WEBVTT 00:00:07.279 --> 00:00:15.759 well it's semiformal review time again guys this is the GeForce GTX 680 from 00:00:12.920 --> 00:00:21.480 Galaxy this is the new flagship single GPU graphics card from NVIDIA by our 00:00:18.199 --> 00:00:25.560 friends at Galaxy this is a 2 GB uh 00:00:21.480 --> 00:00:29.039 frame buffer over 1 GHz core clock and 00:00:25.560 --> 00:00:31.320 over 6 GHz memory clock graphics card it 00:00:29.039 --> 00:00:37.440 has two six pin power connectors it has a PCI Express 3.0 16x slot is compatible 00:00:35.120 --> 00:00:42.120 with uh the variety of SLI configurations that are available using 00:00:39.040 --> 00:00:44.200 these two SLI ports and it is on the 00:00:42.120 --> 00:00:49.039 latest manufacturing process enabling it to be phenomenally power efficient and 00:00:46.239 --> 00:00:54.079 my camera won't focus and perform like the bees knes and supposedly although I 00:00:51.320 --> 00:00:59.039 have not yet tested this card it also overclocks 00:00:55.840 --> 00:01:00.280 like like wo like pretty good so in 00:00:59.039 --> 00:01:06.360 terms of cards that I'm going to be testing against I have a total of seven 00:01:03.079 --> 00:01:10.479 cards to be tested so I've got the 7970 00:01:06.360 --> 00:01:12.560 from AMD this is a stock bone reference 00:01:10.479 --> 00:01:18.560 card to be compared against the reference GTX 680 I have a sapphire 00:01:16.000 --> 00:01:23.000 hd7950 so this is an aftermarket card you're going to see this one's strengths 00:01:20.240 --> 00:01:26.040 when we get to things like the Acoustics and temperature testing that's going to 00:01:24.640 --> 00:01:31.479 be a separate video though this is the performance video we've got the GTX 590 00:01:28.920 --> 00:01:36.320 this is a last generation dual GPU card to compare against the current 00:01:32.799 --> 00:01:38.560 generation single GPU card uh 6990 so 00:01:36.320 --> 00:01:45.320 this is a last generation dual GPU card from AMD then I want to compare the 680 00:01:41.880 --> 00:01:49.000 against the 580 as well as the previous 00:01:45.320 --> 00:01:51.200 generation GTX 570 so that we can see 00:01:49.000 --> 00:01:54.680 where it falls in terms of the overall performance stack now you see 3D Vision 00:01:53.399 --> 00:02:00.719 glasses there that's going to be a separate review this is going to be my 00:01:56.680 --> 00:02:03.039 2D performance at 1080p review since 00:02:00.719 --> 00:02:07.600 that is what the majority of monitors are supporting these days that is what 00:02:05.000 --> 00:02:12.239 we are going to be focused on how now I did change a couple things so instead of 00:02:10.319 --> 00:02:16.440 using my normal benchmarking settings for the games that I ran I did alter 00:02:14.599 --> 00:02:20.959 things a little bit so for Battlefield 3 I'm using the ultra preset for crisis 2 00:02:18.959 --> 00:02:25.640 I'm using Ultra Plus highres textures Witcher 2 is ultra plus the vsync is off 00:02:23.400 --> 00:02:29.640 Dirt 3 is four times anti-aliasing with high and Ultra we available Batman 00:02:28.280 --> 00:02:33.840 Arkham City is going to be using four time multi sample anti-aliasing this is 00:02:32.239 --> 00:02:38.800 important which kind of anti-aliasing I'm using because this new card supports 00:02:36.319 --> 00:02:44.120 a couple of modes of anti-aliasing that have some distinct advantage txaa and 00:02:41.959 --> 00:02:50.000 FXAA but we are not going to be using those for the purposes of this review 00:02:46.280 --> 00:02:52.040 this will be Apples to Apples phys X off 00:02:50.000 --> 00:02:57.280 okay and Skyrim will be at Ultra with four times msaa 8 times AF vsync forced 00:02:55.319 --> 00:03:02.280 off through the inii file and the highres texture pack the official 00:02:59.879 --> 00:03:06.239 highres texture pack you can see I've actually finished the benchmarks for all 00:03:04.280 --> 00:03:09.760 the other cards this takes forever guys it takes absolutely forever and you can 00:03:08.280 --> 00:03:14.560 also see a sneak peek of all the different scenarios I'm planning to run 00:03:11.680 --> 00:03:21.120 1080p review ah yes my test bench so this is my uh lg8 2011 test bench with a 00:03:18.640 --> 00:03:25.200 363k 30 30 3930k 00:03:22.760 --> 00:03:28.799 3930k uh this is running at stock speed so I haven't made any changes to that 00:03:26.760 --> 00:03:35.239 guy right there 16 gigs of Kingston hyperx memory a Seuss P9 x79 Deluxe 00:03:33.040 --> 00:03:41.480 motherboard as well as of course our h100 for my boot SSD I'm using a mushkin 00:03:38.720 --> 00:03:47.239 Kronos Deluxe my steam Drive is running off here and for my power supply I'm 00:03:43.519 --> 00:03:49.439 using an ocz ZX Series 1250 W modular 00:03:47.239 --> 00:03:57.920 power supply that pretty much covers it for the test bench itself and it is time 00:03:54.480 --> 00:04:00.439 to start testing this card because yes 00:03:57.920 --> 00:04:04.280 that is the line that is still blank has not yet been on the test pinch I'll be 00:04:01.840 --> 00:04:11.079 using the latest drivers on all cards so that is uh 12.2 for the AMD cards and 00:04:08.040 --> 00:04:12.799 then the hold on which one is this I'll 00:04:11.079 --> 00:04:18.160 have to look because it's going to be a different driver for the new card than 00:04:15.360 --> 00:04:24.440 it is for the other card so 29610 for every GeForce card except the 680 which 00:04:20.919 --> 00:04:26.880 will'll be using the 300 whatever it is 00:04:24.440 --> 00:04:34.960 sorry for the long intro guys here we go here's Rumble investigating my uh 00:04:30.759 --> 00:04:38.560 test bench while the GTX 680 runs 3D 00:04:34.960 --> 00:04:43.479 Mark 11 my first test of the 00:04:38.560 --> 00:04:43.479 night what a brat he is 00:05:02.639 --> 00:05:05.800 oh he's stepping on a 00:05:06.360 --> 00:05:12.440 6990 get off there go on get 00:05:19.240 --> 00:05:25.520 get here we are in the middle of a combined test so to give you guys some 00:05:22.960 --> 00:05:31.600 idea how power efficient this guy is check this out it's pulling less than 00:05:28.440 --> 00:05:34.800 280 Wat from the wall so that's probably 00:05:31.600 --> 00:05:37.360 around 250 WTS of actual power 00:05:34.800 --> 00:05:43.199 consumption with the 6 core processor high-end single graphics card and uh 00:05:40.479 --> 00:05:47.840 throughout all of this the thing is darn near silent as 00:05:45.960 --> 00:05:55.199 well but more on that later a score of 3172 puts it right in line with the rest 00:05:50.759 --> 00:05:56.880 of or with uh the GTX 590 in 3D Mark 11 00:05:55.199 --> 00:06:03.440 um although significantly behind the 6990 let's see how things go in Real 00:06:00.199 --> 00:06:06.039 World gameplay scenarios so as I always 00:06:03.440 --> 00:06:09.639 do I'm using well I gu guess I shouldn't say always because Dirt 3 is using a 00:06:07.840 --> 00:06:13.840 canned Benchmark and Batman Arkham City is using a canned Benchmark Battlefield 00:06:11.840 --> 00:06:18.080 3 is using my usual run through of the comrades beginning of the comrades level 00:06:16.000 --> 00:06:22.880 crisis 2 is using my usual run through of the first camp in crisis 2 The 00:06:20.919 --> 00:06:26.240 Witcher 2 is using a run through where I go up uh little Hill I fight some guys 00:06:25.039 --> 00:06:32.720 go through a building and then come into a courtyard where I fight some other guys and Skyrim uh the town is called 00:06:31.039 --> 00:06:36.280 white white run I think something like that and I do a couple laps of White Run 00:06:34.840 --> 00:06:41.360 looking at some people looking at some fire looking around so uh stay tuned for 00:06:38.880 --> 00:06:47.120 real world gaming benchmarks and the first results are in 00:06:43.840 --> 00:06:49.400 so while posting power results that are 00:06:47.120 --> 00:06:56.440 very similar to the incredibly power efficient 7970 per it posts also better 00:06:53.919 --> 00:07:01.599 average frame rate and the same minimum frame rate compared to the 7970 it 00:06:59.639 --> 00:07:07.960 actually performs similarly to the GTX 590 once again while consuming over 150 00:07:05.479 --> 00:07:11.440 watts less power at the wall uh something to note here is due to the 00:07:09.759 --> 00:07:15.879 very high graphical settings both the 580 and the 570 kind of choked on this 00:07:13.520 --> 00:07:19.960 Benchmark so you can see there is an enormous difference between this new 00:07:17.639 --> 00:07:25.479 generation premium card and the last generation one now crisis 2 is less of a 00:07:23.160 --> 00:07:30.720 clear win against the lower price 7970 but it is still a win nonetheless 00:07:28.520 --> 00:07:34.720 so it takes the performance Crown again in crisis 2 something that I've noticed 00:07:32.520 --> 00:07:39.039 at this point is that this is actually the first reference GeForce card that 00:07:36.960 --> 00:07:43.080 I've encountered that does have coil wi so it's a little bit of a high-pitched 00:07:41.039 --> 00:07:46.520 noise that you experience under load with the card and and I'm talking about 00:07:44.800 --> 00:07:50.159 in games not in folding at home or anything like that um this is something 00:07:48.440 --> 00:07:54.560 that I've experienced on pretty much every reference AMD card for the last 00:07:52.080 --> 00:08:00.159 few Generations but hadn't seen on a GeForce card as yet so the 680 does does 00:07:58.440 --> 00:08:05.080 exhibit this noise exhibit it um make it The Witcher 2 00:08:03.280 --> 00:08:12.360 looks like a tight battle between the 7970 in terms of average FPS and uh the 00:08:09.720 --> 00:08:16.800 GTX 680 but in actuality you can see my notes here from when I was testing this 00:08:14.560 --> 00:08:23.599 30fps which is considered sort of the minimum of playable was very unplayable 00:08:19.639 --> 00:08:26.800 felt very laggy whereas this 30f 32 FPS 00:08:23.599 --> 00:08:29.080 actually felt significantly better so uh 00:08:26.800 --> 00:08:33.640 so yeah that puts the GTX 680 as far as a single card Solutions go definitely at 00:08:32.000 --> 00:08:39.640 the top this is another title where the last generation products from NVIDIA the 00:08:36.120 --> 00:08:41.599 uh 580 and the 570 really struggled uh 00:08:39.640 --> 00:08:47.440 whether that's due to having less video memory remember the new card has 2 gigs 00:08:43.519 --> 00:08:50.040 versus 1.5 and 1.28 on those older cards 00:08:47.440 --> 00:08:55.040 um and also has significantly more power but uses a similar amount of 00:08:52.720 --> 00:08:59.640 power since I haven't shown you guys my Skyrim test sequence before I'll take 00:08:57.160 --> 00:09:04.079 this opportunity to do that so I'm I'm just running around this town following 00:09:02.360 --> 00:09:09.360 the same path every time looking off into the distance a little bit I always 00:09:07.360 --> 00:09:13.399 stop and look down at this camp for a couple seconds running very smoothly 00:09:11.200 --> 00:09:17.160 although I did see a flicker there yeah there's another one so I've experienced 00:09:15.040 --> 00:09:22.440 flickering now on the 6990 uh 00:09:19.040 --> 00:09:24.040 7970 and uh and this card out of all the 00:09:22.440 --> 00:09:29.399 cards that were tested I've seen this flickering on three of them so uh that's 00:09:27.600 --> 00:09:35.160 very frustrating definitely detracts from gameplay now I could probably get 00:09:32.519 --> 00:09:39.079 the flickering to go away by changing some graphical settings on any of those 00:09:36.800 --> 00:09:44.680 given cards but since what I'm trying to do is find a similarly demanding 00:09:41.360 --> 00:09:46.399 situation for each of the cards I don't 00:09:44.680 --> 00:09:50.320 really want to change any settings and I stop and look at this guard for a second 00:09:48.560 --> 00:09:54.720 stop and look at this guard for a second get my cat to go 00:09:51.880 --> 00:09:57.839 away um and then off we go around again so while I'm doing the second lap I'll 00:09:56.360 --> 00:10:01.519 talk about some other things so first of all is my power testing methodology I 00:10:00.360 --> 00:10:05.839 don't know if you guys can see it through the camera yeah you can you can see my power meter so what I'm doing is 00:10:04.480 --> 00:10:10.480 while I'm benchmarking I'm keeping an eye on that and what I'm looking for is 00:10:08.160 --> 00:10:15.880 sort of something along the lines of of a consistent Peak so if I see it Spike 00:10:13.360 --> 00:10:20.040 to 320 once but then you know otherwise hover more around the when it's when 00:10:17.839 --> 00:10:24.399 it's under high load around the 300 watt range 305 watt range then that's where 00:10:22.800 --> 00:10:28.760 I'll take the reading for that card yeah it's very approximate see this one I'd 00:10:26.800 --> 00:10:33.560 probably go with 305 based on what I've seen so far that's something that that's 00:10:30.839 --> 00:10:37.560 a peak that I've hit several times so you can see Skyrim isn't running 100% 00:10:35.680 --> 00:10:42.200 correctly on this card maybe that's in part due to the early drivers could be 00:10:39.440 --> 00:10:47.000 settings um I also want oh yeah so yeah Power and temperatures instead of 00:10:43.800 --> 00:10:50.040 measuring power and temps uh with like a 00:10:47.000 --> 00:10:52.800 um a synthetic Benchmark what I'm doing 00:10:50.040 --> 00:10:58.399 is I'm looking at it in all of the applications that I'm testing in order 00:10:54.880 --> 00:10:59.920 to get a good solid um average so this 00:10:58.399 --> 00:11:03.000 benchmark's done you can tell cuz the fraps counter has reappeared so I'll 00:11:01.680 --> 00:11:09.040 definitely have to make some notes about the flickering on this one and let's go have a look at the results temperature 00:11:06.839 --> 00:11:13.200 monitoring is done using MSI afterburner so I'm looking for the peak temperature 00:11:10.600 --> 00:11:16.720 which looks like it was 77 here which is consistent with all the other ones this 00:11:14.880 --> 00:11:23.480 kind of stands to reason because what the GTX 680 does is it actually uses its 00:11:20.639 --> 00:11:28.720 thermal uh thermal envelope to determine how much power to throw at a particular 00:11:26.760 --> 00:11:33.880 game uh so it'll actually scale its clock speed let me see if uh if 00:11:31.120 --> 00:11:38.959 afterburner is correctly yeah look at that it is monitoring core clock 00:11:37.040 --> 00:11:43.120 whoops oh no I don't think it's monitoring that correctly well let's not 00:11:41.399 --> 00:11:48.079 worry about that for now then I know that um EVGA Precision is already 00:11:45.880 --> 00:11:52.440 updated for the GTX 680 but none of the other applications are right now so 680 00:11:50.720 --> 00:11:58.160 comes right oh hold on maybe there was another card yeah flickering on the 00:11:54.519 --> 00:12:00.880 590 the 6990 and the 7970 so it looks 00:11:58.160 --> 00:12:06.720 like all of the all the the king performers 00:12:03.680 --> 00:12:08.440 flickering um had the same issue with 00:12:06.720 --> 00:12:16.720 this game so there there you have it so that's the 680 590 580 570 6990 7970 and 00:12:14.360 --> 00:12:21.399 7950 so this is in terms of sheer performance a clear win for the GTX 680 00:12:19.519 --> 00:12:27.079 where it actually beats out everything in this test by slim margin including 00:12:24.600 --> 00:12:30.880 the GTX 590 it definitely felt significantly smoother during use and 00:12:29.199 --> 00:12:34.519 once again that impressive power consumption is right there and we're 00:12:32.760 --> 00:12:39.560 done here Batman Arkham City which ironically is uh the way it's meant to 00:12:36.639 --> 00:12:46.000 be played title is the first game we've seen where the GTX 680 is actually 00:12:42.279 --> 00:12:47.639 outperformed by the Radeon hd7970 could 00:12:46.000 --> 00:12:50.240 be driver issues because one thing I noticed here is that the power 00:12:49.199 --> 00:12:55.279 consumption of the card was significantly lower than it was in the 00:12:52.800 --> 00:13:01.040 other games so what that means to me is maybe the GPU core clock isn't scaling 00:12:57.880 --> 00:13:02.240 in order to uh to reach its maximum TDP 00:13:01.040 --> 00:13:08.079 or maybe there's some driver optimizations that can squeeze some more performance out of this game I guess we 00:13:05.480 --> 00:13:12.839 will have to wait and see so just a reminder this is the 680 the 590 the 580 00:13:11.040 --> 00:13:19.160 the 570 the 6990 the 7970 and the 00:13:16.160 --> 00:13:21.720 7950 which is almost equal with the GTX 00:13:19.160 --> 00:13:26.160 680 in this particular title so that pretty much concludes my performance 00:13:24.120 --> 00:13:29.880 benchmarks so here we go I'm just going to freeze frame for you guys yeah I 00:13:28.120 --> 00:13:33.800 could make graphs but I really don't think that's likely to happen oh idle 00:13:31.639 --> 00:13:39.240 power so idle power all of these new generation cards the 680 as well as the 00:13:36.480 --> 00:13:43.800 7970 and the 7950 are significant improvements over their last generation 00:13:41.399 --> 00:13:49.639 brethren in terms of idle power idle temperatures remain for the most part 00:13:46.199 --> 00:13:54.320 fairly similar though uh so let's go 00:13:49.639 --> 00:13:57.560 ahead and start hiding row so in review 00:13:54.320 --> 00:13:59.880 there's the cards and Battlefield 3 as 00:13:57.560 --> 00:14:04.600 including the power and temperatures of all of these 00:14:01.240 --> 00:14:07.959 games and we can go ahead and hide those 00:14:04.600 --> 00:14:11.759 ones there's crisis 2 including power 00:14:07.959 --> 00:14:11.759 temperature minmax and average 00:14:13.040 --> 00:14:20.720 numbers The Witcher two so there's our power attempts min 00:14:18.120 --> 00:14:24.680 max average remember guys this was completely unplayable whether that's a 00:14:22.320 --> 00:14:31.040 driver issue or whatever else this 30 was not playable unlike the 32 here next 00:14:28.440 --> 00:14:31.040 We've Got Dirt 00:14:36.079 --> 00:14:45.279 three Batman Arkham City which was a Radeon dominated test 00:14:41.639 --> 00:14:48.680 this time around well other than the 590 00:14:45.279 --> 00:14:51.199 but whatever guys all right and finally 00:14:48.680 --> 00:14:57.160 Skyrim which had weird flickering on these four cards and was otherwise uh 00:14:55.120 --> 00:15:02.759 fairly normal so thank you for checking out my performance review of the GeForce 00:14:59.639 --> 00:15:07.199 GTX 680 in a single card configuration 00:15:02.759 --> 00:15:09.480 at 1080p with high SL ultimate or Ultra 00:15:07.199 --> 00:15:13.000 graphic settings on a variety of games don't forget to subscribe to lus 00:15:10.920 --> 00:15:18.399 techtips for more out of focus videos just like this one and thank you for 00:15:15.399 --> 00:15:18.399 watching