1
00:00:00,080 --> 00:00:06,160
Recently, with the help of one of the world's top overclockers, we built the

2
00:00:03,520 --> 00:00:10,559
world's fastest gaming PC, cracking the top 20 on the 3D Mark leaderboards. And

3
00:00:08,639 --> 00:00:15,679
that was all without doing anything particularly horrible to our RTX 5090,

4
00:00:13,440 --> 00:00:20,960
which got us thinking, what if we did do horrible things? There it goes. See? Oh

5
00:00:18,480 --> 00:00:24,800
my god. Got to say, boys, I was expecting the

6
00:00:22,720 --> 00:00:28,960
wicked fast test bench. I was expecting the chunky water block. And I was

7
00:00:26,720 --> 00:00:34,239
expecting the industrial chiller. What I wasn't expecting was for you to put

8
00:00:30,640 --> 00:00:39,360
Blasty Dip all over this ASUS ROG Astral

9
00:00:34,239 --> 00:00:40,960
GPU. Why? Sorry. Go fast.

10
00:00:39,360 --> 00:00:44,879
Well, that's the thing, isn't it? If we want to go fast, we got to go cold. And

11
00:00:42,800 --> 00:00:49,520
if we want to go cold, there is a very real risk of condensation forming on our

12
00:00:47,440 --> 00:00:53,840
components. Now, we could insulate them, but realistically, we probably don't

13
00:00:51,680 --> 00:00:58,320
need to because inside our environmental chamber, we can control not only

14
00:00:55,920 --> 00:01:02,160
temperatures, but also humidity. meaning that we probably won't even need our

15
00:01:00,399 --> 00:01:06,080
makeshift conformal coding. It just looks cool, just like I think it's cool

16
00:01:03,920 --> 00:01:10,479
to tell you about our sponsor, Hner. Hster is based in Germany and has your

17
00:01:08,159 --> 00:01:15,280
data locked down by GDPR and strict privacy laws, even at their locations in

18
00:01:12,880 --> 00:01:20,479
the US and Singapore. Get $20 off a Hexner cloud server today using code

19
00:01:17,600 --> 00:01:24,320
LT25 at our link in the video description. If you don't get why NVIDIA

20
00:01:22,560 --> 00:01:27,520
would care about what we're doing, I don't blame you. I mean, it's our GPU,

21
00:01:26,240 --> 00:01:32,960
and I'm sure that they would agree that we can paint it as we see fit. But what

22
00:01:30,240 --> 00:01:37,520
NVIDIA doesn't like is when users and board partners play around with the

23
00:01:34,960 --> 00:01:44,000
power targets for their cards. And this particular one is running a BIOS that

24
00:01:39,600 --> 00:01:45,759
allows us to run it at over a,000 watts.

25
00:01:44,000 --> 00:01:52,479
That is about double of what the excellent 12vt

26
00:01:49,360 --> 00:01:55,439
2x6 power connector here is rated for.

27
00:01:52,479 --> 00:02:00,880
Not only that, but even with this special hybrid liquid metal compound

28
00:01:57,680 --> 00:02:02,960
that SPL provided for us, running 1,000

29
00:02:00,880 --> 00:02:07,840
watts through this GPU could have a serious impact on its reliability and

30
00:02:05,759 --> 00:02:11,280
being aliveness, which NVIDIA doesn't like. This is cool. I was cleaning off

31
00:02:09,360 --> 00:02:14,560
the hybrid compound for a B-roll shot, and you can actually see the little

32
00:02:13,200 --> 00:02:18,319
balls of metal in it. It's where my nickname comes from.

33
00:02:16,720 --> 00:02:21,760
We've already done a baseline run with our card on air cooling. So, our first

34
00:02:20,080 --> 00:02:28,720
step is to install this Bits Power Nebula GPU.

35
00:02:24,560 --> 00:02:30,720
Wow. Damn, that's beautiful. I don't

36
00:02:28,720 --> 00:02:35,519
know about you guys, but uh I'm going to need a cold shower after after seeing

37
00:02:33,200 --> 00:02:39,519
these machining marks. I'm going to install this on there and then we're

38
00:02:37,840 --> 00:02:42,879
going to do a non-chilled water cooling run to see if we can push things any

39
00:02:41,280 --> 00:02:46,080
further than we already have. What's this block go for? It's pre-production,

40
00:02:44,400 --> 00:02:51,200
so we don't really know yet. Probably around 250. enough that if you can

41
00:02:48,000 --> 00:02:52,800
afford a 5090, you don't care anyway.

42
00:02:51,200 --> 00:02:56,560
All right, let's see how it goes. Now, quick question for you. For your hybrid

43
00:02:54,879 --> 00:03:00,640
liquid metal thermal compound to work properly, do we need to run this at an

44
00:02:59,120 --> 00:03:04,239
above ambient temperature for a little bit? Let that liquid metal melt and

45
00:03:02,720 --> 00:03:09,680
spread out. I would say it couldn't hurt anything. Okay, it's not

46
00:03:08,159 --> 00:03:13,040
Don't show that. No, no, no. Don't show that. NVIDIA really never is going to

47
00:03:11,360 --> 00:03:16,480
send us a GPU ever again. Do you want to take that GPU out? Why? Well, we're

48
00:03:15,280 --> 00:03:21,120
going to take it down anyway, and you're going to spill water on it. There's a stop fitting here, right? So, should be

49
00:03:19,840 --> 00:03:24,879
fine. Did you stop the other stop fitting? I did. Cool. Man, Bits Power

50
00:03:23,280 --> 00:03:28,800
makes some gorgeous stuff. I still remember their OG true gold and true

51
00:03:27,280 --> 00:03:32,000
silver fittings back in the day. The silver ones were brilliant, actually,

52
00:03:30,319 --> 00:03:35,680
because they serve a dual purpose. They look cool and silver acts as a natural

53
00:03:34,400 --> 00:03:41,280
bioside, so you don't have to worry about anything growing in your loop. Wow, these have a lot of restriction in

54
00:03:39,200 --> 00:03:46,799
them. Can't say I love these fittings. Spits power. Also, the threading is not

55
00:03:44,239 --> 00:03:50,080
good. These cards sag enough when you do have them screwed in at the back. This

56
00:03:48,480 --> 00:03:54,319
one's on a whole other level. Like, I cannot I cannot let go of this. It will

57
00:03:52,319 --> 00:03:57,360
probably break the socket off. Oh, wow. It's got all the sizes and everything.

58
00:03:55,920 --> 00:04:01,360
Look at that. Comes with your own calipers. Perfect.

59
00:03:59,680 --> 00:04:07,200
Alex was saying I went too short on this tube, which could be perceived as dumb.

60
00:04:04,319 --> 00:04:11,360
Like a fox. Now, it's a safety net for our our poor GPU in case the SEG bracket

61
00:04:09,519 --> 00:04:15,920
gets bumped. Now we theoretically open up the valves and everything is bueno.

62
00:04:14,319 --> 00:04:20,560
What could go wrong? We couldn't figure out why the GPU is at 0 degrees. It's

63
00:04:17,919 --> 00:04:24,560
because of zero zero deg zero frequency hertz. Yeah. Frequency. That unit of

64
00:04:22,560 --> 00:04:27,280
measurement that everyone uses and such. All right. This does not look any faster

65
00:04:26,080 --> 00:04:32,880
than yesterday. In fact, it looks slower. Yes. You didn't overclock anything, did you? No. I knew we brought

66
00:04:30,639 --> 00:04:37,680
you here for a reason. I want to make sure it's working first. Is that 58°

67
00:04:36,080 --> 00:04:41,040
though? That doesn't seem right. Before we push things any further, Alex wanted

68
00:04:39,280 --> 00:04:46,800
to compare this score with what we got on our stock 5090, which is

69
00:04:44,720 --> 00:04:52,800
actually I don't think this is the right BIOS. But what do you mean by not the

70
00:04:49,280 --> 00:04:55,360
right Oh, we hit 970 W. Okay, I'm giving

71
00:04:52,800 --> 00:05:00,000
myself some credit here. Remember when I was like, "Hey, those temps seem high."

72
00:04:57,360 --> 00:05:02,960
I remember that. I I remember. Do you want to go back to the normal BIOS? I

73
00:05:01,759 --> 00:05:07,039
feel like now that we've skipped the regular BIOS, we've skipped it. So, we

74
00:05:04,800 --> 00:05:10,800
might as well just go for it. plow forward. Yeah. So, now we could probably

75
00:05:08,960 --> 00:05:14,400
make some adjustments to clock speed. Yeah, absolutely. The idea today is to

76
00:05:12,800 --> 00:05:17,759
crack top 10. We haven't even pulled out the chiller yet, and we're already well,

77
00:05:16,560 --> 00:05:22,000
we're not that close, but we'll get there. Just to be clear, none of what

78
00:05:19,840 --> 00:05:25,520
we're doing is necessary or probably even going to result in a difference

79
00:05:23,600 --> 00:05:30,639
that you would feel dayto-day. This is pretty much as fast as hardware gets as

80
00:05:27,919 --> 00:05:34,720
it is. We have a Ryzen 9 9950X 3D processor, which is as fast as it gets

81
00:05:32,479 --> 00:05:39,520
with eight of the cores having 3D vcash and the other eight cores to pick up, I

82
00:05:36,960 --> 00:05:46,320
guess, any slack. We have 32 gigs of DDR5 6000 CL26 and that's tuned to run

83
00:05:43,680 --> 00:05:52,080
at about 6200 MHz. And of course, we have an Astral RTX 5090, which according

84
00:05:49,120 --> 00:05:56,240
to Splave is the best one. Yeah. Most importantly though, it's in a

85
00:05:54,240 --> 00:06:01,360
Lamborghini case that'll make it go way faster. Alex, you see this carbon fiber?

86
00:05:58,720 --> 00:06:04,639
That's plus like 25 FPS for sure. I'm willing to bet it doesn't even affect

87
00:06:02,560 --> 00:06:08,160
the like weight of it that much. Oh, it's a really heavy game. God, it's so

88
00:06:06,160 --> 00:06:12,560
heavy. I bet you could lift it easily. Maybe.

89
00:06:10,960 --> 00:06:17,520
Sorry. Now that I know this thing's running a 1000 watts, I'm really impressed with this water block rather

90
00:06:15,280 --> 00:06:21,440
than disappointed. That's a lot of heat. How much radiator

91
00:06:19,199 --> 00:06:24,720
you got on this? Just a triple. Yep. That's a good thing we're going to be

92
00:06:22,400 --> 00:06:30,000
replacing that. Bits power did not give a thermal pad. I was just checking that.

93
00:06:27,919 --> 00:06:33,600
No, they didn't. I would strongly recommend Bits Power if you're watching

94
00:06:31,759 --> 00:06:36,720
this or if you buy one of these and Bits Power didn't watch this. Throw a little

95
00:06:35,120 --> 00:06:39,759
thermal pad right in the back by where the GPU power connector is. It might

96
00:06:38,400 --> 00:06:45,280
help dissipate some heat if you ever have a loose connection. Now, that's totally cheating. You just set the

97
00:06:43,039 --> 00:06:48,319
anisotropic filtering to crap mode. High performance, baby. They used to call

98
00:06:46,800 --> 00:06:51,440
that cheating. Doesn't have to look good. Just has to be fast. Well, in

99
00:06:50,240 --> 00:06:56,560
fairness, the reason they used to call it cheating was because ATI and NVIDIA

100
00:06:54,560 --> 00:07:00,880
would change it around in their defaults in the driver settings. That's true.

101
00:06:58,560 --> 00:07:05,599
It's fine if I decide I want my image quality to be a little worse. It's not

102
00:07:02,800 --> 00:07:10,160
fine if you decide it should be worse. This looks pretty good. Ah, we cracked

103
00:07:08,080 --> 00:07:13,919
it like 210 there. Oh, yeah. What did we even do? Uh, we just bumped the clock

104
00:07:12,080 --> 00:07:17,120
speed up a little bit. Nice. This legit looks like a really good run. I don't

105
00:07:15,199 --> 00:07:23,520
know about top 10, but good run. We just have to beat Jake. Oh. Easy. Whoa,

106
00:07:21,840 --> 00:07:29,120
buddy. What was JC at? Yeah. 4325. I think

107
00:07:26,319 --> 00:07:34,240
we're 43129. All right, we can do better. Oh, we're not there. 43250.

108
00:07:32,000 --> 00:07:38,080
Let's chill it. Chill it. Yeah, let's not just break him. Let's crush him. Uh,

109
00:07:36,720 --> 00:07:42,800
do you want to do a quick explainer for those who didn't watch the previous video? You know, what all these dials

110
00:07:41,199 --> 00:07:46,479
and knobs mean and what order you should adjust them in. In the meantime, I'm

111
00:07:44,479 --> 00:07:51,120
going to go plug in the chiller. So, the first thing you do clearly is set the

112
00:07:48,080 --> 00:07:53,120
fan speed to 100%. Doesn't matter on our

113
00:07:51,120 --> 00:07:56,800
card, but on yours it probably does. After that, I see you have the power

114
00:07:54,400 --> 00:08:00,400
target only at 95 and have done nothing to the GPU voltage. Would you want to

115
00:07:58,879 --> 00:08:06,319
play with those at all? At home, probably. If you don't have the 1000

116
00:08:02,639 --> 00:08:08,080
watt BIOS, um you're you're going to run

117
00:08:06,319 --> 00:08:13,599
into thermal problems before you hit 100%. At home, you would maybe even use

118
00:08:10,879 --> 00:08:17,919
I think it goes up to like 104 or 105 on a normal card. You just max that out.

119
00:08:16,400 --> 00:08:24,639
Then you could play with the voltage a little bit. Uh and then just I just do

120
00:08:20,879 --> 00:08:27,120
25 megahertz at a time uh on the clock

121
00:08:24,639 --> 00:08:33,279
and run it and if it passes go higher. Easy enough. Easy. The voltage is tricky

122
00:08:29,599 --> 00:08:35,919
on 5090. It gives you a percentage. What

123
00:08:33,279 --> 00:08:40,880
does that mean? I just put it at 100%. Okay. All the volts. 100% voltage. Sure.

124
00:08:39,039 --> 00:08:44,800
And you can see that doesn't really affect much. It caps itself out. They

125
00:08:43,360 --> 00:08:49,920
don't really want you to increase voltage on the card when power limit is

126
00:08:48,160 --> 00:08:54,560
kind of the limiting factor. Makes sense. So we're kind of controlling

127
00:08:51,519 --> 00:08:59,519
power only through power target. So you

128
00:08:54,560 --> 00:09:02,000
can use VF curve to to lower voltage

129
00:08:59,519 --> 00:09:06,959
but the performance is worse. I just avoid messing with voltages and just

130
00:09:04,800 --> 00:09:12,000
play with power target and your core clock. So if it becomes unstable when

131
00:09:09,279 --> 00:09:17,279
you go a tiny bit higher in the boost, you increase the power until you max it

132
00:09:13,760 --> 00:09:20,640
out. Yeah. So say I have increased the

133
00:09:17,279 --> 00:09:25,200
GPU boost clock by 200 MHz and this

134
00:09:20,640 --> 00:09:29,200
failed. So you could try like 95% power

135
00:09:25,200 --> 00:09:31,600
target and see if you pass.

136
00:09:29,200 --> 00:09:36,080
And if you pass, that's about all you get. Would you ever

137
00:09:33,920 --> 00:09:40,000
turn down the power target and try again and see if it was a thermal issue? Uh,

138
00:09:38,160 --> 00:09:44,560
you could turn down power target, but then it's just going to take away your

139
00:09:41,839 --> 00:09:49,200
clock cuz you'll hit the power target cuz Jake says he backed off voltage and

140
00:09:47,120 --> 00:09:53,519
that was how he beat us yesterday. We run the air conditioner

141
00:09:51,440 --> 00:09:57,680
so we would by the end of the run it was the card was really warm whereas here

142
00:09:56,240 --> 00:10:02,080
we're taking the heat away and it's not really going up after we've started the

143
00:09:59,360 --> 00:10:06,000
bench. I see. So, by the end of the bench, we were failing maybe in the last

144
00:10:03,600 --> 00:10:11,680
2 seconds. So, we just really need like 3° to be able to pass. Okay. I got a lot

145
00:10:10,000 --> 00:10:16,000
more than 3° for you. Our chiller's ready. All right. Let's do it. All right. Let's do it. Did you have kind of

146
00:10:14,240 --> 00:10:20,399
a a thought on exactly how we're plumbing this up? I was going to close

147
00:10:17,279 --> 00:10:23,120
these and then just remove a couple of

148
00:10:20,399 --> 00:10:26,079
PCIe brackets, throw them at the back. Oh, okay. Do they just go to the back or

149
00:10:24,880 --> 00:10:30,320
they could just go here? They just go right there, too.

150
00:10:28,959 --> 00:10:35,519
I don't think we need the reservoir on the loop either. I was just going to plum around it. I was basically just

151
00:10:33,200 --> 00:10:38,480
going to go to like here. Oh yeah, we can just totally delete the whole

152
00:10:37,200 --> 00:10:42,640
radiator if you want. What do you need a radiator for? Radiators are for chumps.

153
00:10:40,959 --> 00:10:47,680
Yeah, we've got a much bigger one. Hey, there's a display for coolant temperature in there. Is that what that

154
00:10:45,040 --> 00:10:50,800
is? Yeah. Oh, sick. Does it go sub-zero? We don't know yet. Oh, neat. Okay, we'll

155
00:10:49,200 --> 00:10:54,320
find out. Cool. The last thing that I want to do is drain this reservoir

156
00:10:52,959 --> 00:10:58,640
because if we go and we put this in a sub-zero chamber and it freezes, it will

157
00:10:56,640 --> 00:11:02,560
break. It's about to get real, Alex. Oh, Jesus.

158
00:11:01,040 --> 00:11:06,880
There it goes. See, that's the fastest way to drain a thing. This is really

159
00:11:05,120 --> 00:11:11,360
nice how Le and Lee did this. You just push on that. Full thing comes out,

160
00:11:09,839 --> 00:11:15,200
buddy. That's awesome. It's full of water. Don't forget. Yeah, I know. Cool. Okay.

161
00:11:14,160 --> 00:11:19,600
I'm just making sure, you know, before you like, you know, spill water. I'd hate for any of that to happen on this

162
00:11:18,399 --> 00:11:23,360
shoot. It's been a really tidy, clean shoot with my spilled water so far.

163
00:11:21,760 --> 00:11:28,000
There it goes. Professionals. All right, let's fire up this chiller. All right.

164
00:11:24,959 --> 00:11:30,720
All right. Let's go.

165
00:11:28,000 --> 00:11:33,920
Woo! She stopped pumping. I'm just going to give her the old power cycle here.

166
00:11:32,079 --> 00:11:38,480
Also, that poor pump sounds like it is dying. She moved a little bit of

167
00:11:35,600 --> 00:11:42,160
coolant. And then she stopped. That seems to be full of coolant. Yeah, we

168
00:11:40,240 --> 00:11:45,519
must try again. Okay, we're just trying again. Not going

169
00:11:43,839 --> 00:11:49,519
to lie, little scared. That sounds pretty bad. We might need to prime the

170
00:11:46,959 --> 00:11:53,920
pump, Alex. We shouldn't have to. I mean, I know we shouldn't have to, but

171
00:11:52,000 --> 00:12:03,320
we might want to do it. That is super weird. Why? What are you adjusting?

172
00:11:58,880 --> 00:12:03,320
Oh, this is the pressure. Oh, Jesus.

173
00:12:04,720 --> 00:12:08,200
I got shelled.

174
00:12:08,240 --> 00:12:17,360
I was going to say we should put some hose clamps on those and then I totally forgot. I brought them over. Oh my god,

175
00:12:13,279 --> 00:12:19,120
it's a mess. Is deer toxic? Um, it's all

176
00:12:17,360 --> 00:12:23,839
over me. It's probably not great for my pants, that's for sure.

177
00:12:22,399 --> 00:12:28,880
We're supposed to cool the computer, Alex, not the people. Look at the

178
00:12:26,000 --> 00:12:32,720
screen. Oh no. And this video had so many fun moments

179
00:12:31,120 --> 00:12:36,240
that we couldn't fit them all in this video. So the extras are all on float

180
00:12:34,639 --> 00:12:42,160
plane. Go check it out. Oh, he doesn't see me. He doesn't see me. He doesn't seem No, no, no, no. Come on. First you

181
00:12:40,160 --> 00:12:45,680
call me small balls. Then you say you don't believe in my ability to install a

182
00:12:43,839 --> 00:12:52,160
water block. Alex, were we reading it wrong? Yeah, we're just really dumb. Oh,

183
00:12:48,560 --> 00:12:53,600
this is peak cinema. Fortunately, the

184
00:12:52,160 --> 00:12:58,160
glass panel in the front of this computer took the brunt of it. I don't

185
00:12:55,360 --> 00:13:02,160
think we hit the GPU or motherboard. One drop right there. We hit our screen in

186
00:13:00,320 --> 00:13:08,320
the back, but I think she's pretty clean. Okay, so she was clearly moving

187
00:13:04,880 --> 00:13:11,519
it. Yeah. So, that was awesome. This was

188
00:13:08,320 --> 00:13:13,200
air log. Yeah. Is there a valve shut off

189
00:13:11,519 --> 00:13:18,240
somewhere else? because it seemed like it was pushing against.

190
00:13:16,399 --> 00:13:25,519
Yeah, because that's the out. Oh man. Dude, I am super confused. Oh my god, I

191
00:13:22,240 --> 00:13:28,880
figured it out. Oh, we're idiots. The

192
00:13:25,519 --> 00:13:31,040
shut off. We had the valve closed on the

193
00:13:28,880 --> 00:13:34,720
CPU block. We're lucky we didn't blow up that GPU block. Yeah, it would have been

194
00:13:32,720 --> 00:13:37,839
fine. We They're good for over 200 PSI. That's true. We do. Well, not that one

195
00:13:36,160 --> 00:13:41,760
specifically, but pump only does like 100 PSI. Our Canadian healthcare doesn't

196
00:13:40,000 --> 00:13:49,600
cover your chemical brand. I'm sorry. Yeah, I probably have better coverage here than I do at home. Okay, let's do

197
00:13:45,760 --> 00:13:54,720
this thing. You ready?

198
00:13:49,600 --> 00:13:56,320
I guess it's Oh my god. Oh my god.

199
00:13:54,720 --> 00:14:00,639
All right, we did it. Is there overclocking always this as exciting?

200
00:13:57,920 --> 00:14:04,800
Not as explosive as this usually. What do we think is safe to set the chiller

201
00:14:02,000 --> 00:14:09,360
at? Right now it's at 23. Yeah, lower. I bet we could do like probably 12 or so.

202
00:14:07,920 --> 00:14:14,240
Yeah, I was going to say 12. All right, give her a sec. She's chilling. She's just chilling, you know, jumping right

203
00:14:12,240 --> 00:14:18,000
into a benchmark. Already, I think we're what, 5 degrees cooler than we were when

204
00:14:16,160 --> 00:14:21,519
we were relying on ambient air to cool our loop down. Exactly. And did you

205
00:14:20,160 --> 00:14:26,000
change any settings or anything or we just hoping it's going to more clocks?

206
00:14:24,000 --> 00:14:31,760
Okay. We're still on the 1000 watt BIOS. Hey. Yeah. And we're at 43°. Yeah. Wow.

207
00:14:29,920 --> 00:14:35,600
So, it was really heating up that water before. Well, I think the bigger thing

208
00:14:33,279 --> 00:14:41,600
now is that the pump just is ripping compared to the 85.

209
00:14:37,839 --> 00:14:43,519
43810. There we go. All right. Wa,

210
00:14:41,600 --> 00:14:49,600
buddy. Okay, let's submit that, buddy. LT right here. Let's go. Number 17. Take

211
00:14:46,800 --> 00:14:54,000
that, Jake. Honestly, though, um, Alex, she's struggling to pull down. Yeah.

212
00:14:51,440 --> 00:14:59,279
What's on the go? Well, 1,000 watts of GPU. Oh, that that is a thing. Like,

213
00:14:56,560 --> 00:15:03,600
does this not have 1,000 W of capacity? It has 5,000 watts. Oh, well, on the one

214
00:15:02,079 --> 00:15:07,199
hand, we're pretty upset about our industrial chiller not working. We think

215
00:15:05,440 --> 00:15:12,560
maybe there's been a gas leak of some sort cuz she's pulling down. The

216
00:15:09,519 --> 00:15:15,920
compressor is running, but the condenser

217
00:15:12,560 --> 00:15:17,199
is not hot and it's not pulling down to

218
00:15:15,920 --> 00:15:24,320
the temperature quickly and it's actually going up while we run our benchmark. The good news is that we just

219
00:15:20,880 --> 00:15:27,519
hit 44,000. So, is that still 17th

220
00:15:24,320 --> 00:15:29,680
place? 16th. Oh, wait. We We took your

221
00:15:27,519 --> 00:15:33,199
own spot. Ah, normally I'd be thinking, okay,

222
00:15:31,680 --> 00:15:37,040
we'll just get an HVAC technician to come out and fix our system. But the

223
00:15:35,440 --> 00:15:41,839
problem is that Allan's got a flight to catch. So, if we want to go very, very

224
00:15:40,000 --> 00:15:48,880
subzero, we're going to have to come up with some ideas. Hey, look. I found an

225
00:15:44,320 --> 00:15:51,040
idea. Oh, yeah. Now, that's a radiator.

226
00:15:48,880 --> 00:15:56,959
We throw this with a bunch of high-speed fans inside the environmental chamber,

227
00:15:54,320 --> 00:16:01,199
and we can chill the loop down that way. Ooh. Ooh, that's pretty fast. spicy.

228
00:15:59,519 --> 00:16:05,519
Even though the chiller isn't chilling quite as well as we hoped, the pump in

229
00:16:02,959 --> 00:16:10,000
it is still fantastic, and our scores just keep on going up. I'm trying to

230
00:16:07,440 --> 00:16:15,120
find 3,000 RPM fans, but they're all mixed in together with the 2,000 RPM

231
00:16:12,240 --> 00:16:21,199
fans. Hello.

232
00:16:17,680 --> 00:16:23,680
Huh. Yeah, that'll do nicely. Here's

233
00:16:21,199 --> 00:16:32,639
what I'm kind of vibing, huh? All right, I got a plan.

234
00:16:32,639 --> 00:16:39,440
What are you doing? I am making a shroud. Why? Because if we just put the

235
00:16:38,000 --> 00:16:43,600
fan against the radiator, we're going to miss out on probably a solid 60% of the

236
00:16:42,079 --> 00:16:49,040
actual cooling capacity of that radiator. Dude, this is going to be

237
00:16:45,440 --> 00:16:51,440
awesome. Heck yeah. Oh, dude. See this?

238
00:16:49,040 --> 00:16:58,000
This is what I'm talking about. Okay, so we get the fan off of the fins. Boom.

239
00:16:56,000 --> 00:17:01,920
We're making use of all that airflow. Boom. Oh, I wish I'd centered it. Well,

240
00:17:00,160 --> 00:17:07,439
it's done now. So, this is going to be sketchy.

241
00:17:04,000 --> 00:17:09,439
Shield your virgin eyes.

242
00:17:07,439 --> 00:17:15,600
Gorgeous. All right. So, now theoretically

243
00:17:13,520 --> 00:17:22,000
all the airflow will go through the radiator. We have a

244
00:17:19,120 --> 00:17:27,919
cooling solution. Side quest complete. You have obtained this freaking thing.

245
00:17:24,959 --> 00:17:34,880
Okay, here we go, boys. Oh my god. What the hell is that? That Oh my god, my

246
00:17:31,440 --> 00:17:37,360
friend is a radiator and a fan. Whoa,

247
00:17:34,880 --> 00:17:40,799
buddy. Oh, that's fast. We're working. We're working over here. I was working.

248
00:17:38,960 --> 00:17:44,799
What are you implying? I made this. That's beautiful. Oh, wait. Did I ever

249
00:17:42,480 --> 00:17:49,440
get a pump? Okay. Well, I found a reservoir. So, we have that. Well, what

250
00:17:47,760 --> 00:17:55,760
I was thinking was we still use the chiller even though it's not chilling too good. That pump in it is insane.

251
00:17:53,039 --> 00:18:00,000
Okay. So, we put the chiller back. Are we going to lose a lot of our cooling,

252
00:17:57,200 --> 00:18:04,400
though? Maybe use the res pump thing we were using in here or No.

253
00:18:02,640 --> 00:18:07,280
Yeah, we could just use that one. We could do that. Yeah. So, wait, wait,

254
00:18:06,080 --> 00:18:13,200
wait, wait, wait. Where does this put us? Oh, dude. Let's go. Let's go. We're

255
00:18:10,400 --> 00:18:17,600
top 10, baby. Let's go. Not even cold yet, right? In terms of raw actual

256
00:18:15,840 --> 00:18:21,600
numbers, we're not even that far off some of these scores. Yeah. Do you think

257
00:18:19,760 --> 00:18:26,440
she's got more left to give? I think she does. All right. I'm excited.

258
00:18:27,679 --> 00:18:35,039
She's still spooling. Wow. That is So, I think she's going to cool your uh your

259
00:18:32,880 --> 00:18:39,360
GPU there. Uh I think we're good. So, we're just going to use this and we're

260
00:18:36,720 --> 00:18:46,000
just going to set it up in there. Woo! She's chilly, boys. She's chilly. So,

261
00:18:43,520 --> 00:18:49,760
it's blows out here. I feel like we could do that.

262
00:18:48,080 --> 00:18:53,520
Okay, I'm going to run the power out to Alex.

263
00:18:51,360 --> 00:18:58,880
All right. How we feeling about this? Uh, mid. We call that a Canadian W.

264
00:18:57,120 --> 00:19:03,200
Feeling mid. Okay. Here we go. Wait, should we take the tempered glass off?

265
00:19:00,720 --> 00:19:09,760
No, I just throw it in. Okay. Oh, yeah, bud. No problem, bud. Okay, so we got

266
00:19:06,080 --> 00:19:11,440
the radiator here. Love the O1 case.

267
00:19:09,760 --> 00:19:16,640
It's designed so you can reach right through it when you want to get to your

268
00:19:13,360 --> 00:19:18,480
janky subzero water cooling radiator in

269
00:19:16,640 --> 00:19:22,559
your environmental chamber like that. See, bud? She goes into the radiator, up

270
00:19:20,880 --> 00:19:30,080
the top of the radiator, down to the GPU. These are stoppered. Uh, then we

271
00:19:26,000 --> 00:19:31,360
got this here. Uh, there's some tubing.

272
00:19:30,080 --> 00:19:36,160
And then if you want to give me like a cable tie or something, then we can come

273
00:19:33,679 --> 00:19:39,760
off it here. Never grow up. You just get better toys. Let's get some windshield

274
00:19:38,160 --> 00:19:43,280
deicers. We don't have a funnel by any chance, do we? Man, I really hope that

275
00:19:41,840 --> 00:19:46,720
this doesn't le That's good. We're going to need another bag, I think. All right,

276
00:19:44,960 --> 00:19:50,880
Alex, you want to power cycle it? Dude, the coolant said 16 and that's with the

277
00:19:49,039 --> 00:19:56,240
door open. Reese, are you hitting the fan? Yeah, go for it. And this is how

278
00:19:53,440 --> 00:19:59,760
you set records, I guess. Well, depends. We haven't gotten there yet. Hey, we got

279
00:19:58,000 --> 00:20:04,880
a great view of all the displays and everything. So, our coolant is at 8° and

280
00:20:02,559 --> 00:20:09,120
our GPU is at 12. That actually checks out. That sounds about right that at

281
00:20:06,880 --> 00:20:13,919
idle, our GPU would be a few degrees above our coolant temperature. We're

282
00:20:11,919 --> 00:20:18,240
gonna see that delta grow once we're actually running benchmarks and stuff.

283
00:20:15,440 --> 00:20:23,440
But, uh, this is a very good starting point. You know which delta I'm actually

284
00:20:20,240 --> 00:20:26,320
most impressed by how close our coolant

285
00:20:23,440 --> 00:20:31,120
is to the chamber temperature. Yeah, it is really close. 2° off. That's how much

286
00:20:29,440 --> 00:20:35,440
air we're moving through that radiator. Fear it. How's them temps? You like them

287
00:20:33,120 --> 00:20:41,919
temps? Temps. They seem okay. We're at minus.5 and our loop's at 4°. GPU

288
00:20:39,679 --> 00:20:46,320
sitting chilly at six. This worked surprisingly well given that our chiller

289
00:20:44,480 --> 00:20:51,600
stopped chilling. I can't believe we pulled this solution out of our butts.

290
00:20:48,880 --> 00:20:57,120
Okay, so we're at 30C on our GPU right now. She's running pretty smooth right

291
00:20:53,200 --> 00:21:00,559
now. Eh, 45 349.

292
00:20:57,120 --> 00:21:02,960
We're in seventh. Seventh. That is a

293
00:21:00,559 --> 00:21:07,760
whole gigahertz over the base of like a founders edition. And it didn't even

294
00:21:04,720 --> 00:21:10,080
take very much work, did it? Just for

295
00:21:07,760 --> 00:21:13,679
me. Whoa. This is going to be a pretty nice

296
00:21:11,919 --> 00:21:17,840
run, I think. So, yeah, we haven't even failed yet. Really? Oh. Oh, shouldn't

297
00:21:16,480 --> 00:21:22,880
have said anything. Why did you say that? Did you see that? Why did you say

298
00:21:20,000 --> 00:21:27,440
that? We've been finding that the 50 series does this really weird thing

299
00:21:24,640 --> 00:21:32,080
where when the load drops, like you're not hitting it 100%. The frequency just

300
00:21:30,159 --> 00:21:37,280
goes to the moon for some reason and then it crashes. 45 354. That is a touch

301
00:21:35,360 --> 00:21:41,840
higher, isn't it? Five points. Yeah. Okay. Hey, look, I'll take it. So, you

302
00:21:40,000 --> 00:21:46,960
don't think that giving the CPU anymore will get us any more score? Probably

303
00:21:44,799 --> 00:21:54,960
not. Um, Intel actually scores a lot higher on

304
00:21:50,159 --> 00:21:56,880
this 1400 series for some reason. Oh,

305
00:21:54,960 --> 00:22:00,280
did we? So, we if we would have used one, it would probably

306
00:22:00,720 --> 00:22:06,960
Sorry. This is what we get for building a computer that's actually for gaming

307
00:22:05,280 --> 00:22:14,799
instead of for getting higher benchmark scores. Exactly. Oh,

308
00:22:10,559 --> 00:22:16,480
that is fast. Oh, six. We're going for

309
00:22:14,799 --> 00:22:21,679
number six. Let's go, baby. Let's go. Top five. Actually, we're the top AMD

310
00:22:18,720 --> 00:22:27,760
result now. Top AMD. Top Canadian. Let's go. All right. How much is this drawing?

311
00:22:24,080 --> 00:22:30,799
Oh, it's it's not too crazy. 900 watts

312
00:22:27,760 --> 00:22:32,559
at the moment. Okay. 1,028. 1,000 and a

313
00:22:30,799 --> 00:22:38,080
bit. That's about what we expect. That's what we were doing before. Here we go. I

314
00:22:35,679 --> 00:22:42,320
believe. Come on. We're just shy of minus 20 on the chamber. We don't want

315
00:22:40,400 --> 00:22:47,520
to go too much lower than that. Once you start hitting minus30, um things don't

316
00:22:45,679 --> 00:22:52,320
likey. That's what like the juice in the capacitors of your power supply just

317
00:22:49,200 --> 00:22:53,919
starts turning into solid not juice.

318
00:22:52,320 --> 00:22:56,919
She's artifacting. Nah, she's artifacting.

319
00:22:57,360 --> 00:23:03,760
That's a fun one. Oh wow, we're starting to get some really weird reported

320
00:23:01,760 --> 00:23:07,919
numbers there. Well, our RAM's on the sun, so that's probably going to be a

321
00:23:06,320 --> 00:23:14,320
problem. You ever feel like if you don't move, it'll pass? Yeah.

322
00:23:11,280 --> 00:23:15,919
No, she us. I kind of almost just want

323
00:23:14,320 --> 00:23:19,760
to restart the computer and see if that helps it. Okay. It seems to be a bit

324
00:23:18,159 --> 00:23:23,919
confused. I like it. What makes this a Hail Mary

325
00:23:21,919 --> 00:23:28,559
run exactly? Uh, I don't even know what I said, so we'll just see if it works.

326
00:23:26,400 --> 00:23:32,400
What do we need? Like 50 points, something like that. Yeah. Yeah. Not

327
00:23:30,080 --> 00:23:36,000
much. Always the last 50. We're at our full minus 25 now. So, there's nothing

328
00:23:34,799 --> 00:23:40,720
more we're going to get from more cooling. Like, we've kind of been changing multiple variables as we've

329
00:23:38,880 --> 00:23:45,440
gone here because we've been adding more cooling capacity as we've been tuning

330
00:23:42,960 --> 00:23:53,280
it. Kind of building the train as it's flying down the rails. Come on. Oh,

331
00:23:50,640 --> 00:23:57,280
is that enough, boys? Wa. Now, that's a Hail Mary if I've ever seen one. Is that

332
00:23:55,679 --> 00:24:03,360
her name? The astronaut. Mary. What? Katy Perry. No, it's Mary. 45. Fourth.

333
00:24:01,120 --> 00:24:09,919
Number four. We're top five. Let's go. Let's go, Canada. Let's go. Damn. And

334
00:24:07,039 --> 00:24:15,520
these are LN2 scores. Really? Yeah. We went faster than Len. Yeah, man. Why

335
00:24:12,240 --> 00:24:17,760
were we able to beat an LN2 rig? Yeah.

336
00:24:15,520 --> 00:24:21,200
With a processor that really is not as optimized for this benchmark. What did

337
00:24:19,279 --> 00:24:27,279
we do right? So, what could the people do? They're using a D2. I think is maybe

338
00:24:23,679 --> 00:24:29,919
slightly less performance. Got it. But I

339
00:24:27,279 --> 00:24:36,400
think the cards might shut off at like minus 40. So we're almost there, I

340
00:24:33,440 --> 00:24:42,720
guess, cooling wise. Understood. That makes sense. So do you think if we

341
00:24:39,760 --> 00:24:49,200
had a 14900 in here that we would have a shot at top three podium finish? That's

342
00:24:45,279 --> 00:24:52,080
500 more points. That's a lot. Uh so I

343
00:24:49,200 --> 00:24:56,960
talked to Safe Disc and he said he tested it and you it was about a

344
00:24:53,760 --> 00:25:00,400
thousand more points. So, we maybe would

345
00:24:56,960 --> 00:25:03,360
have been up top here. I'm sorry.

346
00:25:00,400 --> 00:25:08,080
I'll go home and do that. Okay. Okay. I'll send you the chamber. Yeah.

347
00:25:06,559 --> 00:25:12,159
That's a nice card you've got yourself there. Not bad. She seems She seems

348
00:25:10,720 --> 00:25:16,559
pretty binned. It looks like you might have a shot at number one with that thing. Are we happy with this? I'm

349
00:25:15,200 --> 00:25:23,520
really happy. Yeah, that was great. Like, we're as we're as high as we're going to get. We squeakaked past Ralph.

350
00:25:21,039 --> 00:25:27,360
Squeakaked past. It's a really fast computer. Should we game on it? Be kind

351
00:25:25,360 --> 00:25:32,279
of silly if we didn't. What's the cost of the system plus the chiller?

352
00:25:32,400 --> 00:25:38,320
That's not a bad title. 50k gaming PC.

353
00:25:35,520 --> 00:25:42,559
Yeah. How fast is a $50,000 gaming PC? Never mind that most of it is not the

354
00:25:39,919 --> 00:25:48,480
PC. I can't believe we hit top five. Let's go. Never doubted it.

355
00:25:46,880 --> 00:25:53,679
Just we were getting ready to pack up. That's

356
00:25:50,159 --> 00:25:55,919
why he's so big, cuz he's full of

357
00:25:53,679 --> 00:25:59,679
Can I win Counter Strike now? Oh yeah, there we go. There we go. Oh. Oh, now

358
00:25:57,279 --> 00:26:04,600
we're going. We're going fast, boys. Got him. Got him. Look how fast that kill

359
00:26:01,600 --> 00:26:04,600
was.

360
00:26:05,360 --> 00:26:12,640
Damn it. Wait, our FPS isn't even that good. Oh, wait. We're at 1440p.

361
00:26:10,960 --> 00:26:19,080
Yeah, she doesn't look that bad. Let's go. Head shot. Boom.

362
00:26:15,840 --> 00:26:19,080
Damn it.

363
00:26:19,840 --> 00:26:27,600
Oh, how can I miss all three of those shots? And he ain't even a pro. Maybe

364
00:26:24,320 --> 00:26:29,520
his computer's in a giant freezer, too.

365
00:26:27,600 --> 00:26:33,600
That's probably it. He probably out freezered me.

366
00:26:31,919 --> 00:26:38,559
Hopeless. God, I missed that. Come on. Good

367
00:26:36,480 --> 00:26:41,559
movement. Good movement. All right. Come on.

368
00:26:43,039 --> 00:26:51,760
I give up. Uh, Oblivion Remastered. Here we go, boys. Okay, we're running at a

369
00:26:47,840 --> 00:26:54,080
casual 560 frames per second.

370
00:26:51,760 --> 00:26:57,919
Oh, no, no, no, no. She's good. She's good. No, I don't think it's good. Oh,

371
00:26:55,600 --> 00:27:04,480
she's not good. Uh, she's not good. She Oh, she's bad. That's a wrap.

372
00:27:00,799 --> 00:27:07,919
Rap. Just like this. Segue

373
00:27:04,480 --> 00:27:11,279
that you're going to rap. Um, to our

374
00:27:07,919 --> 00:27:14,159
sponsor, MicroEnter, rated number one in

375
00:27:11,279 --> 00:27:18,880
PC Mag's 2024 Reader Choice Awards. It's clear that MicroEnter is the place to

376
00:27:16,400 --> 00:27:22,720
shop for your next gaming or work setup. With close to 30 locations all over the

377
00:27:21,039 --> 00:27:26,640
US, they've been dishing out deals and providing excellent customer service,

378
00:27:24,640 --> 00:27:30,240
technical knowledge, and support for decades. Speaking of deals, from now

379
00:27:28,480 --> 00:27:36,240
until the end of May, they have deep discounts on over 150 desktop PCs, like

380
00:27:33,520 --> 00:27:40,720
this absolute baller gaming rig or this space-saving all-in-one solution, both

381
00:27:38,240 --> 00:27:45,760
from HP. And residents of the Bay Area rejoice. Their new Santa Clara location

382
00:27:43,279 --> 00:27:49,440
is set to open any time now. So, use our links in the description to shop

383
00:27:47,039 --> 00:27:52,720
Microenter's May desktop deals and to make sure you get your free flash drive

384
00:27:51,039 --> 00:27:56,559
when you check out their new Santa Clara location. Huge thanks to Splave for

385
00:27:55,200 --> 00:28:02,399
coming out. Thank you for watching. And if you guys are looking for another video to check out, how about that time

386
00:27:59,360 --> 00:28:05,440
slave came and we got a CPU to a nice

387
00:28:02,399 --> 00:28:05,440
6.9 GHz.