WEBVTT 00:00:07.439 --> 00:00:15.519 welcome to a very exciting unboxing the new single GPU graphics card King is 00:00:12.519 --> 00:00:17.560 here this is the GeForce GTX 680 from 00:00:15.519 --> 00:00:24.680 Galaxy it's got a three-year warranty 2 gigs of gddr5 memory at 6 00:00:21.359 --> 00:00:26.480 GHz uh features direct X11 as have the 00:00:24.680 --> 00:00:31.800 last couple of generations of GeForce graphics cards SLI fizx and 3D Vision 00:00:29.519 --> 00:00:37.760 something that is not mentioned on the box though that it also supports is 00:00:33.840 --> 00:00:41.440 NVIDIA surround in 3D Vision off of a 00:00:37.760 --> 00:00:43.120 single card so uh NVIDIA has added that 00:00:41.440 --> 00:00:48.000 functionality to their card you can hook up up to four displays in three gaming 00:00:46.280 --> 00:00:52.320 plus one auxiliary and you can have 3D division 00:00:50.360 --> 00:00:55.440 running at the same time off a single card so hold on let's see if there's 00:00:53.719 --> 00:01:00.199 anything going on here ah yes key features NVIDIA GPU boost more on that 00:00:58.079 --> 00:01:07.000 in a moment adaptive vertical syn is cool because what it allows you to do is 00:01:04.199 --> 00:01:12.479 turn on NVIDIA's adaptive vertical SN so vsync as you guys may or may not know 00:01:10.159 --> 00:01:16.400 will prevent tearing so that is those artifacts that look like you got half of 00:01:14.439 --> 00:01:21.759 one frame and half of another frame sort of slightly offset as you turn in game 00:01:19.479 --> 00:01:26.799 or or move around in game so vsync prevents that but also locks your frame 00:01:23.479 --> 00:01:29.560 rate at 60 FPS and it can cause some uh 00:01:26.799 --> 00:01:34.960 you know Mouse leg and stuff like that so adaptive vsync turns vsync on to 00:01:32.759 --> 00:01:40.560 prevent tearing and then off when your frame rate is below 00:01:38.200 --> 00:01:44.600 60 so what that means is that when you're below 60 what can happen a lot of 00:01:42.479 --> 00:01:50.079 the time is you'll be locked into a frame rate that is a multiple of 60 like 00:01:46.759 --> 00:01:54.439 30 or 20 or 15 instead of running at you 00:01:50.079 --> 00:01:56.039 know 42 or 25 so you'll be able to 00:01:54.439 --> 00:02:01.759 squeeze the most performance out of your card when it's not running at 60 FPS but 00:01:58.640 --> 00:02:03.560 be locked at 60 when it is capable of it 00:02:01.759 --> 00:02:07.399 in order to prevent tearing so that's kind of a neat feature man that was 00:02:05.600 --> 00:02:10.599 really tight in there okay let's keep going here NVIDIA surround which I 00:02:09.160 --> 00:02:15.319 mentioned already four concurrent displays including two dual link DVI 00:02:12.200 --> 00:02:18.400 HDMI and DisplayPort 1.2 this card also 00:02:15.319 --> 00:02:21.319 includes support for 4K displays so that 00:02:18.400 --> 00:02:29.080 is super high resolution displays direct X11 with direct compute 5.0 fiz X yes 00:02:24.440 --> 00:02:32.080 cool 3D Vision SLI Cuda PCI Express 3.0 00:02:29.080 --> 00:02:35.239 this is the first PCI Express 3.0 00:02:32.080 --> 00:02:37.879 graphics card from NVIDIA's 00:02:35.239 --> 00:02:42.080 Partners it's suggesting a minimum of 550 W power supply but what you guys are 00:02:40.400 --> 00:02:45.000 going to find out about this when I do my review on it which is probably going 00:02:43.519 --> 00:02:50.040 to go up around the same time as this unboxing is that the power consumption 00:02:47.040 --> 00:02:51.480 is actually phenomenal for a high-end 00:02:50.040 --> 00:02:58.640 card so thank you for purchasing something something set up easy quick steps for quick installation guide and 00:02:55.280 --> 00:03:01.080 whatnot okay cool and then we've got the 00:02:58.640 --> 00:03:04.680 card itself so let's hold on let's see what galaxy includes in terms of 00:03:02.760 --> 00:03:09.599 accessories and then we'll start talking about some of the additional technology 00:03:07.080 --> 00:03:16.239 points that I want to cover about this graphics card here we go if I can figure 00:03:13.560 --> 00:03:23.560 out how this comes apart ah very nice okay so we have a DVI to VGA 00:03:19.760 --> 00:03:27.680 adapter we have two Molex to single PCIe 00:03:23.560 --> 00:03:30.040 6 pin two Molex to single PCI six pin 00:03:27.680 --> 00:03:34.200 that's funny guys high-end graphics card two six pin adapters we'll more on that 00:03:32.239 --> 00:03:38.599 in a moment uh graphics card driver installation disc download the latest as 00:03:35.760 --> 00:03:45.200 well as a user manual okay so fairly straightforward accessory package 00:03:42.080 --> 00:03:51.470 here nice packaging overall though looks 00:03:45.200 --> 00:03:59.060 good very black all right now the card 00:04:02.680 --> 00:04:10.680 okay now in terms of its physical appearance it's a very similar looking 00:04:07.120 --> 00:04:13.079 to a GTX 580 in terms of its size in 00:04:10.680 --> 00:04:16.239 terms of the the cooler although there are some subtle differences but let's 00:04:14.680 --> 00:04:22.160 have a look at the overall physical aspects of the card first so we've got two SLI connectors meaning this card 00:04:19.639 --> 00:04:26.720 does support three-way or four-way SLI there's our little Galaxy sticker it's 00:04:23.800 --> 00:04:30.240 got a green GeForce GTX thing going on unlike the GTX 590 it doesn't look like 00:04:28.759 --> 00:04:35.240 it lights up or anything but it does look pretty sharp I'll give them that 00:04:32.039 --> 00:04:38.360 this is cool check this out two six pin 00:04:35.240 --> 00:04:40.160 PCIe connectors also what NVIDIA's done 00:04:38.360 --> 00:04:43.840 by putting the two six pin PCIe connectors in the same place is they've 00:04:41.720 --> 00:04:48.160 bought themselves some extra room on the PCB to do other things you can see 00:04:46.680 --> 00:04:54.000 actually oh look at this looks like it'll be up to uh it'll 00:04:51.440 --> 00:04:57.080 be up to oh that's bizarre it looks like either they changed their mind at some 00:04:55.560 --> 00:05:01.400 point between having a six and an eight pin and then they did like two six pins 00:05:00.000 --> 00:05:05.000 but then looks like you could either put them here and here or here and here 00:05:03.360 --> 00:05:08.080 maybe it's up to the board Partners at some point in the future to decide how 00:05:06.759 --> 00:05:11.600 they want to implement it but there you go this is the OEM One and NVIDIA says 00:05:10.320 --> 00:05:15.280 that bought them some room for additional PCB components versus doing 00:05:14.039 --> 00:05:22.120 it the other way it should also be fairly tidy to Cable manage this way because you can either come up in one 00:05:19.840 --> 00:05:25.960 column or down in one column from a cable management hole so that's very 00:05:23.880 --> 00:05:30.319 very cool we've got the usual NVIDIA blower style fan that we find on most 00:05:27.840 --> 00:05:34.800 OEM cards with Galaxy's graph on the Shroud itself on the back of the 00:05:33.120 --> 00:05:38.600 card we find not a whole lot we find the PCB is Right flush with the cooler this 00:05:36.840 --> 00:05:45.440 is something that's very confusing to me about AMD cards and that is uh why AMD 00:05:43.440 --> 00:05:49.440 insists on continuing to do this where they have the PCB ending here and then 00:05:47.280 --> 00:05:54.080 they just have an additional like random bunch of plastic that makes the card 00:05:51.680 --> 00:05:59.199 unnecessarily up to a centimeter longer than it actually is I ran into trouble 00:05:55.960 --> 00:06:01.919 installing that 7970 into a shuttle xpc 00:05:59.199 --> 00:06:07.599 barebone the other day because of that extra overhang there so NVIDIA does not 00:06:04.360 --> 00:06:09.280 do that here's our PCI Express 3.0 16x 00:06:07.599 --> 00:06:15.160 slot so that's capable of up to two times the bandwidth of a PCI Express 2.0 00:06:12.520 --> 00:06:19.039 16x slot on the back we find a bunch of other components and whatnots so you can 00:06:17.240 --> 00:06:24.280 see where the RAM chips are installed all on the other side there's no shroud 00:06:20.840 --> 00:06:26.759 on the back of this card so you can see 00:06:24.280 --> 00:06:30.800 the bare components you can see the GPU is right under there and then on the 00:06:28.960 --> 00:06:36.000 back we find the io so there's our two dual link DVI uh this guy is a dvii this 00:06:34.039 --> 00:06:41.520 is a DVI D so the difference is that this one supports analog output we've 00:06:38.400 --> 00:06:43.160 got HDMI as well as DisplayPort 1.2 and 00:06:41.520 --> 00:06:48.440 then we've got a little vent here and a bigger vent here and this is kind of cool check this out so you guys see how 00:06:46.280 --> 00:06:52.800 far away the cooling fins are in there can you see that really um the actual 00:06:51.120 --> 00:06:55.879 fins cooling the GPU I guess that's kind of tough isn't it okay I'll try and get 00:06:54.520 --> 00:07:00.039 you guys a better angle okay can you see in there how they're angled they're 00:06:57.800 --> 00:07:03.840 swept back like that yeah excellent okay so NVIDIA's done 00:07:02.080 --> 00:07:09.319 that for a very specific reason it gives some Dead Space Between the actual fins 00:07:06.160 --> 00:07:11.720 of the GPU heat sink and the Shroud at 00:07:09.319 --> 00:07:16.840 the back so this is much like using a spacer on a radiator in order to 00:07:14.039 --> 00:07:22.000 eliminate uh any kind of dead zones and as well as to more evenly push out the 00:07:20.000 --> 00:07:27.479 air so so you give the air a little bit of room to to accumulate and then come 00:07:25.360 --> 00:07:31.879 out evenly so I'm not describing this in the best possible terms honestly it's 00:07:29.080 --> 00:07:36.240 late and I'm tired but but the point is that it you are able to exhaust air more 00:07:33.720 --> 00:07:41.840 efficiently with less resistance from the see the the metal pieces in the 00:07:38.800 --> 00:07:43.199 Shroud and it's generally better some of 00:07:41.840 --> 00:07:47.120 the air is going to come out into your case so there's a gap here there's a gap 00:07:45.280 --> 00:07:51.000 here but most of the exhaust from this blower design is going to go out the 00:07:49.159 --> 00:07:55.120 back of the case so if you do have a case that doesn't have great ventilation 00:07:53.120 --> 00:08:02.720 this is usually a better choice than going with uh the more typical 00:07:57.479 --> 00:08:02.720 aftermarket designs that you see such as 00:08:03.960 --> 00:08:11.919 uh ah these kinds of dual fan implementations that blow all the air 00:08:08.840 --> 00:08:13.800 into the case all right moving on here 00:08:11.919 --> 00:08:20.360 so what else what else we got here ah GPU boost so this card has 195 wat TDP 00:08:17.560 --> 00:08:23.759 so that is thermal design power and what that means is that the clock speed 00:08:22.479 --> 00:08:27.879 whenever you're designing a graphics card you want a particular thermal 00:08:26.240 --> 00:08:31.360 envelope that you have to meet that the reason is a for power delivery you have 00:08:30.440 --> 00:08:36.399 to make sure that it's going to be within spec and B for heat dissipation 00:08:34.279 --> 00:08:41.399 so you can only dissipate so much heat with your reference cooler you don't 00:08:37.680 --> 00:08:42.680 want the uh the actual GPU to start uh 00:08:41.399 --> 00:08:48.320 producing more heat than you can dissipate because then you're going to run into trouble okay so what that means 00:08:46.760 --> 00:08:53.480 is that when you design a graphics card you go okay I can apply this much voltage I can apply this much clock 00:08:50.959 --> 00:08:57.080 speed and it stays within this TDP now NVIDIA took a bit of an Innovative 00:08:54.959 --> 00:09:02.760 solution that uh may or may not have been inspired by something like Intel's 00:08:59.760 --> 00:09:05.959 uh turbo boost and that is they allow 00:09:02.760 --> 00:09:08.320 their GPU to turn its clock speed up in 00:09:05.959 --> 00:09:14.320 less demanding games and applications to achieve overclocked 00:09:11.160 --> 00:09:16.880 settings on its own so their GPU boost 00:09:14.320 --> 00:09:25.000 technology is capable of taking the 1,06 MHz clock speed up to five 00:09:21.360 --> 00:09:27.120 typically 5% or as much as 10% above 00:09:25.000 --> 00:09:31.399 that stock clock speed as long as it stays within the thermal design 00:09:29.480 --> 00:09:36.480 specification so that's one really cool feature GPU boost okay FXAA has been 00:09:34.839 --> 00:09:40.720 added to the NVIDIA control panel so this is a faster form of anti-aliasing 00:09:39.040 --> 00:09:44.000 and you can now force it on in games even the ones that don't support it so 00:09:42.440 --> 00:09:48.959 that's a really neat thing as well we already talked about adaptive vsync txaa 00:09:46.240 --> 00:09:52.640 modes 1 and two so NVIDIA claims and I'm going to be hopefully taking a look at 00:09:50.600 --> 00:09:57.040 this in the future that mode one is about equivalent to a times AA with the 00:09:55.000 --> 00:10:01.880 performance hit of only two times anti-aliasing and mode two provides 00:10:00.079 --> 00:10:06.720 significantly better image quality than mode one with only the performance hit 00:10:03.920 --> 00:10:10.399 of four times um multi- sampling anti-aliasing so the more traditional 00:10:08.440 --> 00:10:16.360 kind so that's really neat as well we've covered 3.1 3+1 displays 4K monitors 00:10:13.800 --> 00:10:20.600 support better physics performance uh nvnc so video encoding Hardware can run 00:10:18.920 --> 00:10:24.839 faster than the Cuda based Solutions in the past and can also run concurrently 00:10:23.160 --> 00:10:29.959 with the Cuda course on the graphics card so you can allow the Cuda course to 00:10:26.240 --> 00:10:32.760 handle pre-processing while the nvnc so 00:10:29.959 --> 00:10:36.240 nvn I guess that's short for encoder uh can handle the encoding the three heat 00:10:34.720 --> 00:10:39.839 pipe dual slot fan I guess we can't really see that but the way this works 00:10:38.160 --> 00:10:43.560 is instead of using a vapor chamber like we've seen with a couple previous uh 00:10:41.839 --> 00:10:48.959 generation graphics cards from both sides uh which is just like a big long 00:10:45.600 --> 00:10:51.079 kind of like a big Square very flat heat 00:10:48.959 --> 00:10:55.279 pipe it uses three heat pipes one in an S shape and then two in C shapes coming 00:10:53.519 --> 00:11:00.200 out from the GPU to keep things nice and flat to evenly distribute heat to the 00:10:57.800 --> 00:11:04.760 aluminum fins inside finally it has 1,532 Kuda cores it 00:11:03.360 --> 00:11:08.480 clocked at oh I mentioned this sort of briefly but it's clocked at just over 1 00:11:06.519 --> 00:11:14.959 GHz stock speed although it'll boost up from there as well as 6 GHz on the 00:11:11.079 --> 00:11:17.399 memory and finally another cool thing 00:11:14.959 --> 00:11:22.639 about GPU boost is that it will actually scale with overclocking on the card so 00:11:20.160 --> 00:11:27.360 you can overclock your stock speed and your GPU boost will actually scale up 00:11:24.959 --> 00:11:31.600 from there as well and while I haven't tested it yet I'm really hoping to to 00:11:29.440 --> 00:11:36.120 see some interesting overclocking results with this particular GPU so 00:11:33.680 --> 00:11:41.160 NVIDIA claims two times the performance per watt of the two generations ago 00:11:38.839 --> 00:11:46.680 Flagship GTX 480 meaning we've got significantly better performance at uh 00:11:44.120 --> 00:11:51.880 what is apparently with based on the 195 wat TDP significantly better power 00:11:49.839 --> 00:11:55.440 consumption so that means us SLI users out there are in for a treat because 00:11:53.600 --> 00:11:59.720 your power bill is about to go down when you upgrade to a couple of these puppies 00:11:57.519 --> 00:12:02.680 so I think that pretty much much covers it thank you for checking out my 00:12:00.920 --> 00:12:07.959 unboxing and first look at the GeForce GTX 680 from Galaxy and don't forget to 00:12:06.040 --> 00:12:11.200 subscribe for more unboxings reviews and other computer videos