1
00:00:00,000 --> 00:00:05,359
Another day, another new graphics card.

2
00:00:03,280 --> 00:00:13,920
Welcome to the Lionus Tech Tips performance overview of the GTX 980 from

3
00:00:08,559 --> 00:00:13,920
NVIDIA, their most advanced GPU

4
00:00:18,279 --> 00:00:23,279
yet. DKL's new Rip Jaws 4 series is

5
00:00:21,279 --> 00:00:27,439
their ultimate DDR4 memory with speeds of up to

6
00:00:24,359 --> 00:00:29,519
3,333 MHz available. Click now to learn

7
00:00:27,439 --> 00:00:34,000
more. As usual, we'll start with the unboxing and physical overview. The

8
00:00:31,519 --> 00:00:40,719
Media Sample GTX 980s come packaged in this sweet box that has the card upright

9
00:00:37,920 --> 00:00:47,200
looking all majestic with its magnesium alloy fan housing, aluminum frame, and

10
00:00:43,719 --> 00:00:48,600
trivalent chromium plating with H.

11
00:00:47,200 --> 00:00:53,199
What's that? Sorry, you've seen this user

12
00:00:50,800 --> 00:00:58,320
configurable green LED lit GeForce logo before. NVIDIA hasn't really changed

13
00:00:55,680 --> 00:01:03,359
their cooler design since the 690. Ah, yes. No, I can see how you might think

14
00:01:00,399 --> 00:01:08,479
that, but you haven't seen the back yet. Check it out. Now, we get a full metal

15
00:01:06,479 --> 00:01:12,799
back plate with a classy looking design and surface. Now, the issue with back

16
00:01:11,040 --> 00:01:17,439
plates, and presumably the reason NVIDIA has never done a reference card with one

17
00:01:14,479 --> 00:01:22,000
before, is that it restricts the intake of a card installed in the next slot

18
00:01:19,920 --> 00:01:26,159
over in tightly packed SLI configurations. So NVIDIA made this

19
00:01:24,400 --> 00:01:32,000
strategic chunk of the back plate removable for users with lots of cards

20
00:01:29,280 --> 00:01:37,280
right next to each other in SLI. Not that you need to run SLI just to hook up

21
00:01:35,040 --> 00:01:42,159
enough monitors anymore. As we move over to the IO, yeah, that's some pretty good

22
00:01:39,600 --> 00:01:47,439
stuff right there. Three DisplayPort ports, so that's up to three G-Sync

23
00:01:44,400 --> 00:01:50,479
displays off of a single card. An HDMI

24
00:01:47,439 --> 00:01:54,320
2.0 port. So yeah, that's 4K 60 Hz over

25
00:01:50,479 --> 00:01:56,720
HDMI. and a dual link DVI port can all

26
00:01:54,320 --> 00:02:01,280
be found here. Any four of which can be used at one time on this beast of a

27
00:01:59,280 --> 00:02:07,920
graphics card. Which leads us to the next notable thing physically. This is a

28
00:02:04,040 --> 00:02:12,319
flagship 80 class card, but like the GTX

29
00:02:07,920 --> 00:02:13,920
680, it only has two six pin connectors.

30
00:02:12,319 --> 00:02:19,040
A hint at the incredible power efficiency of the Maxwell architecture

31
00:02:16,319 --> 00:02:24,160
at the core of the GM 204 chip that powers this card. The GTX 980 not only

32
00:02:22,480 --> 00:02:30,080
hung around at the top of our performance charts, but did it while

33
00:02:26,480 --> 00:02:33,519
consuming less power than GK 110based

34
00:02:30,080 --> 00:02:35,519
cards like the 780 and 780 Ti and while

35
00:02:33,519 --> 00:02:39,840
running extremely quietly. So, let's talk about Maxwell for a bit here.

36
00:02:37,599 --> 00:02:43,519
Without shrinking the manufacturing process, so they're still on the same

37
00:02:41,519 --> 00:02:49,440
28nmter node that they used to build the Kepler based GTX 680, NVIDIA has

38
00:02:46,319 --> 00:02:51,239
effectively reduced TDP by 30 W,

39
00:02:49,440 --> 00:02:56,800
increased the number of CUDA cores by 33%, increased base clocks and memory

40
00:02:54,239 --> 00:03:01,760
speed by around 12% and doubled the number of render output units, as well

41
00:02:59,200 --> 00:03:06,640
as adding some other cool hardware bits to enable some really cool technology.

42
00:03:04,080 --> 00:03:11,120
Pretty amazing, actually. So, first up in the new technology list is their

43
00:03:09,040 --> 00:03:15,280
improved third generation delta color compression. A technique that allows

44
00:03:13,120 --> 00:03:20,400
data coming out of the core to be compressed by a factor of 2:1, 4:1, or

45
00:03:18,000 --> 00:03:25,120
even 8:1 before being committed to memory by storing the value of a change

46
00:03:23,120 --> 00:03:30,159
versus the previous frame rather than the value itself. This combined with

47
00:03:27,519 --> 00:03:36,959
improved caching allows the GTX 980, in spite of its mere 256-bit memory

48
00:03:33,280 --> 00:03:39,280
interface, to achieve up to around 30%

49
00:03:36,959 --> 00:03:44,319
better effective memory performance compared to the GTX 680. In this scene,

50
00:03:42,400 --> 00:03:48,640
the pink highlighted screen elements are taking advantage of this technology.

51
00:03:46,560 --> 00:03:52,319
Next up is a feature that's all about getting the most out of older games

52
00:03:50,239 --> 00:03:58,640
where you have lots of GPU horsepower to spare. DSR or dynamic super

53
00:03:55,799 --> 00:04:04,120
resolution is mostly just another name for a feature that has existed forever

54
00:04:00,879 --> 00:04:06,319
but is now easier to use. Super sampling

55
00:04:04,120 --> 00:04:10,239
anti-aliasing. The way it works is it renders the scene at a higher

56
00:04:08,000 --> 00:04:14,640
resolution. So you can choose 2x, 3x, or 4x in the control panel. Then

57
00:04:12,480 --> 00:04:19,120
intelligently downscales it with a filter whose sharpness or softness can

58
00:04:16,560 --> 00:04:24,880
be adjusted to let you use your extra GPU horsepower that otherwise was just

59
00:04:22,160 --> 00:04:29,680
sitting there to yield potentially a much more pleasing gaming experience.

60
00:04:27,040 --> 00:04:35,440
It's not the same as actually running on a 4K display, but it's much better than

61
00:04:32,960 --> 00:04:39,680
running native 1080p. The effect is especially noticeable on the grass in

62
00:04:37,840 --> 00:04:44,080
Dark Souls II here in this footage that I took at the NVIDIA event. The next

63
00:04:42,400 --> 00:04:49,880
feature is about getting more performance when we don't have enough,

64
00:04:46,000 --> 00:04:52,199
and it's called MFAA or frame

65
00:04:49,880 --> 00:04:57,040
anti-aliasing. That's short for multi-frame sampled anti-aliasing. Grow

66
00:04:54,720 --> 00:05:01,919
up. The way this works is instead of taking a large number of samples with a

67
00:04:59,840 --> 00:05:08,000
static pattern to smooth the edges of on-screen objects or textures, MFAA

68
00:05:05,040 --> 00:05:13,440
takes fewer samples per frame, half as many from what we've been told, but uses

69
00:05:10,160 --> 00:05:16,400
new hardware in the GTX 980 and 970 to

70
00:05:13,440 --> 00:05:21,199
alternate the positions with each frame. Then it applies a clever filter. So the

71
00:05:18,479 --> 00:05:26,479
performance impact is about equivalent to two times MSAA, but the resulting

72
00:05:24,320 --> 00:05:31,759
image looks pretty much identical to four times MSAA. This feature isn't

73
00:05:29,600 --> 00:05:34,639
enabled in the driver yet, so expect an update once we get a chance to play

74
00:05:33,199 --> 00:05:38,960
around with it some more, but the initial results are impressive. Check

75
00:05:37,199 --> 00:05:43,360
out the effect on the transparent textures like these wires in this

76
00:05:41,120 --> 00:05:46,880
Battlefield 4 demo that NVIDIA was showing off at the launch event. Another

77
00:05:45,199 --> 00:05:50,720
thing we won't be able to test for a while is the improvements that NVIDIA

78
00:05:48,800 --> 00:05:55,680
has been making to the VR experience on their cards. The most notable of which

79
00:05:52,400 --> 00:05:57,680
is called auto asynchronous warp. A

80
00:05:55,680 --> 00:06:03,840
cheat, I guess, that allows the head tracking of a VR headset like the Oculus

81
00:05:59,840 --> 00:06:06,560
Rift to sample its position again right

82
00:06:03,840 --> 00:06:11,360
before a frame is finished being drawn to reduce the player's perception of the

83
00:06:08,960 --> 00:06:13,919
delay between physical motion and on-screen motion. And this will be

84
00:06:12,720 --> 00:06:18,639
important for folks who have trouble with motion sickness in VR. Which leads

85
00:06:16,080 --> 00:06:22,639
us to the final feature, their VxGI global illumination technique that we

86
00:06:20,479 --> 00:06:26,080
also can't really test here because while Unreal Engine 4 is getting

87
00:06:24,319 --> 00:06:30,319
support, it'll be a while before we get real games with the tech, but it looks

88
00:06:28,000 --> 00:06:34,400
so freaking cool. NVIDIA has a couple of demos to show it off. One where they

89
00:06:31,759 --> 00:06:37,680
render a Cornell box, so you can see the way the direct light enters at the top,

90
00:06:36,400 --> 00:06:43,520
so the tops of the shapes are the brightest with realistic soft shadows and color bleeding from reflective

91
00:06:40,720 --> 00:06:48,240
surfaces. and one where they recreated as closely as possible the conditions of

92
00:06:45,759 --> 00:06:53,600
this iconic photo from the 1969 lunar landing to demonstrate that it's

93
00:06:50,560 --> 00:06:56,400
completely realistic and wasn't taken

94
00:06:53,600 --> 00:07:00,720
somewhere in Area 51 or whatever. Now, these two demos wouldn't be impressive

95
00:06:58,240 --> 00:07:06,080
at all except that they're being done in real time instead of using path tracing,

96
00:07:03,759 --> 00:07:09,759
a phenomenally computationally intensive global illumination technique that

97
00:07:07,520 --> 00:07:14,000
involves following light rays as they bounce around the scene. like all of

98
00:07:11,440 --> 00:07:19,360
them. NVIDIA is using dedicated hardware to split the scene into cubes called

99
00:07:16,319 --> 00:07:22,160
voxels that act as dynamic light sources

100
00:07:19,360 --> 00:07:28,880
that approximate the results of a more accurate rayraced rendering method but

101
00:07:25,280 --> 00:07:32,080
in real time. All right, so as cool as

102
00:07:28,880 --> 00:07:34,240
all that technology is, how does the

103
00:07:32,080 --> 00:07:38,240
card actually perform? That's what you want to know. Great question, Luke. Take

104
00:07:36,560 --> 00:07:41,440
it away. Before we get into things, I want to point out that overclocking all

105
00:07:39,680 --> 00:07:45,039
of our cards is yet again backed by popular demand. And because of that, you

106
00:07:43,520 --> 00:07:48,880
will be able to find the overclocking results and settings in the graphics

107
00:07:46,960 --> 00:07:51,840
card overclocking dock, which should be linked in the description below this

108
00:07:50,240 --> 00:07:56,560
video. Also, when reading the settings for the GTX 980 in the overclocking dock

109
00:07:54,319 --> 00:08:00,319
below, there's no reason to worry. There is a reason why I ignore the memory

110
00:07:58,080 --> 00:08:04,400
clock speed. The 980, much like the 780 Ti before it, features a fantastic

111
00:08:02,319 --> 00:08:08,319
technology called autobalancing power rails, which in its most basic

112
00:08:06,560 --> 00:08:12,080
description, allows the card to take extra power overhead from parts of the

113
00:08:10,319 --> 00:08:16,400
card that aren't using it and move it elsewhere, allowing you to focus on one

114
00:08:14,080 --> 00:08:20,639
specific thing, which in this case resulted in the GPU clock speed being

115
00:08:18,639 --> 00:08:24,160
cranked through the freaking roof, which is awesome because then you can access

116
00:08:22,479 --> 00:08:28,080
all of that FPS that you wouldn't otherwise be able to. Performance of the

117
00:08:25,919 --> 00:08:32,080
GTX 980 is honestly a little bit surprising in a good way. To give proper

118
00:08:30,720 --> 00:08:37,760
context for that statement, we have to look at the rumor mill. Personally, I like to avoid rumors, but sometimes I

119
00:08:35,919 --> 00:08:42,159
don't really have a choice. For instance, on the September 12th version

120
00:08:39,360 --> 00:08:47,040
of the WAN Show, my brother at Cold SE on Twitter, and I discussed how the 980

121
00:08:44,480 --> 00:08:51,200
was, according to rumors, not supposed to beat out the 780 Ti. So, when I

122
00:08:49,760 --> 00:08:58,399
finally got the card in hand, I was delighted to see that more often than not, it was able to beat the 780 Ti in

123
00:08:55,120 --> 00:09:00,560
both 1080p and 4K benchmarks. Now, my

124
00:08:58,399 --> 00:09:07,120
overclock for the 980 feels a little bit high, running at a range of 1503 to 1492

125
00:09:04,040 --> 00:09:08,880
MHz, but I really have no idea as I have

126
00:09:07,120 --> 00:09:14,000
no one else's overclocks to compare to yet. And to be fair, my 780 Ti overclock

127
00:09:11,680 --> 00:09:19,200
is very aggressive anyways. So, I haven't exactly done the 980 any favors

128
00:09:16,000 --> 00:09:20,800
in regards to the 780 Ti's clock speed.

129
00:09:19,200 --> 00:09:24,560
But don't forget, all of this performance awesomeness doesn't just

130
00:09:22,560 --> 00:09:28,880
stand on its own. Not only are you paying considerably less for this card

131
00:09:26,560 --> 00:09:33,120
than you would pay for a 780 Ti, but you also pay less in the future. Underload,

132
00:09:31,120 --> 00:09:37,920
the 980 has considerably less power draw than the 780 Ti, which is nice if you're

133
00:09:35,680 --> 00:09:40,720
actually paying for your power bill. And if you're not paying for your power

134
00:09:39,360 --> 00:09:45,360
bill, maybe you can use that little tidbit of information to convince your parents that it might be time for an

135
00:09:43,680 --> 00:09:48,880
upgrade. Anyways, thanks for watching my segment, guys. Back to you, Linus.

136
00:09:46,959 --> 00:09:52,320
Thanks, Luke. All right, guys. So, you've seen the results and you've seen

137
00:09:50,560 --> 00:09:56,360
the pricing. Links as always in the video description for pricing. So, let

138
00:09:54,240 --> 00:09:59,519
us know what you think. For non-overclockers, as I'm sure you'll

139
00:09:58,240 --> 00:10:08,480
have noticed from the other reviews around the web, GTX 980 is a total tank.

140
00:10:04,399 --> 00:10:11,519
But when overclocking the GTX 780 Ti, we

141
00:10:08,480 --> 00:10:14,079
observed that this older GK 1110-based

142
00:10:11,519 --> 00:10:18,320
card becomes much more competitive if you're willing to sacrifice features and

143
00:10:16,240 --> 00:10:23,440
power consumption. So, let us know what you think. If you had over 500 bucks to

144
00:10:20,959 --> 00:10:27,680
spend on a graphics card, which one would get your money? Also, let us know

145
00:10:25,920 --> 00:10:31,839
in the comments what other coverage you'd like to see of this card. An

146
00:10:29,360 --> 00:10:35,920
overclocking guide, um, SLI performance scaling. Any other ideas? let us know.

147
00:10:34,480 --> 00:10:39,040
Thanks for watching, guys. Like this video if you liked it. Dislike it if you

148
00:10:37,760 --> 00:10:43,440
disliked it. You can also leave a comment if you have feelings that are just more complicated than this. Also,

149
00:10:42,399 --> 00:10:48,320
check out the link in the video description to support us if you want to buy a cool t-shirt, give us a monthly

150
00:10:46,880 --> 00:10:51,680
contribution, or change your Amazon bookmark to one with our affiliate code.

151
00:10:49,920 --> 00:10:54,800
So, when you buy your new graphics card, we'll get a small kickback as a, you

152
00:10:53,360 --> 00:10:58,240
know, thank you for providing us with these videos, Lionus. We sure do think

153
00:10:56,560 --> 00:11:02,320
you're awesome. Here's your affiliate code being used in Amazon. Helps us out

154
00:11:00,160 --> 00:11:24,010
a lot. Thanks again for watching. And as always, don't forget to subscribe.