AMD Ryzen 7 1700 Undervolting Revisited with Wraith Spire Cooler

I did undervolt test on AMD Ryzen 7 1700 on mid-April 2017. The result was kinda disappointing due the power consumption reduction being fairly low. Back then I changed the voltage on the UEFI from 1.1875 V to 1.0875 V. But actually in CPU-Z the voltage reading when running at stock was only 1.072 V (compared to 1.1875 V in UEFI). Setting it to 1.0875 V in UEFI ended with 1.04 something in CPU-Z reading. the reduction is just around 0.03 V which is the reason why power consumption reduction was also very little. So, I decided to re-run the test, now with even lower Voltage. This time I set the voltage to 1.000 V in UEFI. This resulted with 0.976 V in CPU-Z reading, making it close to 0.1 V reduction or 0.0996 V to be more precise.

Any, I changed the setup a little bit. I decided to use Wraith Spire cooler instead of OCZ Vendetta 2 to see the whether the temperature reduction is more visible or not.

CPU AMD Ryzen 7 1700
Motherboard ASRock AB350 Pro4
RAM 2x Crucial 8GB DDR4
2666 MHz 16-16-16-36 1T 1.2V
Graphics Card AMD Radeon R9 290 Reference, undervolted by 0.1 V
Storage Samsung SSD 960 EVO 250 GB
Samsung Spinpoint F3 HD103SJ 1TB
Cooling AMD Wraith Spire
Fans Only in Wraith Spire
PSU Seasonic P760 (80 Plus Platinum)
Other Components attached Power Logic Xenon Sabre
Cooler Master CMStorm Sentinel Advance II
SADES Wolfgang
ADATA MyFlash 1 GB USB Flash Drive

Now let’s jump to the power consumption summary. This time the result is more pleasing. R7_1700_undervoliting_Power_Consumption_Charts

The reduction in power consumption during active load varies from 13 Watts to 19 Watts. This means 10% to 13% power consumption reduction depending the type of work load but this only applies to applications that utilize all cores.

Moving to temperature, as I stated before, I switched the cooler to Wraith Spire to see whether the temperature difference will be more noticeable or not. To make it even more obvious, I disabled every fan in my Chassis but I kept my side panel open. Ambient room temperature is 28.8 °C which is not ideal if you were doing a lab test. But I am not doing it in lab anyway. I am doing it in my room with Air Conditioning turned off to give this CPU a taste of summer heat.


The result turns out to be quite noticeable. You’d be surprised with the temperature during Handbrake workload. 83.5 °C at stock settings is amazing, isn’t it? Don’t worry, that is only the peak temperature. Most of the time, temperature is around 75 °C during Handbrake. Undervolting provides significant decrease on the peak temperature by 16 °C. Anyway, the average temperature during Handbrake when undervolted is around 63 °C. So, it is 12 °C difference in average and 16 °C at peak with 19 Watts difference in power consumption. Moving to Cinebench, the temperature when undervolted is only 2 °C lower than on stock settings. Then, we see a 5 °C difference in CPU-Z stress test. You might be notice why temperature during Cinebench is lower than CPU-Z stress test eventhough the power consumption during Cinebench is higher than CPU-Z stress test. That is because I run CPU-Z benchmark at longer time than Cinebench.

Speaking of cooling, I can understand why AMD does not include Wraith Spire with Ryzen 7 1700X and Ryzen 7 1800X. Simply because the cooler won’t be able to handle the CPU during intense workload, especially with a slight taste of summer heat. But I still don’t understand why they didn’t include Wraith Max in the package from the start? And why do they change the cooler mounting of AM4? The only reason I can think is AMD wanted to share some profit with 3rd party cooler maker because people will buy new cooler. Overclockers don’t need stock cooler because they use after market cooler and cooler maker will be happy.

Any drawback in performance?
Take a look at this result of Cinebench and CPU-Z Stress Test.



See the score? They are with margin of error each other, virtually identical. But actually, there is a drawback. To make this undervolt possible, I actually overclock to 3.2 GHz from 3.0 GHz making it similar to stock settings when running on Precision Boost. However, once I changed the default clock speed, Precision Boost is disabled. So, there will be no Boost and XFR up to 3.75 GHz under single threaded load. That means you lose 0.55 GHz of speed during light works. Does it matter? In my case, I don’t care because single threaded applications usually ended up much quicker than multi threaded applications. Since there is no difference in multi threaded applications, then I am fine with it.

How about gaming?
Well… that is a question that I cannot answer at the moment. You know… The problem is I am stilll using AMD Radeon R9 290 which is very power hungry compared to newer Graphics Card from Nvidia. So, gaming power consumption test is not so good with this Graphics Card. I might do this once I have a much more efficient Graphics Card.

Why don’t you just overclock it?
I have a good reason. My work finished much faster in this CPU at stock speed compared to when I was using my old Core i5-3570K even when overclocked to 4.5 GHz. I don’t feel there is a need to overclock this CPU. For gaming, Radeon R9 290 takes the hit before Ryzen 7. Unless I use Geforce GTX 1080 or anything faster, there is no need to overclock the CPU.

Anyway, I am now happy with the result. This CPU provide a fantastic performance at stock speed, even better when overclock. Power consumption is fairly low at stock settings, even better when undervolted. Now, you can see the benefit of undervolting Ryzen 7 1700. At best case, it reduces the power consumption by up to 19 Watts. Splendid. This is the most efficient 8-Core CPU for Desktop PC and you can make it even more efficient!


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s