1
00:00:00,080 --> 00:00:06,480
in our review of the core i9 10 900k we were impressed by the performance of

2
00:00:04,319 --> 00:00:12,480
Intel's 10 core beast i mean this thing turbos just shy of 5 gigahertz across

3
00:00:09,760 --> 00:00:17,119
all cores and if you set your motherboard to allow it to ignore the

4
00:00:14,639 --> 00:00:21,359
time limits on its turboing it'll actually do that basically indefinitely

5
00:00:19,760 --> 00:00:26,880
crazy right well crazy hot the one thing we didn't tell

6
00:00:25,359 --> 00:00:31,840
you guys because we accidentally left that in the slide out of the slide

7
00:00:29,199 --> 00:00:35,920
it the one thing we didn't tell you guys because we accidentally left that out of

8
00:00:33,520 --> 00:00:40,960
the slide is what kind of cooling it takes to get that sort of performance so

9
00:00:38,320 --> 00:00:46,079
to rectify that error and take things a little bit deeper we are going to try it

10
00:00:43,520 --> 00:00:50,719
out with everything from Intel's pin or old stock cooler to a more sensible air

11
00:00:48,960 --> 00:00:55,280
cooler to a high-end air cooler to a top-tier air cooler and see just what

12
00:00:53,680 --> 00:00:59,600
kind of performance we can get out of this chip depending

13
00:00:57,360 --> 00:01:03,359
on the cooling solution and this video is brought to you by

14
00:01:01,280 --> 00:01:07,119
glasswire glasswire is the tool that tells you when a strange new app

15
00:01:04,799 --> 00:01:11,600
accesses your network in real time and lets you block it use offer code Linus

16
00:01:09,040 --> 00:01:13,840
to get 25 off at the link in the video description

17
00:01:21,360 --> 00:01:28,240
Anthony do we want to replicate our initial test from the review with a

18
00:01:25,759 --> 00:01:32,240
Corsair 280 millimeter aio this is standard for our CPU reviews or do we

19
00:01:30,400 --> 00:01:35,600
want to go straight to the pain train

20
00:01:34,320 --> 00:01:41,040
let's do the pain drink let's do the pain training all right we'll shut it down we're shutting we're shutting the

21
00:01:38,560 --> 00:01:47,520
thing down okay hey well we're turboing to 4.9 4.95 of

22
00:01:45,360 --> 00:01:51,520
course we're not doing anything yet first order of business is to establish

23
00:01:49,280 --> 00:01:56,560
a baseline we've got all of Intel's thermal and power limits in place we're

24
00:01:53,920 --> 00:02:00,719
going to go ahead and run a blender classroom benchmark and we're going to

25
00:01:58,240 --> 00:02:06,159
monitor our CPU clock speeds to see where we end up now i suspect it won't

26
00:02:03,680 --> 00:02:10,080
be an issue for this cooler but with the bigger ones and particularly the water

27
00:02:07,840 --> 00:02:15,120
cooler we're gonna do two runs of classroom one just to

28
00:02:12,720 --> 00:02:19,440
heat up the heat sink and then a second one to actually take our values from

29
00:02:17,760 --> 00:02:24,480
we're at that uh got that 98 degrees you know good band

30
00:02:22,959 --> 00:02:29,280
let's check them out very uh popular with the kids today i'm

31
00:02:26,959 --> 00:02:33,040
always in sync with the youth that's why uh you know on friday nights i hang out

32
00:02:31,120 --> 00:02:36,040
in the back street you know with the boys

33
00:02:37,599 --> 00:02:43,840
no no you don't want to hang out with the new kids on the block or chris hanson might show up and

34
00:02:42,480 --> 00:02:48,160
give you trouble unsurprisingly we experienced

35
00:02:45,599 --> 00:02:54,000
significant thermal throttling with 100 degree temperatures and our lowest core

36
00:02:50,720 --> 00:02:57,120
managing a sustained 3.6 gigahertz i'm

37
00:02:54,000 --> 00:02:59,920
not expecting much here but let's see

38
00:02:57,120 --> 00:03:03,519
if unlocking the power limit changes anything about that situation here we go

39
00:03:02,000 --> 00:03:09,680
ladies and gentlemen we are removing all limits and going

40
00:03:06,159 --> 00:03:09,680
auto s vid behavior

41
00:03:17,360 --> 00:03:25,519
i really wasn't expecting this we're running at the same 100 degrees

42
00:03:22,000 --> 00:03:28,319
you know 97 to 99 degrees

43
00:03:25,519 --> 00:03:32,480
but curiously our clock speeds are actually higher

44
00:03:30,400 --> 00:03:37,879
how about that do you have an explanation for that that's bizarre our

45
00:03:35,280 --> 00:03:42,159
lowest core sustained clock speed is now 3.9 gigahertz in spite of the fact that

46
00:03:40,720 --> 00:03:46,879
we didn't change out our cooling solution now that's still not impressive

47
00:03:44,799 --> 00:03:52,959
with power limits fully removed we would expect this ship to maintain 4.9

48
00:03:49,680 --> 00:03:55,280
gigahertz not 3.9 gigahertz over an

49
00:03:52,959 --> 00:03:59,680
extended period of time so now then it's time for us to swap out for something

50
00:03:57,920 --> 00:04:04,159
a little bit better i don't know how it is the cooler

51
00:04:02,799 --> 00:04:08,400
i don't know how it is that cooler master has been making heat sinks this

52
00:04:06,080 --> 00:04:12,560
long and yet they have such obtuse mounting mechanisms still

53
00:04:11,120 --> 00:04:16,320
i've been saying it for years and i'll say it again everyone else

54
00:04:14,480 --> 00:04:19,680
just copy Noctua thanks for coming to my ted talk you

55
00:04:17,919 --> 00:04:24,240
know well i can't say much positive about the finishing on the fins or about

56
00:04:21,919 --> 00:04:29,520
the mounting mechanism you can't fault the hyper 212 family when it comes to

57
00:04:26,160 --> 00:04:31,680
price to performance it did spike up to

58
00:04:29,520 --> 00:04:36,960
around 100 degrees during the turbo window under stock operation but it

59
00:04:34,320 --> 00:04:42,400
managed an average of just under 69 degrees during the rest of it netting it

60
00:04:39,759 --> 00:04:46,400
a 400 megahertz improvement over the stock cooler

61
00:04:43,759 --> 00:04:53,199
now we're running our second round with the limits removed and yes we are seeing

62
00:04:49,520 --> 00:04:54,000
anywhere from 93 to 99 degrees

63
00:04:53,199 --> 00:05:00,240
but our sustained turbo speeds are 4.7

64
00:04:58,000 --> 00:05:04,639
gigahertz that's just 200 megahertz shy of what we're expecting with water

65
00:05:02,720 --> 00:05:10,560
cooling of course most people don't spend 500 bucks on a CPU to leave any of

66
00:05:08,000 --> 00:05:14,080
its performance on the table so we're gonna splurge we're gonna spend

67
00:05:12,240 --> 00:05:19,039
another twenty dollars and we're gonna move up to a Noctua nhu 12

68
00:05:17,840 --> 00:05:23,039
s mostly the differences in performance

69
00:05:21,440 --> 00:05:28,160
that i'm expecting here are going to be down to Noctua's superior fan design

70
00:05:26,080 --> 00:05:32,240
also they have a much better mounting mechanism strongly recommended and there

71
00:05:31,039 --> 00:05:36,720
it is how much cooling do you need to get the

72
00:05:34,320 --> 00:05:42,479
most out of your 10900k assuming that you are not manually overclocking the

73
00:05:38,960 --> 00:05:44,000
answer is somewhere between a hyper 212

74
00:05:42,479 --> 00:05:47,840
assuming you're using a high quality thermal paste rather than one of our

75
00:05:46,160 --> 00:05:53,039
graphite thermal pads and an nhu 12s

76
00:05:50,400 --> 00:05:56,880
with that said even with the 5 to 10 degrees of improvement that we would

77
00:05:54,479 --> 00:06:02,000
expect here going to a thermal paste i would say that our load temperatures in

78
00:05:59,039 --> 00:06:07,520
our fully unfettered mode are higher than i would be comfortable with so

79
00:06:04,479 --> 00:06:09,199
i would still consider an upgrade if you

80
00:06:07,520 --> 00:06:13,759
plan to do things like render out scenes like this in blender

81
00:06:11,440 --> 00:06:18,240
on the regular so let's see how much improvement we get here every time i go

82
00:06:16,160 --> 00:06:23,919
a while without mounting one of these i forget how utterly

83
00:06:20,360 --> 00:06:26,319
chunk-tastic the nht15 is like look at

84
00:06:23,919 --> 00:06:30,560
this the performance may actually be better than water cooling but it sure

85
00:06:28,639 --> 00:06:37,120
does clutter up your CPU socket like even just getting at your memory is a

86
00:06:33,039 --> 00:06:39,600
chore oh that's better those are under

87
00:06:37,120 --> 00:06:43,680
control thermals all right for your convenience i'm going to talk through

88
00:06:40,960 --> 00:06:47,840
the h115i so that's the 280 millimeter aio water cooler and the nhd15 air

89
00:06:46,240 --> 00:06:52,319
cooler at the same time because their performance profiles were somewhat

90
00:06:50,080 --> 00:06:58,080
similar under stock operation both of them actually didn't manage to

91
00:06:54,639 --> 00:07:00,240
outperform the nhu 12s by a significant

92
00:06:58,080 --> 00:07:05,039
margin which i was a little bit surprised to see especially that

93
00:07:02,319 --> 00:07:10,160
equivalency in the hottest core while the system was turboing but when

94
00:07:07,759 --> 00:07:14,880
it came time to unlock the power limits with multi-core enhancement

95
00:07:12,000 --> 00:07:19,199
both these solutions clearly separated themselves from everything that we've

96
00:07:16,800 --> 00:07:24,000
seen so far not just managing 4.9 gigahertz across the board but doing it

97
00:07:22,080 --> 00:07:28,639
with core temperatures that i would consider to be maybe not

98
00:07:26,319 --> 00:07:34,160
the most comfortable we are still up in the 80 high 90 low degree range but once

99
00:07:32,479 --> 00:07:38,479
you sub out our thermal pad for a high quality thermal grease

100
00:07:35,759 --> 00:07:43,039
definitely acceptable and around that 80 degrees that we talked about in our

101
00:07:40,560 --> 00:07:47,520
original review what all of that means is that if i was looking for additional

102
00:07:45,199 --> 00:07:51,680
overclocking headroom i would absolutely recommend spending the hundred plus

103
00:07:49,440 --> 00:07:55,199
dollars on a thermal solution like this but if all i wanted was to get the most

104
00:07:53,440 --> 00:07:57,440
out of my CPU and i knew that i wasn't going to be hitting it this hard all the

105
00:07:56,639 --> 00:08:02,960
time around 50 to 60 dollars for a good

106
00:08:00,639 --> 00:08:09,039
single tower aftermarket air cooler is going to do the trick just nicely

107
00:08:06,400 --> 00:08:14,560
just like our sponsor does the trick just nicely micro center is open for all

108
00:08:12,400 --> 00:08:19,440
of your work from home or learn from home technology needs and they want you

109
00:08:16,319 --> 00:08:21,680
to check out the lienlee landcool 205

110
00:08:19,440 --> 00:08:26,400
mid tower tempered glass case it supports atx m80x and mini itx

111
00:08:24,560 --> 00:08:30,000
motherboards with extra space behind the motherboard tray for cable management

112
00:08:28,080 --> 00:08:34,399
it's got side ventilation for improved airflow with easy access to the power

113
00:08:32,000 --> 00:08:38,240
supply and hard drive bay from the side it's got rear removable SSD mounting

114
00:08:36,399 --> 00:08:42,800
brackets and high performance magnetic dust filters at the top and front with a

115
00:08:40,479 --> 00:08:46,720
dust filter tray at the bottom you can check out it and other micro center

116
00:08:44,720 --> 00:08:50,240
specials at the link in the video description if you guys enjoyed this

117
00:08:48,720 --> 00:08:54,800
video and you like crazy cooling experiments maybe check out our water

118
00:08:52,560 --> 00:08:59,279
cooling the macbook air 2020 video we use an ice bath to explore whether it's

119
00:08:56,880 --> 00:09:04,040
actually a thermal limit or maybe a power limit holding back its performance