1 00:00:00,000 --> 00:00:04,920 NVIDIA's RTX 4090 was a massively 2 00:00:03,000 --> 00:00:08,099 powerful GPU with massive power consumption to match something that's 3 00:00:06,480 --> 00:00:13,139 been making headlines since before launch with or without combusting 4 00:00:09,960 --> 00:00:15,179 adapter cables now the RTX 4080 is here 5 00:00:13,139 --> 00:00:18,000 and NVIDIA claims that it's going to double the performance of the previous 6 00:00:16,560 --> 00:00:22,500 generation which would put it dangerously close to the RTX 4090s 7 00:00:20,580 --> 00:00:27,359 performance level and what they say will be a much lower power draw and yet 8 00:00:24,720 --> 00:00:31,740 NVIDIA's claims were far higher than the power draw we actually see 9 00:00:29,699 --> 00:00:36,000 the kind of lie that I don't mind hearing the question now is can this 10 00:00:34,739 --> 00:00:41,640 thing live up to those performance claims or at twelve hundred dollars well 11 00:00:39,000 --> 00:00:46,260 the 40 series inflated price tags make it a skip-it generation like the RTX 20 12 00:00:44,280 --> 00:00:49,260 series cards it's only one way to find out 13 00:00:47,340 --> 00:00:54,840 and there's only one way to segue to our sponsor Karma save money while shopping 14 00:00:51,960 --> 00:00:58,860 with Karma's new pay with K feature with it you'll get the lowest price and get 15 00:00:56,820 --> 00:01:03,920 cash back bonuses on all of the Best Brands learn more about Karma later on 16 00:01:01,199 --> 00:01:03,920 in this video 17 00:01:09,439 --> 00:01:16,560 thank you the RTX 4080 16 gig or I guess 18 00:01:14,880 --> 00:01:21,479 I don't really need to call it that anymore the 4080 has roughly 60 percent 19 00:01:19,740 --> 00:01:26,820 of the Cuda cores compared to the bigger 4090 with slightly lower clocks to go 20 00:01:24,659 --> 00:01:30,960 with them this is coupled with a 256 bit instead of a 384-bit memory bus which 21 00:01:29,040 --> 00:01:35,880 means effective bandwidth is cut to just two-thirds given that the starting price 22 00:01:33,060 --> 00:01:40,560 is roughly two-thirds of the 40-90 the math is mathing although the rated total 23 00:01:38,700 --> 00:01:45,720 Graphics power is a bit higher than you'd expect given the spec Cuts still 24 00:01:42,780 --> 00:01:50,579 it's the same 320 watt tgp as the outgoing RTX 3080. so if we're gaining 25 00:01:48,720 --> 00:01:56,100 performance like NVIDIA says we are we're not paying for it there at least 26 00:01:52,860 --> 00:01:58,079 in theory let's test it in practice we 27 00:01:56,100 --> 00:02:02,100 grab the RTX 4080 along with its predecessor and what we consider to be 28 00:01:59,880 --> 00:02:06,060 the closest competing gpus in terms of price and performance and slap them into 29 00:02:03,960 --> 00:02:10,560 our 13th gen core bench to give them the best chance to stretch their legs 30 00:02:07,500 --> 00:02:13,500 starting with non-ray Trace 4K the RTX 31 00:02:10,560 --> 00:02:18,780 4080 pulls a respect 25 lead over the RTX 3090 TI in Forza Horizon 5 and 32 00:02:16,739 --> 00:02:24,060 nearly reaches 80 percent of the RTX 4090s performance which puts it on solid 33 00:02:21,480 --> 00:02:28,020 footing out of the gate f-122 sees the 6950 XT pulling close thanks to the 34 00:02:26,640 --> 00:02:32,160 Monstrous traditional rendering performance that cart has and the 3090 35 00:02:30,180 --> 00:02:35,940 TI closes the Gap substantially thanks to its greater memory bandwidth the 3090 36 00:02:34,500 --> 00:02:40,319 TI's performance carries over to cyberpunk somewhat but the other cards 37 00:02:38,280 --> 00:02:44,940 fall behind NVIDIA's 40 series cards Hitman 3 sees the RTX 4090 pull 38 00:02:43,019 --> 00:02:50,400 significantly ahead of the 4080 again likely thanks to its memory bandwidth as 39 00:02:46,800 --> 00:02:52,560 the 3090 TI and 6950xt both trade blows 40 00:02:50,400 --> 00:02:57,300 in this title Modern Warfare 2 pushes the 4080 back again compared to the 4090 41 00:02:55,260 --> 00:03:01,680 but it's still commanding that roughly 50 lead over its predecessor and sits 42 00:02:59,760 --> 00:03:05,879 comfortably ahead of the 3090 TI and 6950xt older titles like Red Dead 43 00:03:04,200 --> 00:03:10,019 Redemption 2 and Shadow of the Tomb Raider both display similar results 44 00:03:07,560 --> 00:03:14,879 across the board with the RTX 4080 clearly topping the last generation's 45 00:03:12,300 --> 00:03:19,080 best cards but trailing the 4090 by about 30 percent 1440p results are 46 00:03:17,220 --> 00:03:22,739 unsurprisingly much tighter thanks to the lighter load it places on the GPU 47 00:03:20,819 --> 00:03:27,060 although not in the way you might expect while the 4080 closes in on the 4090 the 48 00:03:25,560 --> 00:03:31,920 other gpus remain where they were relative to the 4K results suggesting 49 00:03:29,400 --> 00:03:35,220 that it's only the RTX 4090 that's CPU Bound in these tests there are cases 50 00:03:33,599 --> 00:03:39,840 like cyberpunk and Modern Warfare 2 where the extra horsepower still makes a 51 00:03:37,080 --> 00:03:43,860 big difference but for users of 1440p displays and especially in older titles 52 00:03:41,760 --> 00:03:48,360 like Red Dead Redemption 2 and Shadow of the Tomb Raider 53 00:03:45,120 --> 00:03:50,280 the RTX 4080 may be more than enough to 54 00:03:48,360 --> 00:03:55,099 achieve close to 40 90 level performance while costing substantially less money 55 00:03:52,739 --> 00:04:00,480 and drawing substantially less power which does also cost money we'll get to 56 00:03:58,080 --> 00:04:05,459 the power draw but first Ray tracing this is AMD's Achilles heel and it shows 57 00:04:04,019 --> 00:04:11,220 even in Shadow of the Tomb Raider where the 6950 XT is only roughly equal to the 58 00:04:08,700 --> 00:04:15,599 lower tier RTX 3080. even in this older title NVIDIA's new RT cores appear to 59 00:04:13,799 --> 00:04:21,239 perform slightly faster than the 3090 TI's giving the 4080 a broader lead than 60 00:04:18,419 --> 00:04:25,620 it had in traditional rendering f-122 by comparison had the 4080 lose relative 61 00:04:23,699 --> 00:04:29,699 performance against the Ford 90 versus when Ray tracing was turned off but it 62 00:04:27,900 --> 00:04:34,380 pulls ahead substantially from the RTX 3080 making for frame rates that are 63 00:04:32,100 --> 00:04:38,580 playable at 4K rather than cinematic cyberpunk unfortunately remains 64 00:04:36,479 --> 00:04:43,259 cinematic on all of our cards at 4K with Ray tracing enabled though it's here 65 00:04:40,560 --> 00:04:48,300 that we really do see the RTX 4080 pull over double the performance of its 66 00:04:44,699 --> 00:04:51,120 predecessor and AMD's 6950xt and while 67 00:04:48,300 --> 00:04:55,199 dlss 3.0 is still early days it appears that the two cards handle its frame 68 00:04:52,680 --> 00:04:59,400 generation roughly the same way which is unsurprising but we had to test it to be 69 00:04:57,300 --> 00:05:04,320 sure for the same reasons as 1440p rendering performance the RTX 4080 seems 70 00:05:02,160 --> 00:05:08,400 to benefit more from dlss overall than the 4090 does thanks to the latter 71 00:05:06,479 --> 00:05:11,340 becoming more CPU bound and this is something we should see more of as the 72 00:05:10,020 --> 00:05:15,180 rest of the lineup continues to launch just like we recently launched these 73 00:05:13,320 --> 00:05:18,840 various Nazi Channel hoodies on ltdstore.com they feel like stealth 74 00:05:17,040 --> 00:05:23,100 hoodies but they're anything but moving on to productivity the RTX 4080 is 75 00:05:21,419 --> 00:05:28,080 nearly double the performance of the 3080m blender rendering thanks both to 76 00:05:25,440 --> 00:05:31,560 the additional and upgraded RT cores and it seems to come within three quarters 77 00:05:29,759 --> 00:05:35,940 of the speed of the 4090. DaVinci Resolve favors the 4090 by about a 78 00:05:33,660 --> 00:05:40,800 minute with the 4080 coming in just a little slower than the 3090 TI however 79 00:05:38,100 --> 00:05:45,720 its addition of av1 encoding means that any 40 series GPU I mean any of them is 80 00:05:43,620 --> 00:05:49,620 going to make your PC sub substantially faster at this kind of rendering 81 00:05:47,220 --> 00:05:54,419 compared to any of the other gpus we've tested here Matlab GPU bench is NVIDIA 82 00:05:52,560 --> 00:05:58,620 only but we can see that depending on the algorithm the RTX 4080 can come 83 00:05:56,699 --> 00:06:03,539 within spitting distance of the 4090 in single Precision or be a good 40 behind 84 00:06:01,680 --> 00:06:08,280 in double Precision either way it's consistently well ahead of the RTX 3090 85 00:06:05,759 --> 00:06:12,900 TI and 3080 making it an excellent choice for GPU compute it's less 86 00:06:10,560 --> 00:06:17,460 clear-cut for topaz AI where our test Suite finished soon around the 4080 than 87 00:06:14,880 --> 00:06:21,900 the 3090 TI and within 10 minutes of the 4090 but AMD actually pulled ahead here 88 00:06:20,039 --> 00:06:26,039 suggesting that topaz is geared more towards raw GPU compute than AMD's AI 89 00:06:24,600 --> 00:06:31,259 accelerating tensor cores or at least the AI model we used was finally 90 00:06:28,680 --> 00:06:35,639 specviewperf puts the RTX 4080 in the good overall position handling 91 00:06:32,759 --> 00:06:41,039 outperforming both the RTX 3080 and 3090 TI although AMD pulls off some wins and 92 00:06:38,220 --> 00:06:45,000 Katia Creo energy medical and especially Siemens NX though the latter is likely 93 00:06:43,139 --> 00:06:49,380 because that version that's used in spec viewperf doesn't support acceleration on 94 00:06:47,220 --> 00:06:53,819 NVIDIA yet power consumption has been a Hot Topic since the earliest rumors of 95 00:06:51,060 --> 00:06:59,520 the RTX 40 series launch and the RTX 4090 proved our concerns were valid 96 00:06:56,400 --> 00:07:01,740 however the RTX 4080 is 97 00:06:59,520 --> 00:07:07,080 surprisingly tame yeah during our testing it almost never reached its 320 98 00:07:04,620 --> 00:07:11,340 watt rated total Graphics power instead hovering around the 300 watt Mark or 99 00:07:09,000 --> 00:07:15,960 usually lower we were so confused by this that we asked NVIDIA what was up 100 00:07:13,500 --> 00:07:20,400 and they responded by saying Oh Yeah we actually we're changing what tgp means 101 00:07:18,840 --> 00:07:25,440 sure okay 102 00:07:21,900 --> 00:07:27,840 so what did they say basically tgp has 103 00:07:25,440 --> 00:07:32,039 popularly been thought of as a power Target until recently it's the power 104 00:07:29,940 --> 00:07:36,180 level that the GPU tried to adhere to when under load opportunistic boost 105 00:07:33,780 --> 00:07:41,280 meant that gpus would often Spike higher for short periods but usually settled in 106 00:07:37,979 --> 00:07:43,919 around their tgp rating now tgp is being 107 00:07:41,280 --> 00:07:48,360 looked at more like a power limit the maximum the card should hit under normal 108 00:07:45,840 --> 00:07:52,500 conditions and NVIDIA's built themselves a buffer Igor's lab recently published 109 00:07:50,940 --> 00:07:56,699 an article showing this in Greater depth with the RTX 4090 which I'll have linked 110 00:07:55,020 --> 00:08:01,979 below if you want to check it out NVIDIA themselves claimed that the RTX 4080 111 00:07:58,680 --> 00:08:04,800 draws roughly 251 Watts on average while 112 00:08:01,979 --> 00:08:09,240 gaming though it's closer to 300 watts at 4K depending on the game and yeah 113 00:08:07,020 --> 00:08:14,099 even when running a power virus we couldn't get the 4080 to go much higher 114 00:08:11,340 --> 00:08:18,240 than the rated 320 Watts despite the GPU load remaining High and the core clocks 115 00:08:15,900 --> 00:08:22,680 remaining above 2 gigahertz this leads to some interesting questions because 116 00:08:20,099 --> 00:08:26,639 each and every RTX 4880 you've tested has had a massively overbuilt cooler 117 00:08:25,020 --> 00:08:30,360 that resulted in the card barely ever getting warm like seriously Under full 118 00:08:28,560 --> 00:08:35,940 synthetic load we're looking at around the low 60 degrees at worst and in games 119 00:08:33,479 --> 00:08:40,860 it sits around in the 50s the fans never even began to approach 50 percent let 120 00:08:38,820 --> 00:08:44,760 alone higher even the internal case temperature was the lowest of any GPU in 121 00:08:43,200 --> 00:08:50,519 our test Suite with a delta T over ambient of between just 10 and 11 122 00:08:47,040 --> 00:08:53,220 degrees unlike the RTX 4090 this GPU 123 00:08:50,519 --> 00:08:58,019 would be well suited to a small form factor chassis I'm wondering if NVIDIA 124 00:08:55,500 --> 00:09:03,300 made this adjustment to ggp latent development and nobody got the memo in 125 00:09:00,120 --> 00:09:05,459 time the question now is why though 126 00:09:03,300 --> 00:09:09,540 NVIDIA didn't volunteer that information when we asked when we overclocked the 127 00:09:07,620 --> 00:09:13,980 cards we noted that there wasn't a lot of Headroom for the GPU core clocks our 128 00:09:11,760 --> 00:09:17,880 cards could only push between 110 and 150 megahertz over stock and while this 129 00:09:16,260 --> 00:09:21,779 is obviously going to vary from card to card that's only about one percent which 130 00:09:19,740 --> 00:09:25,380 is much lower than the roughly 15 percent you might find back on the 10 131 00:09:23,399 --> 00:09:29,339 series cards this might be one reason why NVIDIA dialed back the GPU power a 132 00:09:27,779 --> 00:09:32,220 bit in fact even when we pushed the memory and the core clock as high as 133 00:09:30,839 --> 00:09:36,180 they would go we still didn't get substantially higher frame rates however 134 00:09:34,019 --> 00:09:41,040 we did substantially increase power consumption letting further Credence the 135 00:09:38,700 --> 00:09:46,320 idea that NVIDIA saw no reason to push the GPU to its rated 320 Watts under the 136 00:09:43,800 --> 00:09:50,279 normal operation what we've observed is its sweet spot with that in mind what 137 00:09:49,080 --> 00:09:54,420 are you getting for twelve hundred dollars U.S you're getting a GPU that's 138 00:09:52,680 --> 00:09:58,920 much faster than anything the previous generation can muster with dual envy 139 00:09:56,760 --> 00:10:02,519 encoders that support av-1 a thermally very tame design with boatloads of 140 00:10:00,959 --> 00:10:07,019 thermal Headroom that could potentially be cut down to fit lower profile chassis 141 00:10:04,860 --> 00:10:10,980 and official support for deep learning frame generation to boost perceived 142 00:10:08,880 --> 00:10:16,500 frame rates where it needs to however unlike the RTX 4090 which is a Halo 143 00:10:14,040 --> 00:10:22,320 product with no price pair currently the RTX 4080 has to compete with those last 144 00:10:19,019 --> 00:10:24,720 gen cards and while the RTX 3090 TI is 145 00:10:22,320 --> 00:10:28,440 close in its current retail price it's not as close in performance or power 146 00:10:26,880 --> 00:10:32,580 consumption you're paying an extra hundred dollars for the 4080 and getting 147 00:10:30,480 --> 00:10:39,300 in my opinion more than a hundred dollars in value the RX 6950 XT and RTX 148 00:10:35,940 --> 00:10:40,920 3080 however those do represent a threat 149 00:10:39,300 --> 00:10:44,820 in terms of value if the extra features simply don't matter to you some of these 150 00:10:42,720 --> 00:10:50,100 can be purchased for as little as 800 brand new with used cards routinely 151 00:10:47,519 --> 00:10:54,240 going for Less sometimes even 600 at half the price they definitely do not 152 00:10:51,959 --> 00:10:57,899 represent just half the value based on our testing bear in mind the AMD card 153 00:10:55,980 --> 00:11:01,380 draws substantially more power so if that's a concern go with the 3080 for 154 00:10:59,640 --> 00:11:05,279 now we'll have all of those linked below none of which is to say that you 155 00:11:02,820 --> 00:11:10,079 shouldn't buy the RTX 4080 it's a solid card and it's only just getting started 156 00:11:07,079 --> 00:11:12,240 just maybe wait for reviews of AMD's 157 00:11:10,079 --> 00:11:16,680 rdna 3 gpus to drop before pulling the trigger who knows you'll probably even 158 00:11:14,040 --> 00:11:20,459 go on sale if AMD does a good enough job competition is beautiful 159 00:11:18,899 --> 00:11:25,320 and hopefully I did a good enough job segueing to our sponsor karma karma is a 160 00:11:23,579 --> 00:11:29,579 browser extension and mobile app that's designed to help you be a better online 161 00:11:27,120 --> 00:11:34,019 Shopper and save you money it has features like coupon scanning price drop 162 00:11:31,680 --> 00:11:37,980 alerts real-time stock updates and tons of other tools are included to help you 163 00:11:35,579 --> 00:11:42,480 get whatever product you want you can even organize the stuff you want into 164 00:11:39,720 --> 00:11:47,160 wish lists to save for a later time and get cash back bonuses on all the Best 165 00:11:44,700 --> 00:11:50,579 Brands Karma's newest feature pay with Karma or pay with K allows you to 166 00:11:49,500 --> 00:11:55,200 connect several different payment options like credit cards and debit 167 00:11:52,680 --> 00:11:58,980 cards to Apple pay and Google pay so you can shop at all your favorite Brands 168 00:11:56,940 --> 00:12:03,720 once your info is loaded up you can experience One swipe shopping all within 169 00:12:01,320 --> 00:12:09,300 karma for this week only sign up with Karma using our link and get 10 added to 170 00:12:06,660 --> 00:12:13,560 your account instead of five dollars and don't forget to also check out their 171 00:12:11,279 --> 00:12:16,800 commercial linked below it will blow your mind how it works thanks for 172 00:12:15,300 --> 00:12:21,360 watching guys go check out our response to AMD's upcoming rdna 3 graphics card 173 00:12:18,899 --> 00:12:27,200 cards for more on how the GPU landscape is changing NVIDIA launched first but 174 00:12:23,760 --> 00:12:27,200 they might not keep the lead