Temperature & Noise (Stock Cooler)
Tctl Sensor Readings Leave Many Open Questions
One of the lessons learned during the Ryzen 7 launch was that AMD's temperature sensor output has its share of problems. The company configured some of the models to report an additional 20°C on top of the actual temperature in order to guarantee adequate cooling. We’re still puzzled by this choice. Read more about this in our Ryzen 5 1600X coverage.
We’re primarily using Alphacool's Eiszeit 2000 Chiller with its 1/4” threads and powerful pump yet again. We’re also employing the stock cooler to determine the remaining temperature differences in a more real-world setting.
Temperature vs. Power Consumption
Our cooling solution operates completely independently from the waste heat put into it during all of the tests. Once again, this allows us to plot the Tctl values against actual power consumption.
In order to compensate for various jumps and delays, we use a tool that we programmed ourselves to generate different loads over time and distribute them as evenly as possible among the different threads. The results of our efforts confirm what we’d suspected:
Below 20 to 25W, the temperatures are complete nonsense. However, there are large differences between AMD’s Ryzen 3, Ryzen 5, and Ryzen 7 when it comes to interpreting their Tctl sensor readings. We were somewhat surprised to see that there doesn’t seem to be an offset for the Ryzen 3 1300X. Then again, this could also be due to the motherboard’s new BIOS, though.
It’s our suspicion that the Tctl values are calculated by the CPU internally, and that its division gets less accurate with decreasing core count. All of the values look much too high to us. Consequently, we'll use a corrected difference that's based on our comparison curves.
Implications For Air Cooling
The Ryzen 3’s stock cooler gives us a more practical look at operating temperatures, and we manually set it to the speed it reached during our first stress test. Two sharp bends are still apparent in the line graph, even though the overall curve has a much steeper slope. We weren’t able to detect any meaningful leakage currents.
Corrected Temperature Results With Water Cooling
We now take the predicted differences, or deltas, for the three disciplines we test into account.
Stock Cooler Noise Level
The Ryzen 3 processor fits squarely into the 70W category. Even a simple aluminum block doesn’t have a hard time dealing with this level of waste heat. Depending on your motherboard and fan profile, the cooler should only generate between 33 to 36 dB(A). This is a solid result; it's not loud at all.
MORE: Best CPUs
MORE: Intel & AMD Processor Hierarchy
MORE: All CPUs Content
I don't remember when it was last time when we saw a review where, practically, in every chart AMD was at the top, and Intel at the bottom. It is refreshing really. Hopefully this will drive Intel to cleanup the mess it created in 5+ years of unchallenged leadership, and get its acts in order.
I'm enjoying the moment though. With my Ryzen 5 1600 :)
What I'm really curious to see for the mainstream/office segment is the Ryzen-Vega APUs.
Which is a great great thing after Intel's dominance & stranglehold for the last 6 years.
Intel price war ??
No that's not going to happen if it hasn't already ,they're way too arrogant to ever do that.
Too bad their Vega card seems like another Fury. I sold my stock and wait for the first benches before rebuying it. Also, indication is that RX will cost 600-650$ and barely beating a 1080 GTX. However Vega core are going to be incridilbe with AMD APU on the laptop segment. Nvidia MXM cards are going to be a bad investment and cost a fortune while everything will be on the same chip for AMD.