F5F Stay Refreshed Hardware Desktop The higher limiting voltage improved cooling and performance by allowing better thermal management.

The higher limiting voltage improved cooling and performance by allowing better thermal management.

The higher limiting voltage improved cooling and performance by allowing better thermal management.

Pages (2): 1 2 Next
F
64
10-23-2019, 07:47 PM
#1
I'm just starting out with Ryzen and was impressed by my new Ryzen 3600. It was quite unexpected how much cooler it ran compared to my previous Intel 4930k. The temperatures dropped significantly—from 27°C at idle to 33-40°C during stress. With my Kraken x62, I expected better cooling, but this chip actually performed hotter. I wanted a quieter system for rendering in Blender, so I tried undervolting. The results were surprising: under load temps went down to 71-73°C, and performance stayed stable over time. I haven’t seen any issues with voltage limits or instability. My questions are: Why does lowering peak voltage make the CPU faster without any drawbacks? How does it actually work? And why does this chip, which has half the TDP of my old one, run hotter? Could there be hidden downsides? For instance, does a lower voltage improve scores in Cinebench r20, or could it hurt performance under real workloads?
F
FuturzLeCanape
10-23-2019, 07:47 PM #1

I'm just starting out with Ryzen and was impressed by my new Ryzen 3600. It was quite unexpected how much cooler it ran compared to my previous Intel 4930k. The temperatures dropped significantly—from 27°C at idle to 33-40°C during stress. With my Kraken x62, I expected better cooling, but this chip actually performed hotter. I wanted a quieter system for rendering in Blender, so I tried undervolting. The results were surprising: under load temps went down to 71-73°C, and performance stayed stable over time. I haven’t seen any issues with voltage limits or instability. My questions are: Why does lowering peak voltage make the CPU faster without any drawbacks? How does it actually work? And why does this chip, which has half the TDP of my old one, run hotter? Could there be hidden downsides? For instance, does a lower voltage improve scores in Cinebench r20, or could it hurt performance under real workloads?

G
GreenLightFabi
Senior Member
696
10-23-2019, 08:11 PM
#2
The results came from various frequencies, and the improvement continued regardless of your voltage limits.
G
GreenLightFabi
10-23-2019, 08:11 PM #2

The results came from various frequencies, and the improvement continued regardless of your voltage limits.

I
ItzEthqn
Member
68
10-24-2019, 12:52 AM
#3
It depends on frequency; if clock stretching isn't considered, heat output remains low, but the smaller process node dissipates less heat, leading to higher internal temperatures. Zen 2's thermal efficiency would likely lag behind Intel's if they rely on paste instead of solder. Performance may drop in other workloads, especially with lower power consumption (R20), making stability a concern.
I
ItzEthqn
10-24-2019, 12:52 AM #3

It depends on frequency; if clock stretching isn't considered, heat output remains low, but the smaller process node dissipates less heat, leading to higher internal temperatures. Zen 2's thermal efficiency would likely lag behind Intel's if they rely on paste instead of solder. Performance may drop in other workloads, especially with lower power consumption (R20), making stability a concern.

M
Mokai_Mo
Member
212
10-31-2019, 04:54 AM
#4
I agree, you're right. It turns out that choosing 4.2 GHz actually increased core counts beyond what Cinebench boosts. I ran tests at different voltages and confirmed no impact on scores.
M
Mokai_Mo
10-31-2019, 04:54 AM #4

I agree, you're right. It turns out that choosing 4.2 GHz actually increased core counts beyond what Cinebench boosts. I ran tests at different voltages and confirmed no impact on scores.

M
Maxavo91
Member
158
10-31-2019, 06:04 AM
#5
Sorry about that! I usually do that when I tweak my Ryzen processors... hehe.
M
Maxavo91
10-31-2019, 06:04 AM #5

Sorry about that! I usually do that when I tweak my Ryzen processors... hehe.

P
pillowfight
Junior Member
20
10-31-2019, 07:37 AM
#6
The 3600 generally maintains a consistent overclock around 4.1 to 4.2 at speeds between 1.3 and 1.35v. Both of my systems stabilized at 4.1 at 1.32v during the Prime95 stress test. Most motherboards I've used have voltage regulation that pushes them higher than 1.5v, leading to increased heat. I'm not certain if this is due to PBO or another factor.
P
pillowfight
10-31-2019, 07:37 AM #6

The 3600 generally maintains a consistent overclock around 4.1 to 4.2 at speeds between 1.3 and 1.35v. Both of my systems stabilized at 4.1 at 1.32v during the Prime95 stress test. Most motherboards I've used have voltage regulation that pushes them higher than 1.5v, leading to increased heat. I'm not certain if this is due to PBO or another factor.

K
Kacper_Bored
Senior Member
389
10-31-2019, 01:58 PM
#7
I noticed some unusual patterns I hadn’t observed before. It turns out I needed to adjust the frequency to control peak voltage, since the standard setting changes depending on the workload.
K
Kacper_Bored
10-31-2019, 01:58 PM #7

I noticed some unusual patterns I hadn’t observed before. It turns out I needed to adjust the frequency to control peak voltage, since the standard setting changes depending on the workload.

Y
ylyes4
Senior Member
572
10-31-2019, 03:16 PM
#8
Which motherboard are you referring to?
Y
ylyes4
10-31-2019, 03:16 PM #8

Which motherboard are you referring to?

T
toivo8
Junior Member
3
11-04-2019, 10:57 AM
#9
Impressive spec! Most users expect under 1.30v at 4.2ghz, but this is quite high.
T
toivo8
11-04-2019, 10:57 AM #9

Impressive spec! Most users expect under 1.30v at 4.2ghz, but this is quite high.

C
209
11-04-2019, 02:42 PM
#10
Thank you for your response, I believe it addresses all your concerns. Appreciate the help!
C
CaptainFurioux
11-04-2019, 02:42 PM #10

Thank you for your response, I believe it addresses all your concerns. Appreciate the help!

Pages (2): 1 2 Next