1
00:00:00,160 --> 00:00:04,960
This video is brought to you by CM Storm, preferred choice of mice,

2
00:00:03,360 --> 00:00:08,160
headsets, and mechanical keyboards of Curse Gaming. Visit

3
00:00:09,160 --> 00:00:16,720
www.coolermaster-usea.com/team_curse for more details. Guys, welcome to part four

4
00:00:13,040 --> 00:00:18,320
of our 3570K versus FX8350 gaming

5
00:00:16,720 --> 00:00:23,920
showdown. I'm going to start with our platforms. We've got a Crosshair 5

6
00:00:20,160 --> 00:00:25,680
formula with an FX8350 at 4.6 GHz. This

7
00:00:23,920 --> 00:00:29,679
overclock should be attainable by pretty much anyone running this chip on an

8
00:00:27,279 --> 00:00:33,120
overclocking ready platform. Uh, but there is some headroom to keep going.

9
00:00:31,519 --> 00:00:38,000
However, we have an equally conservative overclock on our 3570K running on a

10
00:00:36,000 --> 00:00:41,680
Maximus 5 formula. This guy's running at 4.2 GHz. Now, they're both running

11
00:00:40,160 --> 00:00:46,960
Mushkin Blackline memory. There's copperhead heat spreaders on here, but there's nothing special about those.

12
00:00:44,239 --> 00:00:52,000
It's just regular 1600 MHz memory, and we're running a 128 gig SSD for our boot

13
00:00:49,840 --> 00:00:57,520
drive, an AX850 for our power supply, and an H100 for our cooler. Now, I

14
00:00:55,199 --> 00:01:01,760
specifically called out that you needed an overclocking ready platform. This is

15
00:00:59,520 --> 00:01:05,280
something I wanted to bring up because the cost of the platform, the cost

16
00:01:03,600 --> 00:01:10,400
difference between these platforms is not sheerely dictated by the cost of the

17
00:01:07,600 --> 00:01:15,040
chip itself. So, the FX8350 is about $20 cheaper on average than a 3570K.

18
00:01:12,880 --> 00:01:19,920
However, these two boards actually have a $50 price difference. And if you look

19
00:01:17,360 --> 00:01:25,040
at what the minimum bare minimum Z77 board costs, it's around $120. Whereas

20
00:01:22,720 --> 00:01:28,640
you can get a 970A board as long as you're not planning to run a high-end

21
00:01:26,479 --> 00:01:32,880
dual GPU setup that you need. So even then it supports 8x8x for crossfire

22
00:01:31,119 --> 00:01:36,880
operation. Um you can get one of those for around 80 bucks. So that's something

23
00:01:34,960 --> 00:01:40,320
to consider when you look at the performance numbers we're going to show

24
00:01:38,159 --> 00:01:46,079
you. Methodology wise we're using our usual approach. So our standard card is

25
00:01:42,400 --> 00:01:47,920
an MSI GeForce GTX 660Ti Power Edition.

26
00:01:46,079 --> 00:01:51,520
This guy right here. We run it at a standard overclock that we always run

27
00:01:50,240 --> 00:01:55,840
this card on. You can check out all of our GPU overclocks on any of our video

28
00:01:53,520 --> 00:02:00,079
card reviews. We are using FRAPs to record average frame rates. However, we

29
00:01:58,240 --> 00:02:04,000
are aware of the new methodology of measuring frame times. However, there's

30
00:02:02,240 --> 00:02:08,800
also been I mean it's been an explosive time in GPU benchmarking. Um, and the

31
00:02:06,640 --> 00:02:12,480
correct tools for measuring frame times actually don't yet exist except with

32
00:02:10,879 --> 00:02:17,120
some very specialized equipment. So, we're doing what pretty much everyone else is doing, and we're still using

33
00:02:14,800 --> 00:02:20,160
Fraps to measure frame rates, but expect that to change sometime over the next

34
00:02:18,720 --> 00:02:23,840
little while as more tools become available. We do use in-game runs, so

35
00:02:22,800 --> 00:02:28,640
you guys are going to see little messages pop up if you want to see exactly how we benchmark each game. Just

36
00:02:27,520 --> 00:02:33,280
go ahead and click that, check it out, and then you can run your game the same way that we do and find out how your

37
00:02:31,760 --> 00:02:37,280
system stacks up. So, you're going to see four scenarios on each graph.

38
00:02:34,879 --> 00:02:40,959
Windows 7 with Intel, Windows 7 with AMD, Windows 8 with Intel, Windows 8

39
00:02:39,120 --> 00:02:44,239
with AMD. We're running anti-aliasing on all of the games this time around. We

40
00:02:42,480 --> 00:02:48,480
have a separate video where we ran with no anti-aliasing. And we're going to

41
00:02:46,080 --> 00:02:53,120
start with Far Cry 3. So, in Far Cry 3, the only real change um that happened

42
00:02:51,440 --> 00:02:58,400
was when we went from Windows 7 to Windows 8, both platforms improved in

43
00:02:55,599 --> 00:03:02,640
performance, particularly the Intel one went from 15 frames per second minimum

44
00:03:00,800 --> 00:03:06,640
to 24 frames per second minimum, delivering a much smoother experience,

45
00:03:04,239 --> 00:03:10,400
while the AMD platform was at 20 plus frames per second minimum on both

46
00:03:08,720 --> 00:03:14,959
Windows 7 and Windows 8. So, it looks like Windows 8 comes ahead, but the

47
00:03:12,879 --> 00:03:18,000
actual results in terms of the average frames per second was about the same on

48
00:03:16,640 --> 00:03:23,879
both cards, giving you about the same gaming experience regardless of which chip you choose. Crisis 3 is pretty much

49
00:03:22,080 --> 00:03:29,840
nothing to report here, guys. So, the FX8350 and the 3570K perform about the

50
00:03:27,200 --> 00:03:34,480
same, and they performance doesn't really change from Windows 7 to Windows

51
00:03:31,760 --> 00:03:38,720
8. In Battlefield 3, we saw a bit of a different story with both of them

52
00:03:36,480 --> 00:03:43,280
recording about the same average frame rates. So, a little bit higher with the

53
00:03:40,720 --> 00:03:46,640
8350 edging out the 3570K. Actually, I shouldn't say edging out. It's about,

54
00:03:44,799 --> 00:03:50,959
you know, 5% performance difference. However, it should be noted, but because

55
00:03:48,080 --> 00:03:53,760
we're using real runs, the uh margin of error for our tests is a little bit

56
00:03:52,400 --> 00:04:00,959
higher, although we do run them multiple times to get them as accurate as possible. However, both platforms

57
00:03:57,760 --> 00:04:05,040
increased in minimum frame rates from 35

58
00:04:00,959 --> 00:04:07,360
to 43 on the 8350 and from 37 to 51 when

59
00:04:05,040 --> 00:04:11,360
moving from Windows 7 to Windows 8. So, there was uh there was less less of a

60
00:04:09,599 --> 00:04:15,760
dip when you run into an intensive scenario on Windows 8 versus Windows 7.

61
00:04:14,080 --> 00:04:20,720
In Skyrim, there wasn't really much to report. The 3570K gets a clear win here

62
00:04:18,400 --> 00:04:24,880
versus the 8350, and performance didn't change going from Windows 7 to Windows

63
00:04:22,880 --> 00:04:28,240
8. It should be noted that our Skyrim benchmark is run with about 18 custom

64
00:04:26,800 --> 00:04:31,919
mods enabled, and we haven't released our Skyrim benchmarking guide yet, but

65
00:04:30,320 --> 00:04:35,440
that will hopefully be coming sometime soon. In Dirt 3, this was one of the

66
00:04:34,000 --> 00:04:39,280
more dramatic changes. Without anti-aliasing, both chips were running

67
00:04:37,199 --> 00:04:43,680
at an average around 150 frames per second, enabling anti-aliasing on this

68
00:04:41,360 --> 00:04:48,080
title had them both tanked to around half of what we were seeing previously.

69
00:04:45,680 --> 00:04:51,759
It should also be noted that they didn't really change in performance from

70
00:04:49,199 --> 00:04:55,440
Windows 7 to Windows 8. Once again, Witcher 2, we saw performance

71
00:04:53,759 --> 00:04:59,199
improvements going from Windows 7 to Windows 8 in terms of the minimum frame

72
00:04:57,199 --> 00:05:05,600
rates. So on the AMD platform, we went from 52 to 57 and on the Intel platform,

73
00:05:01,440 --> 00:05:08,639
we went from 51 to 64. So Witcher 2 ran

74
00:05:05,600 --> 00:05:11,120
better on Intel, particularly on well,

75
00:05:08,639 --> 00:05:14,400
not on Windows 7, but on Windows 8, it ran better on Intel, but not on Windows

76
00:05:12,880 --> 00:05:18,960
7. So that was a very interesting observation. It also took the averages

77
00:05:16,960 --> 00:05:22,479
and put about a 10% difference in performance between the two chips.

78
00:05:20,320 --> 00:05:26,960
Whereas on Windows 7, they were much much closer, pretty much within margin

79
00:05:24,160 --> 00:05:32,240
of error. In Metro 2033, we saw a performance improvement for the 8350 of

80
00:05:29,600 --> 00:05:36,000
about 10%. And we saw average frame rates for the Intel stay pretty much the

81
00:05:34,639 --> 00:05:41,680
same. They actually dropped a little bit, but again, margin of error, anything, you know, a few percentage

82
00:05:39,039 --> 00:05:46,400
points, we should pretty much uh not read too much into. Crisis 2, we didn't

83
00:05:44,320 --> 00:05:51,600
really see any optimizations go on from Windows 7 to Windows 8. So, that pretty

84
00:05:49,680 --> 00:05:56,000
much leaves us with Intel being the clear winner for Crisis 2 if you still

85
00:05:53,360 --> 00:06:00,560
play that game, which I suspect most people do not, but we like to include it

86
00:05:58,400 --> 00:06:06,080
just because of the old, can it run Crisis? So, there you go, guys. Windows

87
00:06:03,440 --> 00:06:10,479
8 on the surface looks like the optimal gaming platform whether you go AMD or

88
00:06:08,479 --> 00:06:14,319
Intel. And I mean, quite frankly, it's up to the individual viewer whether you

89
00:06:12,560 --> 00:06:16,800
watch a video like this and go, "This performance difference is worth this

90
00:06:15,759 --> 00:06:21,120
number of dollars to me." We're not going to make that decision for you. We're going to let you go do that. But

91
00:06:19,440 --> 00:06:24,560
it looks like Windows 8 is the optimal gaming platform. However, there were

92
00:06:22,880 --> 00:06:28,240
some issues that we encountered running our games on Windows 8, such as the

93
00:06:26,160 --> 00:06:32,479
requirement to install games for Windows Live before actually installing any

94
00:06:30,319 --> 00:06:36,080
games, otherwise it causes all kinds of halaloo and games just not launching and

95
00:06:34,560 --> 00:06:39,280
things like that, and some other random glitchy behavior. It should be noted

96
00:06:37,840 --> 00:06:44,759
though that with some googling we did get all of the games working. So if you

97
00:06:42,080 --> 00:06:48,960
want that extra little bit of performance, that's me, then it might be

98
00:06:47,520 --> 00:06:55,680
the way to go. Thanks for checking out this video. Don't forget to subscribe to Linus Tech Tips for unboxings, reviews,

99
00:06:52,319 --> 00:06:55,680
and other computer videos.