Our GPU benchmarks hierarchy ranks all the present and former technology graphics playing cards by efficiency, together with the entire greatest graphics playing cards. Whether or not it is enjoying video games or doing high-end inventive work like 4K video enhancing, your graphics card sometimes performs the most important position in figuring out efficiency, and even the greatest CPUs for Gaming take a secondary position.
We have revamped our GPU testbed and up to date all of our benchmarks for 2022, and are actually completed retesting practically each graphics card from the previous a number of generations, plus some even older GPUs. Our full GPU hierarchy utilizing conventional rendering comes first, and beneath that we’ve got our ray tracing GPU benchmarks hierarchy. These in fact require a ray tracing succesful GPU so solely AMD’s RX 6000-series and Nvidia’s RTX playing cards are current.
Our newest addition to the tables is AMD’s Radeon RX 6400, a “true finances” graphics card. About the one devoted GPU’s we’re at the moment lacking are the brand new GeForce GTX 1630 (trace: not good!), and Intel’s Arc A380 (higher video codec assist, in all probability sooner than the 1630 as effectively, however slower than RX 6400).
Beneath our fundamental tables, you will discover our 2020–2021 benchmark suite, which has the entire earlier technology GPUs operating our older check suite operating on a Core i9-9900K testbed. We even have the legacy GPU hierarchy (with out benchmarks) on the backside of the article for reference functions.
The next tables kind every little thing solely by our performance-based GPU gaming benchmarks, at 1080p “extremely” for the primary suite and at 1080p “medium” for the DXR suite. Components together with value, graphics card energy consumption, general effectivity, and options aren’t factored into the rankings right here. We have now added the most recent AMD RX 6950 XT, RX 6750 XT, and RX 6650 XT to the charts, together with the RTX 3090 Ti. All of these use customized third-party AIC fashions with manufacturing unit overclocks, so they are not fairly on the identical enjoying discipline as the remainder of the reference playing cards.
We’re nonetheless ready for the Intel Arc Alchemist desktop launch, however in any other case we’re now trying ahead to the latter a part of the 12 months when the Nvidia Ada and AMD RDNA 3 architectures ought to arrive. Really, the most recent rumors counsel Ada may land as early as August (however in all probability not), so we’d not have that lengthy to attend.
We have switched to a brand new Alder Lake Core i9-12900K testbed, modified up our check suite, and retested the entire previous a number of generations of GPUs. Now let’s hit the benchmarks and tables.
Graphics Card Offers
The element shortages and skyrocketing demand which were killing graphics card availability and GPU costs appear to be coming to an finish. You could find lots of the newest GPUs at costs approaching and even beating MSRP.
One different to purchasing a brand new graphics card is to buy a prebuilt desktop that has your GPU of alternative inside. You may both pull the cardboard and resell the PC or, select a gaming desktop you actually like and make it a whole improve. Listed below are some fast hyperlinks that can assist you discover a graphics card deal on a system should you go this route, together with a direct hyperlink so you’ll be able to see what the most recent costs are on the open market.
GPU Benchmarks Rating 2022
For our newest benchmarks, we check all GPUs at 1080p medium and 1080p extremely, and type the desk by the 1080p extremely outcomes. The place it is smart, we additionally check at 1440p extremely and 4K extremely. All the scores are scaled relative to the top-ranking 1080p extremely card, which in our new suite is the Radeon RX 6950 XT (at the very least at 1080p and 1440p).
You may as well see the above abstract chart displaying the relative efficiency of the playing cards we have examined throughout the previous a number of generations of {hardware} at 1080p extremely. There are just a few lacking choices (e.g., the GTX 1650 GDDR6, GT 1030, and a number of other Titan playing cards), however in any other case it is practically full now.
The eight video games we’re utilizing for our normal GPU benchmarks hierarchy are Borderlands 3 (DX12), Far Cry 6 (DX12), Flight Simulator (DX12), Forza Horizon 5 (DX12), Horizon Zero Daybreak (DX12), Crimson Useless Redemption 2 (Vulkan), Whole Warfare Warhammer 3 (DX11), and Watch Canine Legion (DX12). The fps rating is the geometric imply (equal weighting) of the eight video games.
Graphics Card | 1080p Extremely | 1080p Medium | 1440p Extremely | 4K Extremely | Specs |
---|---|---|---|---|---|
Radeon RX 6950 XT | 100.0% (137.3fps) | 100.0% (190.1fps) | 100.0% (115.4fps) | 100.0% (70.3fps) | Navi 21, 5120 shaders, 2310MHz, 16GB GDDR6@18Gbps, 576GB/s, 335W |
GeForce RTX 3090 Ti | 96.5% (132.4fps) | 94.8% (180.1fps) | 98.7% (113.9fps) | 107.6% (75.7fps) | GA102, 10752 shaders, 1860MHz, 24GB GDDR6X@21Gbps, 1008GB/s, 450W |
Radeon RX 6900 XT (opens in new tab) | 94.5% (129.7fps) | 97.1% (184.6fps) | 91.4% (105.5fps) | 89.7% (63.1fps) | Navi 21, 5120 shaders, 2250MHz, 16GB GDDR6@16Gbps, 512GB/s, 300W |
GeForce RTX 3090 (opens in new tab) | 92.2% (126.6fps) | 93.7% (178.1fps) | 92.3% (106.5fps) | 97.8% (68.8fps) | GA102, 10496 shaders, 1695MHz, 24GB GDDR6X@19.5Gbps, 936GB/s, 350W |
GeForce RTX 3080 12GB (opens in new tab) | 90.7% (124.5fps) | 93.8% (178.2fps) | 90.1% (104.0fps) | 94.3% (66.3fps) | GA102, 8960 shaders, 1845MHz, 12GB GDDR6X@19Gbps, 912GB/s, 400W |
Radeon RX 6800 XT (opens in new tab) | 90.0% (123.5fps) | 94.2% (179.1fps) | 86.5% (99.8fps) | 83.2% (58.5fps) | Navi 21, 4608 shaders, 2250MHz, 16GB GDDR6@16Gbps, 512GB/s, 300W |
GeForce RTX 3080 Ti (opens in new tab) | 89.9% (123.4fps) | 92.0% (174.9fps) | 89.6% (103.4fps) | 94.5% (66.5fps) | GA102, 10240 shaders, 1665MHz, 12GB GDDR6X@19Gbps, 912GB/s, 350W |
GeForce RTX 3080 (opens in new tab) | 84.7% (116.3fps) | 91.2% (173.4fps) | 82.8% (95.5fps) | 86.2% (60.6fps) | GA102, 8704 shaders, 1710MHz, 10GB GDDR6X@19Gbps, 760GB/s, 320W |
Radeon RX 6800 (opens in new tab) | 80.7% (110.7fps) | 90.9% (172.7fps) | 75.9% (87.5fps) | 71.9% (50.6fps) | Navi 21, 3840 shaders, 2105MHz, 16GB GDDR6@16Gbps, 512GB/s, 250W |
GeForce RTX 3070 Ti (opens in new tab) | 75.8% (104.1fps) | 85.4% (162.4fps) | 71.5% (82.6fps) | 66.6% (46.8fps) | GA104, 6144 shaders, 1770MHz, 8GB GDDR6X@19Gbps, 608GB/s, 290W |
Radeon RX 6750 XT (opens in new tab) | 73.7% (101.2fps) | 88.4% (168.0fps) | 65.3% (75.4fps) | 59.6% (41.9fps) | Navi 22, 2560 shaders, 2600MHz, 12GB GDDR6@18Gbps, 432GB/s, 250W |
Titan RTX (opens in new tab) | 73.6% (101.0fps) | 83.2% (158.2fps) | 69.7% (80.5fps) | 68.7% (48.3fps) | TU102, 4608 shaders, 1770MHz, 24GB GDDR6@14Gbps, 672GB/s, 280W |
GeForce RTX 3070 (opens in new tab) | 72.7% (99.8fps) | 82.9% (157.7fps) | 67.2% (77.5fps) | 61.4% (43.2fps) | GA104, 5888 shaders, 1725MHz, 8GB GDDR6@14Gbps, 448GB/s, 220W |
GeForce RTX 2080 Ti (opens in new tab) | 69.9% (96.0fps) | 79.8% (151.6fps) | 65.3% (75.3fps) | 63.4% (44.6fps) | TU102, 4352 shaders, 1545MHz, 11GB GDDR6@14Gbps, 616GB/s, 250W |
Radeon RX 6700 XT (opens in new tab) | 69.8% (95.8fps) | 84.1% (159.8fps) | 61.3% (70.8fps) | 56.1% (39.4fps) | Navi 22, 2560 shaders, 2581MHz, 12GB GDDR6@16Gbps, 384GB/s, 230W |
GeForce RTX 3060 Ti (opens in new tab) | 66.7% (91.5fps) | 78.8% (149.7fps) | 60.4% (69.7fps) | GA104, 4864 shaders, 1665MHz, 8GB GDDR6@14Gbps, 448GB/s, 200W | |
GeForce RTX 2080 Tremendous (opens in new tab) | 61.9% (84.9fps) | 72.5% (137.8fps) | 56.3% (64.9fps) | 49.1% (34.5fps) | TU104, 3072 shaders, 1815MHz, 8GB GDDR6@15.5Gbps, 496GB/s, 250W |
GeForce RTX 2080 (opens in new tab) | 59.9% (82.2fps) | 70.1% (133.1fps) | 54.1% (62.4fps) | TU104, 2944 shaders, 1710MHz, 8GB GDDR6@14Gbps, 448GB/s, 215W | |
Radeon RX 6650 XT (opens in new tab) | 58.2% (79.9fps) | 72.8% (138.4fps) | 49.2% (56.7fps) | Navi 23, 2048 shaders, 2635MHz, 8GB GDDR6@18Gbps, 280GB/s, 180W | |
Radeon RX 6600 XT (opens in new tab) | 56.9% (78.1fps) | 71.8% (136.5fps) | 47.6% (54.9fps) | Navi 23, 2048 shaders, 2589MHz, 8GB GDDR6@16Gbps, 256GB/s, 160W | |
GeForce RTX 2070 Tremendous (opens in new tab) | 55.7% (76.4fps) | 65.3% (124.1fps) | 49.8% (57.4fps) | TU104, 2560 shaders, 1770MHz, 8GB GDDR6@14Gbps, 448GB/s, 215W | |
Radeon RX 5700 XT (opens in new tab) | 53.7% (73.7fps) | 66.2% (125.8fps) | 46.2% (53.3fps) | 41.6% (29.3fps) | Navi 10, 2560 shaders, 1905MHz, 8GB GDDR6@14Gbps, 448GB/s, 225W |
GeForce RTX 3060 (opens in new tab) | 51.1% (70.2fps) | 62.5% (118.8fps) | 45.6% (52.6fps) | GA106, 3584 shaders, 1777MHz, 12GB GDDR6@15Gbps, 360GB/s, 170W | |
Radeon VII (opens in new tab) | 50.8% (69.7fps) | 60.0% (114.0fps) | 45.9% (53.0fps) | 44.7% (31.4fps) | Vega 20, 3840 shaders, 1750MHz, 16GB HBM2@2.0Gbps, 1024GB/s, 300W |
GeForce RTX 2070 (opens in new tab) | 49.5% (67.9fps) | 58.2% (110.7fps) | 44.2% (51.0fps) | TU106, 2304 shaders, 1620MHz, 8GB GDDR6@14Gbps, 448GB/s, 175W | |
Radeon RX 6600 (opens in new tab) | 48.6% (66.7fps) | 62.0% (117.8fps) | 40.0% (46.1fps) | Navi 23, 1792 shaders, 2491MHz, 8GB GDDR6@14Gbps, 224GB/s, 132W | |
GeForce GTX 1080 Ti (opens in new tab) | 48.5% (66.5fps) | 58.2% (110.6fps) | 43.6% (50.3fps) | 42.0% (29.5fps) | GP102, 3584 shaders, 1582MHz, 11GB GDDR5X@11Gbps, 484GB/s, 250W |
GeForce RTX 2060 Tremendous (opens in new tab) | 47.4% (65.1fps) | 55.7% (105.9fps) | 41.8% (48.2fps) | TU106, 2176 shaders, 1650MHz, 8GB GDDR6@14Gbps, 448GB/s, 175W | |
Radeon RX 5700 (opens in new tab) | 47.2% (64.8fps) | 58.5% (111.3fps) | 40.9% (47.2fps) | Navi 10, 2304 shaders, 1725MHz, 8GB GDDR6@14Gbps, 448GB/s, 180W | |
Radeon RX 5600 XT (opens in new tab) | 42.3% (58.1fps) | 52.9% (100.6fps) | 36.4% (42.0fps) | Navi 10, 2304 shaders, 1750MHz, 8GB GDDR6@14Gbps, 336GB/s, 160W | |
Radeon RX Vega 64 (opens in new tab) | 41.4% (56.8fps) | 49.6% (94.3fps) | 36.1% (41.6fps) | 33.4% (23.5fps) | Vega 10, 4096 shaders, 1546MHz, 8GB HBM2@1.89Gbps, 484GB/s, 295W |
GeForce RTX 2060 (opens in new tab) | 40.2% (55.2fps) | 50.9% (96.8fps) | 33.6% (38.7fps) | TU106, 1920 shaders, 1680MHz, 6GB GDDR6@14Gbps, 336GB/s, 160W | |
GeForce GTX 1080 (opens in new tab) | 38.7% (53.1fps) | 47.3% (90.0fps) | 34.2% (39.4fps) | GP104, 2560 shaders, 1733MHz, 8GB GDDR5X@10Gbps, 320GB/s, 180W | |
GeForce RTX 3050 (opens in new tab) | 37.5% (51.4fps) | 47.0% (89.4fps) | 32.6% (37.6fps) | GA106, 2560 shaders, 1777MHz, 8GB GDDR6@14Gbps, 224GB/s, 130W | |
GeForce GTX 1070 Ti (opens in new tab) | 37.2% (51.1fps) | 45.1% (85.8fps) | 32.9% (37.9fps) | GP104, 2432 shaders, 1683MHz, 8GB GDDR5@8Gbps, 256GB/s, 180W | |
Radeon RX Vega 56 (opens in new tab) | 36.9% (50.6fps) | 44.4% (84.4fps) | 32.0% (37.0fps) | Vega 10, 3584 shaders, 1471MHz, 8GB HBM2@1.6Gbps, 410GB/s, 210W | |
GeForce GTX 1070 (opens in new tab) | 32.6% (44.8fps) | 33.7% (64.0fps) | 33.6% (38.8fps) | GP104, 1920 shaders, 1683MHz, 8GB GDDR5@8Gbps, 256GB/s, 150W | |
GeForce GTX 1660 Tremendous (opens in new tab) | 32.4% (44.4fps) | 43.6% (82.8fps) | 27.3% (31.5fps) | TU116, 1408 shaders, 1785MHz, 6GB GDDR6@14Gbps, 336GB/s, 125W | |
GeForce GTX 1660 Ti (opens in new tab) | 32.0% (43.9fps) | 43.1% (81.9fps) | 27.4% (31.6fps) | TU116, 1536 shaders, 1770MHz, 6GB GDDR6@12Gbps, 288GB/s, 120W | |
GeForce GTX 1660 (opens in new tab) | 29.1% (39.9fps) | 39.5% (75.1fps) | 24.7% (28.5fps) | TU116, 1408 shaders, 1785MHz, 6GB GDDR5@8Gbps, 192GB/s, 120W | |
Radeon RX 5500 XT 8GB (opens in new tab) | 29.0% (39.8fps) | 38.2% (72.6fps) | 24.7% (28.5fps) | Navi 14, 1408 shaders, 1845MHz, 8GB GDDR6@14Gbps, 224GB/s, 130W | |
Radeon RX 590 (opens in new tab) | 28.7% (39.4fps) | 36.1% (68.6fps) | 25.2% (29.1fps) | Polaris 30, 2304 shaders, 1545MHz, 8GB GDDR5@8Gbps, 256GB/s, 225W | |
GeForce GTX 980 Ti (opens in new tab) | 26.1% (35.9fps) | 32.9% (62.6fps) | 23.1% (26.7fps) | GM200, 2816 shaders, 1075MHz, 6GB GDDR5@7Gbps, 336GB/s, 250W | |
Radeon R9 Fury X (opens in new tab) | 25.8% (35.4fps) | 33.9% (64.4fps) | Fiji, 4096 shaders, 1050MHz, 4GB HBM2@2Gbps, 512GB/s, 275W | ||
Radeon RX 580 8GB (opens in new tab) | 25.8% (35.3fps) | 32.5% (61.7fps) | 22.5% (26.0fps) | Polaris 20, 2304 shaders, 1340MHz, 8GB GDDR5@8Gbps, 256GB/s, 185W | |
Radeon RX 5500 XT 4GB (opens in new tab) | 24.4% (33.5fps) | 35.2% (66.9fps) | Navi 14, 1408 shaders, 1845MHz, 4GB GDDR6@14Gbps, 224GB/s, 130W | ||
GeForce GTX 1650 Tremendous (opens in new tab) | 24.2% (33.2fps) | 35.7% (67.9fps) | 19.9% (23.0fps) | TU116, 1280 shaders, 1725MHz, 4GB GDDR6@12Gbps, 192GB/s, 100W | |
GeForce GTX 1060 6GB (opens in new tab) | 23.5% (32.2fps) | 30.5% (58.0fps) | 20.0% (23.0fps) | GP106, 1280 shaders, 1708MHz, 6GB GDDR5@8Gbps, 192GB/s, 120W | |
Radeon RX 6500 XT (opens in new tab) | 22.2% (30.4fps) | 34.4% (65.4fps) | 15.6% (18.0fps) | Navi 24, 1024 shaders, 2815MHz, 4GB GDDR6@18Gbps, 144GB/s, 107W | |
Radeon R9 390 (opens in new tab) | 21.7% (29.8fps) | 26.9% (51.2fps) | Grenada, 2560 shaders, 1000MHz, 8GB GDDR5@6Gbps, 384GB/s, 275W | ||
GeForce GTX 980 (opens in new tab) | 21.1% (28.9fps) | 28.2% (53.7fps) | GM204, 2048 shaders, 1216MHz, 4GB GDDR5@7Gbps, 256GB/s, 165W | ||
Radeon RX 570 4GB (opens in new tab) | 20.6% (28.3fps) | 28.2% (53.6fps) | 17.3% (20.0fps) | Polaris 20, 2048 shaders, 1244MHz, 4GB GDDR5@7Gbps, 224GB/s, 150W | |
GeForce GTX 1060 3GB (opens in new tab) | 20.2% (27.8fps) | 27.7% (52.6fps) | GP106, 1152 shaders, 1708MHz, 3GB GDDR5@8Gbps, 192GB/s, 120W | ||
GeForce GTX 1650 (opens in new tab) | 19.4% (26.6fps) | 26.9% (51.1fps) | TU117, 896 shaders, 1665MHz, 4GB GDDR5@8Gbps, 128GB/s, 75W | ||
GeForce GTX 970 (opens in new tab) | 19.3% (26.5fps) | 25.9% (49.1fps) | GM204, 1664 shaders, 1178MHz, 4GB GDDR5@7Gbps, 256GB/s, 145W | ||
Radeon RX 6400 (opens in new tab) | 17.2% (23.7fps) | 27.3% (52.0fps) | Navi 24, 768 shaders, 2321MHz, 4GB GDDR6@16Gbps, 128GB/s, 53W | ||
GeForce GTX 780 (opens in new tab) | 16.0% (22.0fps) | 20.3% (38.5fps) | GK110, 2304 shaders, 900MHz, 3GB GDDR5@6Gbps, 288GB/s, 230W | ||
GeForce GTX 1050 Ti (opens in new tab) | 14.4% (19.8fps) | 20.0% (38.0fps) | GP107, 768 shaders, 1392MHz, 4GB GDDR5@7Gbps, 112GB/s, 75W | ||
GeForce GTX 1050 (opens in new tab) | 10.8% (14.8fps) | 15.7% (29.8fps) | GP107, 640 shaders, 1455MHz, 2GB GDDR5@7Gbps, 112GB/s, 75W | ||
Radeon RX 560 4GB (opens in new tab) | 10.8% (14.8fps) | 16.8% (31.8fps) | Baffin, 1024 shaders, 1275MHz, 4GB GDDR5@7Gbps, 112GB/s, 60-80W | ||
Radeon RX 550 4GB (opens in new tab) | 0.7% (1.0fps) | 10.3% (19.6fps) | Lexa, 640 shaders, 1183MHz, 4GB GDDR5@7Gbps, 112GB/s, 50W |
*: GPU could not run all exams, so the general rating is barely skewed at 1080p extremely.
Our up to date check suite and testbed favor AMD’s GPUs barely, notably at 1080p and even 1440p — which is probably another reason the RTX 3090 Ti exists, because it largely retakes the throne in any respect resolutions, although the brand new 6950 XT reclaims prime honors. Needless to say we’re not together with any ray tracing or DLSS leads to the above desk, as we intend to make use of the identical check suite with the identical settings on all present and former technology graphics playing cards.
AMD’s RX 6950 XT does not massively enhance efficiency, but it surely’s sufficient to make up the hole with the 3090 Ti, and it does so whereas costing over 40% much less. AMD additionally wins, fairly simply, within the efficiency per watt metric. Stepping down the record, the 3090 and 3080 12GB — an overclocked MSI mannequin, since there aren’t any reference 3080 12GB playing cards — place simply forward of the 6800 XT, adopted by the 3080 Ti. The RX 6800 additionally beats the RTX 3070 Ti, whereas the RTX 3070 and RX 6700 XT are successfully tied.
The rankings favor AMD much less on the decrease portion of the chart, with the RTX 3060 and RX 6600 additionally tied, and the RTX 3050 simply eclipses the RX 6500 XT — not that it is tough to take action, as each the 4GB and 8GB RX 5500 XT additionally beat AMD’s newest finances providing.
Turning to the earlier technology GPUs, the RTX 20-series and GTX 16-series chips find yourself scattered all through the outcomes, together with the RX 5000-series. The overall rule of thumb is that you just get one or two “mannequin upgrades” with the newer structure, so for instance the RTX 2080 Tremendous is available in just under the RTX 3060 Ti, whereas the RX 5700 XT lands just a few % behind the RX 6600 XT.
Return far sufficient, and you may see how trendy video games at extremely settings severely punish playing cards that do not have greater than 4GB VRAM. We have been saying for just a few years now that 4GB is simply scraping by, and 6GB or extra is fascinating. The GTX 1060 3GB, GTX 1050, and GTX 780 truly didn’t run a few of our exams, which skews their outcomes a bit, despite the fact that they do higher at 1080p medium.
Now let’s change over to the ray tracing hierarchy.
Ray Tracing GPU Benchmarks Rating 2022
Enabling ray tracing, notably with demanding video games like these we’re utilizing in our DXR check suite, could cause framerates to drop off a cliff. We’re testing with “medium” and “extremely” ray tracing settings. Medium means utilizing medium graphics settings however turning on ray tracing results (set to “medium” if that is an choice), whereas extremely activates the entire RT choices at kind of most high quality.
As a result of ray tracing is a lot extra demanding, we’re sorting these outcomes by the 1080p medium scores. That is additionally as a result of the RX 6500 XT mainly cannot deal with ray tracing even at these settings, and testing at something greater than 1080p medium can be fruitless. We have completed testing all the present ray tracing succesful GPUs, although there can be extra sooner or later. We’re undoubtedly curious to see if Intel’s Arc GPUs can do any higher than the RX 6500 XT, and suspect the reply could be “nope” on the decrease tier A300 sequence.
The six ray tracing video games we’re utilizing are Shiny Reminiscence Infinite, Management Final Version, Cyberpunk 2077, Fortnite, Metro Exodus Enhanced, and Minecraft — all of those use the DirectX 12 / DX12 Final API. The fps rating is the geometric imply (equal weighting) of the six video games, and the share is scaled relative to the quickest GPU within the record, which on this case is the GeForce RTX 3090 Ti.
Graphics Card | 1080p Medium | 1080p Extremely | 1440p Extremely | 4K Extremely | Specs |
---|---|---|---|---|---|
GeForce RTX 3090 Ti (opens in new tab) | 100.0% (118.2fps) | 100.0% (84.4fps) | 100.0% (57.2fps) | 100.0% (29.1fps) | GA102, 10752 shaders, 1860MHz, 24GB GDDR6X@21Gbps, 1008GB/s, 450W |
GeForce RTX 3090 (opens in new tab) | 91.7% (108.4fps) | 89.7% (75.7fps) | 88.7% (50.8fps) | 87.2% (25.4fps) | GA102, 10496 shaders, 1695MHz, 24GB GDDR6X@19.5Gbps, 936GB/s, 350W |
GeForce RTX 3080 Ti (opens in new tab) | 89.3% (105.6fps) | 87.6% (73.9fps) | 86.0% (49.2fps) | 84.6% (24.7fps) | GA102, 10240 shaders, 1665MHz, 12GB GDDR6X@19Gbps, 912GB/s, 350W |
GeForce RTX 3080 12GB (opens in new tab) | 88.5% (104.7fps) | 85.8% (72.4fps) | 83.7% (47.9fps) | 81.4% (23.7fps) | GA102, 8960 shaders, 1845MHz, 12GB GDDR6X@19Gbps, 912GB/s, 400W |
GeForce RTX 3080 (opens in new tab) | 81.5% (96.3fps) | 78.5% (66.3fps) | 76.3% (43.7fps) | 72.2% (21.0fps) | GA102, 8704 shaders, 1710MHz, 10GB GDDR6X@19Gbps, 760GB/s, 320W |
Radeon RX 6950 XT (opens in new tab) | 70.4% (83.2fps) | 66.7% (56.2fps) | 62.9% (36.0fps) | 59.0% (17.2fps) | Navi 21, 5120 shaders, 2310MHz, 16GB GDDR6@18Gbps, 576GB/s, 335W |
GeForce RTX 3070 Ti (opens in new tab) | 66.3% (78.4fps) | 63.0% (53.1fps) | 59.2% (33.9fps) | GA104, 6144 shaders, 1770MHz, 8GB GDDR6X@19Gbps, 608GB/s, 290W | |
Radeon RX 6900 XT (opens in new tab) | 63.0% (74.5fps) | 59.0% (49.8fps) | 55.2% (31.6fps) | 51.7% (15.1fps) | Navi 21, 5120 shaders, 2250MHz, 16GB GDDR6@16Gbps, 512GB/s, 300W |
Titan RTX (opens in new tab) | 62.5% (73.9fps) | 58.2% (49.1fps) | 55.4% (31.7fps) | 52.5% (15.3fps) | TU102, 4608 shaders, 1770MHz, 24GB GDDR6@14Gbps, 672GB/s, 280W |
GeForce RTX 3070 (opens in new tab) | 62.1% (73.4fps) | 58.7% (49.6fps) | 54.9% (31.4fps) | GA104, 5888 shaders, 1725MHz, 8GB GDDR6@14Gbps, 448GB/s, 220W | |
GeForce RTX 2080 Ti (opens in new tab) | 59.2% (70.0fps) | 55.1% (46.5fps) | 52.0% (29.7fps) | TU102, 4352 shaders, 1545MHz, 11GB GDDR6@14Gbps, 616GB/s, 250W | |
Radeon RX 6800 XT (opens in new tab) | 59.0% (69.7fps) | 54.6% (46.1fps) | 51.3% (29.4fps) | 48.2% (14.0fps) | Navi 21, 4608 shaders, 2250MHz, 16GB GDDR6@16Gbps, 512GB/s, 300W |
GeForce RTX 3060 Ti (opens in new tab) | 55.2% (65.3fps) | 51.3% (43.3fps) | 47.8% (27.4fps) | GA104, 4864 shaders, 1665MHz, 8GB GDDR6@14Gbps, 448GB/s, 200W | |
Radeon RX 6800 (opens in new tab) | 50.4% (59.6fps) | 46.6% (39.3fps) | 43.6% (24.9fps) | Navi 21, 3840 shaders, 2105MHz, 16GB GDDR6@16Gbps, 512GB/s, 250W | |
GeForce RTX 2080 Tremendous (opens in new tab) | 49.6% (58.6fps) | 45.0% (37.9fps) | 41.6% (23.8fps) | TU104, 3072 shaders, 1815MHz, 8GB GDDR6@15.5Gbps, 496GB/s, 250W | |
GeForce RTX 2080 (opens in new tab) | 47.5% (56.2fps) | 42.5% (35.9fps) | 39.1% (22.4fps) | TU104, 2944 shaders, 1710MHz, 8GB GDDR6@14Gbps, 448GB/s, 215W | |
GeForce RTX 2070 Tremendous (opens in new tab) | 43.6% (51.5fps) | 39.2% (33.1fps) | 35.5% (20.3fps) | TU104, 2560 shaders, 1770MHz, 8GB GDDR6@14Gbps, 448GB/s, 215W | |
Radeon RX 6750 XT (opens in new tab) | 41.7% (49.3fps) | 39.1% (33.0fps) | 35.6% (20.4fps) | Navi 22, 2560 shaders, 2600MHz, 12GB GDDR6@18Gbps, 432GB/s, 250W | |
GeForce RTX 3060 (opens in new tab) | 41.2% (48.7fps) | 38.3% (32.3fps) | 35.1% (20.1fps) | GA106, 3584 shaders, 1777MHz, 12GB GDDR6@15Gbps, 360GB/s, 170W | |
Radeon RX 6700 XT (opens in new tab) | 38.8% (45.9fps) | 36.4% (30.7fps) | 32.9% (18.8fps) | Navi 22, 2560 shaders, 2581MHz, 12GB GDDR6@16Gbps, 384GB/s, 230W | |
GeForce RTX 2070 (opens in new tab) | 38.5% (45.5fps) | 34.9% (29.4fps) | 31.6% (18.1fps) | TU106, 2304 shaders, 1620MHz, 8GB GDDR6@14Gbps, 448GB/s, 175W | |
GeForce RTX 2060 Tremendous (opens in new tab) | 36.9% (43.6fps) | 33.0% (27.9fps) | 29.9% (17.1fps) | TU106, 2176 shaders, 1650MHz, 8GB GDDR6@14Gbps, 448GB/s, 175W | |
GeForce RTX 2060 (opens in new tab) | 31.8% (37.6fps) | 26.7% (22.5fps) | TU106, 1920 shaders, 1680MHz, 6GB GDDR6@14Gbps, 336GB/s, 160W | ||
Radeon RX 6650 XT (opens in new tab) | 31.6% (37.3fps) | 29.0% (24.5fps) | Navi 23, 2048 shaders, 2635MHz, 8GB GDDR6@18Gbps, 280GB/s, 180W | ||
Radeon RX 6600 XT (opens in new tab) | 30.8% (36.4fps) | 28.0% (23.6fps) | Navi 23, 2048 shaders, 2589MHz, 8GB GDDR6@16Gbps, 256GB/s, 160W | ||
GeForce RTX 3050 (opens in new tab) | 29.4% (34.8fps) | 27.0% (22.8fps) | GA106, 2560 shaders, 1777MHz, 8GB GDDR6@14Gbps, 224GB/s, 130W | ||
Radeon RX 6600 (opens in new tab) | 25.8% (30.5fps) | 23.3% (19.6fps) | Navi 23, 1792 shaders, 2491MHz, 8GB GDDR6@14Gbps, 224GB/s, 132W | ||
Radeon RX 6500 XT (opens in new tab) | 9.4% (11.2fps) | Navi 24, 1024 shaders, 2815MHz, 4GB GDDR6@18Gbps, 144GB/s, 107W | |||
Radeon RX 6400 (opens in new tab) | 7.6% (9.0fps) | Navi 24, 768 shaders, 2321MHz, 4GB GDDR6@16Gbps, 128GB/s, 53W |
The place AMD can declare the highest spot in our normal check suite at 1080p and 1440p, as soon as we allow ray tracing, the most effective AMD can do is sixth place, simply forward of the RTX, RTX 3070 Ti. It is a precipitous drop, and we’re not even utilizing DLSS, which all six of our DXR video games assist.
You may see what DLSS High quality mode did for efficiency on the Asus RTX 3090 Ti in our evaluation, however the brief abstract is that it boosted efficiency by 48% at 1080p extremely, 62% at 1440p extremely, and 89% at 4K extremely — with that final taking efficiency from a borderline unplayable 31 fps common as much as a cushty 60 fps. You may as well legitimately use the Balanced and Efficiency modes with out killing picture high quality, particularly at 4K, which is able to ship even bigger beneficial properties.
AMD’s FSR 2.0 would show useful right here, if AMD can get widespread adoption — AMD’s quickest GPUs can barely handle 1440p at greater than 30 fps. Additionally word that not one of the GPUs can deal with native 4K in the entire video games, although the RTX 3080 was 45% sooner than the RX 6900 XT, and the RTX 3090 Ti was 93% sooner. Hopefully the upcoming Nvidia Ada and AMD RDNA 3 GPUs will be capable of deal with 4K at native decision whereas reaching playable framerates, however even then we count on DLSS or FSR 2.0 can be essential for 60 fps or extra.
The midrange GPUs just like the RTX 3070 and RX 6700 XT mainly handle 1080p extremely and never far more, whereas the underside tier of DXR-capable GPUs barely handle 1080p medium — and the RX 6500 XT cannot even try this, with single digit framerates in most of our check suite, and one sport that would not even work at our chosen “medium” settings. (Management requires at the very least 6GB VRAM to allow you to enabled ray tracing.)
It is also fascinating to take a look at the generational efficiency of Nvidia’s RTX playing cards. The slowest 20-series GPU, the RTX 2060, nonetheless outperforms the brand new RTX 3050 by a bit, however the quickest RTX 2080 Ti is available in a bit behind the RTX 3070. The place the 2080 Ti mainly doubled the efficiency of the 2060, the 3090 delivers about triple the efficiency of the 3050.
2020-2021 GPU Benchmarks Rating
The outcomes beneath are from our earlier model of the GPU benchmarks hierarchy, utilizing a unique check suite and mixing outcomes from 9 video games with six decision and setting mixtures. All the scores are mixed (through a geometrical imply calculation) right into a single general end result, which tends to penalize the quickest and slowest GPUs — CPU bottlenecks come into play at 1080p medium, whereas VRAM limitations can kill efficiency at 4K extremely.
These outcomes haven’t been up to date since early 2022, after we added the RTX 3050 and RX 6500 XT to the record. We can’t be including future GPUs to this desk, so there is not any 3090 Ti, 6950 XT, 6750 XT, or 6650 XT, but it surely does assist to supply a have a look at a barely much less demanding suite of video games, the place 6GB or extra VRAM is not usually required at 1080p extremely settings. You should use these older outcomes to assist inform your buy choices, should you do not sometimes run the most recent video games at maxed out settings.
Selecting a Graphics Card
Which graphics card do you want? That will help you resolve, we created this GPU benchmarks hierarchy consisting of dozens of GPUs from the previous 4 generations of {hardware}. Not surprisingly, the quickest playing cards use both Nvidia’s Ampere structure or AMD’s Huge Navi. AMD’s newest graphics playing cards carry out effectively with out ray tracing, however are likely to fall behind as soon as RT will get enabled — much more so should you allow DLSS, which it is best to. GPU costs are nonetheless an enormous hindrance to upgrading, sadly, and we’re possible at the very least six months away from hitting considerably affordable ranges.
In fact it is not nearly enjoying video games. Many functions use the GPU for different work, and we lined some skilled GPU benchmarks in our RTX 3090 Ti evaluation. However a very good graphics card for gaming will sometimes do equally effectively in complicated GPU computational workloads. Purchase one of many prime playing cards and you may run video games at excessive resolutions and body charges with the consequences turned all the way in which up, and you can do content material creation work equally effectively. Drop all the way down to the center and decrease parts of the record and you may want to begin dialing down the settings to get acceptable efficiency in common sport play and GPU benchmarks.
It isn’t nearly high-end GPUs both, in fact. We examined Intel’s Xe Graphics DG1, which mainly competes with built-in graphics options. The outcomes weren’t fairly, and we did not even strive operating any of these at settings past 1080p medium. Nonetheless, you’ll be able to see the place these GPUs land on the very backside of the 2020-2021 GPU benchmarks record. Fortunately, Intel’s Arc Alchemist, aka DG2, seems to be reduce from completely completely different fabric. We hope, anyway.
In case your fundamental aim is gaming, you’ll be able to’t neglect concerning the CPU. Getting the absolute best gaming GPU will not aid you a lot in case your CPU is underpowered and/or outdated. So you should definitely take a look at the Greatest CPUs for gaming web page, in addition to our CPU Benchmarks Hierarchy to be sure you have the fitting CPU for the extent of gaming you are seeking to obtain.
Check System and How We Check for GPU Benchmarks
We have used two completely different PCs for our testing. The newest 2022 and later configuration makes use of an Alder Lake CPU and platform, whereas our earlier testbed makes use of Espresso Lake and Z390. Listed below are the main points of the 2 PCs.
Tom’s {Hardware} 2022 GPU Testbed
Intel Core i9-12900K (opens in new tab)
MSI Professional Z690-A WiFi DDR4 (opens in new tab)
Corsair 2x16GB DDR4-3600 CL16 (opens in new tab)
Essential P5 Plus 2TB (opens in new tab)
Cooler Grasp MWE 1250 V2 Gold (opens in new tab)
Cooler Grasp PL360 Flux (opens in new tab)
Cooler Grasp HAF500
Home windows 11 Professional 64-bit
Tom’s {Hardware} 2020–2021 GPU Testbed
Intel Core i9-9900K (opens in new tab)
Corsair H150i Professional RGB (opens in new tab)
MSI MEG Z390 Ace (opens in new tab)
Corsair 2x16GB DDR4-3200 (opens in new tab)
XPG SX8200 Professional 2TB (opens in new tab)
Home windows 10 Professional (opens in new tab) (21H1)
For every graphics card, we comply with the identical testing process. We run one move of every benchmark to “heat up” the GPU after launching the sport, then run at the very least two passes at every setting/decision mixture. If the 2 runs are mainly similar (inside 0.5% or much less distinction), we use the sooner of the 2 runs. If there’s greater than a small distinction, we run the check at the very least twice extra to find out what “regular” efficiency is meant to be.
We additionally have a look at all the information and verify for anomalies, so for instance RTX 3070 Ti, RTX 3070, and RTX 3060 Ti all usually going to carry out inside a slim vary — 3070 Ti is about 5% sooner than 3070, which is about 5% sooner than 3060 Ti. If we see video games the place there are clear outliers (i.e. efficiency is greater than 10% increased for the playing cards simply talked about), we’ll return and retest no matter playing cards are displaying the anomaly and determine what the “appropriate” end result can be.
Because of the size of time required for testing every GPU, up to date drivers and sport patches inevitably will come out that may impression efficiency. We periodically retest just a few pattern playing cards to confirm our outcomes are nonetheless legitimate, and if not, we undergo and retest the affected sport(s) and GPU(s). We might also add video games to our check suite over the approaching 12 months, if one comes out that’s fashionable and conducive to testing — see our what makes a very good sport benchmark for our choice standards.
GPU Benchmarks: Particular person Sport Charts
The above tables present a abstract of efficiency, however for those who need to see the person sport charts, for each the usual and ray tracing check suites, we have these as effectively. These charts have been up-to-date as of Could 19, 2022, with testing performed utilizing the most recent Nvidia and AMD drivers normally, although a few of the playing cards have been examined with barely older drivers.
Observe that we’re solely together with the previous two generations of {hardware} in these charts, as in any other case issues get too cramped — and you may argue that with 35 playing cards within the 1080p charts, we’re already effectively previous that time. (Trace: Click on the enlarge icon should you’re on PC.)
Additionally word that we have switched from DX12 to DX11 for Microsoft Flight Simulator testing, partly as a result of DX12 began to have points lately, partly as a result of DX12 continues to be listed as “beta,” however largely as a result of we have decided DX11 runs sooner — somethings by greater than 10% — on most GPUs. We have retested the entire playing cards in DX11 mode now.
Energy, Clocks, Temperatures, and Fan Speeds
Whereas our GPU benchmarks hierarchy types issues solely by efficiency, for these fascinated about energy and different facets of the GPUs, listed below are the suitable charts.
Legacy GPU Hierarchy
Beneath is our legacy desktop GPU hierarchy relationship again to the late Nineteen Nineties. We’ve not examined most of those playing cards in a few years, driver assist has ended on most of those fashions, and the relative rankings are fairly coarse. Observe that we additionally do not consider reminiscence bandwidth or options like AMD’s Infinity Cache. The record beneath is usually supposed to indicate relative efficiency between architectures from an analogous time interval.
We sorted the desk by the theoretical GFLOPS, although on architectures that do not assist unified shaders, we solely have information for “Gops/s” (giga operations per second). That is GeForce 7 and Radeon X1000 and earlier — mainly something from earlier than 2007. We have put an asterisk (*) subsequent to the GPU names for these playing cards, they usually comprise the latter a part of the desk. Evaluating pre-2007 GPUs towards one another ought to be comparatively significant, however attempting to check these older GPUs towards newer GPUs will get a bit convoluted.
* – Denotes efficiency measured in gigaoperations per second, versus GFLOPS. Older GPU architectures with out unified shader assist aren’t straight comparable with newer architectures.
Discovering Reductions on the Greatest Graphics Playing cards
With all of the GPU shortages lately, you are unlikely to see large gross sales on a graphics card, however you could discover some financial savings by trying out the most recent Newegg promo codes, Greatest Purchase promo codes and Micro Middle coupon codes.
For much more data, take a look at our Graphics Card Purchaser’s Information.
MORE: Greatest Graphics Playing cards for Gaming
MORE: Graphics Card Energy Consumption Examined
MORE: Find out how to Stress-Check Graphics Playing cards (Like We Do)
MORE: CPU Benchmarks
Need to touch upon this story? Tell us what you suppose within the Tom’s {Hardware} Boards.