the mid-range of next year it's not at the high end and it never really has been the high end is nice for Halo marketing but that's not where most people buy even still the 4090 remains an impressive card with a nearly comical lead over nvidia's own 4080. the 700 XTX leads even the 3090 TI by 21 sorry early 30 90 TI buyers but the 3090 TI was around two thousand dollars at launch is double the price of the 700 XTX MSRP today compared to the 700 or so-ish dollar 6900 XT if you buy it today that is the XTX leads by about 42 percent for an equal amount more money but that's versus the modern reduced price not the launch on SRP we tested 1440p but this game is just too easy to run and it was bound on guards even from last generation so let's move on the next one has a very limited data set because we just updated the game for F1 2022 we only recently pushed a few gpus through to start building your charts but we got the 700 XTX on there the 4090 predictably leads it's at 205 if he has average at 4K with a 700 XCX in second place at 164. lows are scaling as expected here and the 4090 has about a 25 lead in this one with the 700xcx managing to outperform the 4080 once again but this time by a more significant margin than previously now at nine percent that's outside of the functional equivalence we saw earlier and it's at 4K no less so that's pretty good spacing the lead over the 6950xt is about 23 percent with the lead over previous Flags rips like the 5700 XT and the Vega 64. back into that territory of anything as an upgrade at 1440p in our limited Suite here for F1 the 4090 leads the 700 XTX by just 9 nine percent it's becoming CPU bound that's useful though because it tells us that the 4K results were not CPU bound yet the 700 XTX architecture scales disproportionately well in this game compared to the previous Tomb Raider and total Warhammer benchmarks the lead of XTX over the 4080 it's about 9.5 percent here growing from nine percent previously in Rainbow Six Siege at 4K the 4090 ran at 355 FPS average to the xt-axis to 70. supporting a 31 lead in the 40 90. the XTX is almost exactly tied with the 4080 for average FPS here technically speaking the 700xcx is running disproportionately lower one percent 0.1 metrics than its RTX 40 flanks with the 4080 supporting better frame time consistency than the XTX but again that's technically speaking realistically speaking these differences aren't that meaningful when factoring in the inherently wider error on low calculation and the fact that these numbers are so high anyway there's just no tangible change the next card down is the 3090 TI About 30 FPS behind the 4080 and the 700 XTX obviously owners of a 6950 XT 6900 xt380 or a similar card have no need to upgrade at this point those are all still great devices and you really shouldn't be doing single generation upgrades unless there's some serious need but jumping from a 20 series or a 5700 XT would net pretty large improvements at 1440p the 4090 leads the 700 XTX by 17 percent we know our ceilings about 677 FPS average here but even though we have some room between the 624 and the 677 number some of the higher performing scenarios in this test still hit that limiter the 4080 and the XTX are Unbound though and that's good for this comparison the XTX outguns the 4080 by 30 FPS average which is a six percent lead that's in line with some previous tests the XCX is compelling as a more value-oriented rasterization focused solution compared to the clearly too expensive 4080 it helps to look at something of Vulcan too so he ran strange Brigade as a function test strange Brigade is extremely well optimized it has Vulcan and dx12 and it's an interesting Benchmark for evaluating performance even if it's not that common of a game or ever was strange Brigade had the 49 he had 291 FPS average with the 4080 and about 202 FPS average that 4080 number is about the same as the 700 XTX once again establishing that equivalence is the worst case scenario for the 700 XCX in general rasterization performance thus far that's a fine place to be when your card is cheaper in Horizon zero Dawn at 4K for the last purely rasterization test the rx700 XTX ran at 121 FPS average sitting just behind the RTX 4080 the 4090 leads at 164 FPS average which is a 36 advantage to the 700 XTX as for the rest the next closest card is the 39 DTI at 107 FPS average and then it's the 6950 XT at 99. now we're moving on to Ray tracing tests these are our last gaming benchmarks for this suite and they're really difficult to run now because for Ray tracing you have not just RT performance but you have the FSR or rsr or dlss or X ESS Solutions where these technologies will allow super resolution or super sampling to achieve a sort of fake to higher resolution and we need to test with all of those depending on which game it is in cyberpunk 2077 for the first one at 4K with Ray tracing we're using the equivalent dlss and FSR Technologies between AMD and Nvidia so neither is rendering at the pure unsupported 4K and both are supported by their super resolution Tech but we've configured their quality settings so that they are scaling from equivalent starting points that makes it like for like comparable the rx7900 XTX loses a lot of ground here falling down to the performance level of an RTX 3090 or a TI at best the RTX 4080 leads by 40 percent a staggering gain compared to the rasterization tests and the 490 leads by 93 percent it's not even close generationally over the 6950 XT Andy has at least gained 46 percent in RT performance which is a lot but it's just not enough to get anywhere close to Nvidia if you care about RT performance in this game when we went to double check the performance numbers we noticed that amd's own reviewer reference data suggested that the 7500 XTX would be significantly worse than cyberpunk with rt on so they got that right this is expected Behavior at least at 1440p the RX 7900 XTX ran at 78 FPS average in the same test which has it between the 3090 and the ti this establishes another large lead for the 40 series at 41 for the 4080 and 75 percent for the 490. now that the cyberpunk bloodbath is over we can move on to Tomb Raider which is an easier RT test it only implements raytrained Shadows which means that it's still primarily rasterization driven with some Ray tracing on top and that should be beneficial for AMD but it's also representative of a lot of the earlier titles for RT at 4K the 700 xdx is still behind the RTX 4080. the 4080 leads by 11 this time which is outside the range we saw for purely rasterization but not as large a lead as heavier games like cyberpunk the 4090 runs at 120 FPS average so the xdx is 78 and the 390 runs at 60fps average again to the same 78 for the XTX so at least against the last gen the AMD card looks good but it does lose ground with rt at 1440p the 700 XTX ran at 140 FPS average and allowed another 11 lead to the 4080. the results scale the same here as they did at 4K despite amd's lower resolution performance swings we saw in some non-rt titles control is up next back to more heavily ranged raised games our community voted this as one of their top played RT titles alongside cyberpunk in this one the RX 7900 XTX ran at 45 FPS average at 4K with the 4080 at 52 FPS average a lead of 14 for the 4080 the 40 90s 76 FPS average establishes it 70 ahead which is one of the few times that the price is increasing roughly in step with the performance increase and it doesn't fall into a negative scale for this comparison at 1440p control positioned the 40 80 13 ahead of the 7900 XTX the 4080 really needs to come down in price to be more competitive here because only having a lead sometimes in raid tracing isn't good enough given the overall performance of the 700 XTX even remotely we wanted to bring you the full depth we bring you for any other GPU reviews so we also ran thermal testing for this we used a mix of gpuz and Hardware info engineering Edition to collect internal sensor information we typically also mount thermocouples to the board for further validation however since I'm in Taiwan right now we'll rely on the software it's normally pretty accurate though here's a chart of results at equilibrium the GPU temperature aligns with amd's software and with GPU Z and Hardware info and it plots at 64 degrees Celsius for the edge temperature or the lowest of the GPU measurement the GPU has multiple hotspots now with its chiplet design with the gcd at 80 degrees Celsius and the mcds all around 64 to 67 degrees Celsius these numbers are acceptable the mcds especially the 80 degree number though is pushing the warm side and you'll have some clock loss in a hotter case so make sure you either buy a more powerful partner model cooler or you get a really good case although amd's new cooler is better than its past coolers this is by no means one that we'd classify as impressive it's at best pretty standard Partners will still out do AMD here which is fine that's sort of their job and how it's supposed to work memory is more of a problem the memory hotspot at 84 degrees Celsius when at 21 degrees Celsius ambient means that we have very limited scaling room and warmer case configurations because the fan curve follows GPU temperature it won't compensate on memory thermals independently this can and has caused problems with cards in the past the thermistor and the fan blade although a very cool idea and one we're excited about isn't working in an ideal way it's detected at 33 degrees Celsius even though the actual Inlet temperature if monitored by an external probe is 21 degrees Celsius that's because the thermistor is so close to the warm motor in The Fan Hub and the heat on the fin stack that it's picking up more GPU temperature than anything else in terms of ambient an approximate 10 degree Delta of true ambient open air to internal GPU ambient is enough to know how at least to use this number in your case this can still be useful as a number for configuring your case bands because it's ultimately all relative just know that it's not an absolute indicator of true case ambient in the region of Inlet for the GPU but as stated if it changes when you move the fans around that's still useful you just need to look at the Deltas not the absolute amd's cooler is okay it's better than the last two generations it's just again not something that we'd say is anything beyond acceptable for a reference model this one follows the older trend of buying partner models to see which ones can really knock it out of the park if they can stick close to MSRP and if they can't well then at least this one's not overheating as for the noise that's more of a problem aside from the coil line issues this card also ramps the fan extremely hard and very early in the workload fan hysteresis is not properly configured in vbios because you can see it spikes from 0 to 700 RPM nearly instantaneously when the load starts it then drops a little bit and then within just 60 more seconds it hits its Max fan speed at 1700 RPM this type of Rapid ramp is very noticeable to the end user and is too aggressive because in scenarios where the load might not sustain more than a minute it's not enough to justify the sudden noise hike and then drop and these should just adjust the fan wrap curve to be finer we also ran frequency testing in fire strike Ultra with custom settings we observe the frequency in the range of 2600 to 2720 megahertz it's spikier than we see on the modern Nvidia hardware and in this case that's because AMD is running up against both power and voltage reliability limits that's normal it has to hit some kind of limit it's just a question of which limiter it is and it's better to hit something like vrel or power than it is to hit thermal so Behavior here is what we classify as expected so the 7900 XTX after looking at all that it kicks Nvidia in the ass for rasterization really it's cheaper at worst it's about equivalent which is a really good spot to be if you're already cheaper and Nvidia is kind of a wake-up call for them because they haven't ever sold out of 4080s they've been on the shelves in at least a few places ever since launch and they still are today so Nvidia needs to drop the prices and it really needs to get another card out at the cheaper end as well to start fighting AMD harder here's a quick recap of some of the numbers we looked at today though so we observed about 350 Watts power draw nominally on the XTX which is fair it's not particularly high it's not particularly impressive either on the efficiency side but it's a big generational Improvement in the power envelope over our DNA too a rasterization had the 700 xdx often equal with the RTX 4080 they went back and forth in single digit percentages and the ray Trace testing was more of a blowout victory for NVIDIA and cyberpunk which has known issues for amd's cards per AMD directly we validated FSR was working and it still performed poorly even with that for control however in Tomb Raider we observed that the 4080s lead was significantly reduced and more comparable between them and he still isn't winning an RT performance and it does need to catch up soon but it's closed the Gap considerably over the years now despite the pretty competitive rasterization performance one area where AMD really needs to do some work is its power management on the board the 100 microsecond spiked the 725 Watts was egregious but not nearly as bad as the 3080 was or the 30 90 gigabyte Master where it was like sometimes two and a half X the voltage drop under a high frame rate scenario was also uncharacteristically high for this type of card that's on the card side it's not the power supply Ai and it's something that is probably mostly unique to the reference 700 xdx so board Partners can fix this one between the coil line we demonstrated the voltage regulation and especially the aggressive V bios fan ramp with improper hysteresis that causes noisy fan Behavior we think you should probably skip the reference model and get a partner model instead so that's not to say the GPU itself should be skipped but there will be better cards that should be at or just barely above MSRP in theory we know that's going to hurt some feelings because people have a really strong attachment to AMD reference cards for some reason but that's just the truth of it so a couple things here as we wrap up first of all again we're obviously filming this on the other side of the world it's cool that we can do this we're back home Patrick's been testing it and we've got vitaliano team editing most of this review remotely but I'm analyzing the data from here and that means we're doing our best with I mean still the same amount of data we always have but uh there are some some angles we want to look at once I get back over to the us so we can evaluate the card further so we'll have some follow-ups there if you feel like something's missing from this review it's coming up soon it's just we need to well I need to physically get more hands on with the card once we're back so just to get that out of the way now with what we have here we think the 700x tx's value is better than the 4080 for non-ray traced gaming so like rasterization it's pretty competitive and a lot of the worst case scenarios it's roughly equal you lose by maybe single digit percentages so they're kind of plus or minus eight percent between the two cards depending what it is not really that big of a difference but two hundred dollars is a big difference and even though the 700 XTX might not commonly be available for MSRP neither is the 4080. so they're kind of on equal footing in that regard as well we think that this is not a huge blowout victory for AMD because so Ray tracing they're still behind there and then there's still just not anything in that mid-range affordable where gpus should probably cost for gaming category of the market right now the 6900 XT for example we just checked quickly they're still out there for like 650 to 700 bucks it's worth considering if you are cash constrained to the point that an extra 200 300 is simply too much uh so the 3080 is also closer to 700 depending on where you look these days and that's still worth considering as well but there's a clear lack of cards in the affordable market and that's a problem as for the 4090 it's still uncontested at the high end so AMD doesn't have anything to fight there and the one advantage you always have when you are the literal best as a video card is that you're the literal best so the 4090 has that angle to fight on where it doesn't need to compete on price as much as quality because well that's how it's positioned but the 4080 is the one that needs to it loses here and it does need to fight a little harder Andy's done well to push on video here uh next time we'd really like it if AMD didn't talk about 8K gaming because they have a lot of actually meaningful advantages even if only a price that they could be marketing instead and those are much stronger points than pretending that 8K gaming is a thing anyone cares about when the vast majority of the market isn't even on 4K screens yet it didn't perform that well either so other than that the card looks pretty competitive we think it's worth considering and we'll be benchmarking it further once we're back home but we have a 700 XT review and an XTX teardown coming up next so check back for those subscribe for more as always you go to store.camersexis.net or patreon.com Gamers Nexus to help support us directly because we're doing these reviews even when we're traveling and when I have to shout into a mattress to do a moldy mattress that's leaning on the wall that I probably shouldn't be breathing near I have to shout into that to do the voiceover so thanks for watching we'll see you all next time

Channel: Gamers Nexus

Views: 1,000,061

Keywords: gamersnexus, gamers nexus, computer hardware, amd rx 7900 xtx review, amd rx 7900 xtx benchmarks, amd rx 7900 xtx reference thermals, amd rx 7900 xtx vs nvidia rtx 4080, rtx 4090 vs amd rx 7900 xtx, amd rx 7900xtx, amd 7900xtx, amd rx 7900xt, amd rx 7900 xt, amd rx 7900 xtx vs rtx 4090, best gpus 2022

Id: We71eXwKODw

Length: 31min 36sec (1896 seconds)

Published: Mon Dec 12 2022