Remember the infamous RTC bug that made AMD Ryzen 3000 (codename Matisse) chips run seemingly faster than expected when in reality, they weren't? It now looks like the bug never went away, and it occasionally affects AMD's latest Ryzen 7000 (Raphael) processors that rival the best CPUs on the market.
Russian tech blogger Pro Hi-Tech (opens in new tab) shared a video from one of his subscribers showing a Ryzen 9 7950X coming out of sleep mode to reach a clock speed of 6.28 GHz. It's clearly a bug since the Ryzen 9 7950X has a 5.7 GHz clock speed, so it's impossible that the 16-core chip would surpass 6 GHz on its own without manual overclocking. Unfortunately, the user didn't share the model of the motherboard or the firmware version. Therefore, we're uncertain if a specific AMD AGESA firmware is causing the bug.
The trick reportedly consists of having the Ryzen processor reach its maximum clock speed and then putting the system to sleep. As a result, the clock speed magically skyrockets exponentially. The processor even appears to run faster in benchmarks, such as Cinebench or Corona. In actuality, the processor isn't operating beyond its specifications; instead, what appears to be the return of the RTC bug is fooling the benchmarks.
The RTC bug dates back to 2013, around the Windows 8 era. The flaw was more prominent on Intel processors then, including Sandy Bridge, Ivy Bridge, and Haswell. AMD processors were affected but not to the degree of their Intel counterparts. The RTC bug mainly involved altering the processor's bus frequency. Changes made inside the BIOS were acceptable, but modifying the frequency on a software level would mess up Windows 8's timer.
Benchmarks, including Cinebench and Corona, measure the processor's time to render a scene. A slower internal timer makes the software think the processor takes less time to complete the benchmark. If users were to measure the time with a stopwatch and compare the time exhibited by the software, a significant disparity would be more than evident. The user, who sent in the video, observed a 15% difference in rendering times between the two measurements. Therefore, an abnormally high clock speed combined with the RTC bug creates an illusion that the Ryzen processor is delivering more performance, but it's not.
Microsoft reportedly solved the RTC bug with later versions of Windows, but Zen 2 chips were still susceptible to the issue. We're at Zen 4 now, and the bug apparently hasn't disappeared. It's perplexing that a bug discovered almost ten years ago still lingers in the shadows. The bug doesn't break your system or cause a performance hit. On the contrary, it convinces less-experienced users that their processor is over-performing. At the end of the day, it's a placebo effect, and users and reviewers should be careful, especially the latter since the bug will incline the scales toward Ryzen processors.
And similarly, the following systems have been 100% trouble-free for me:
Ryzen 5600X + RX 6700 XT
Ryzen 5600X + RX 6600 XT
Ryzen 5600 + RX 5700
Ryzen 5600G using iGPU
Ryzen 1600AF + various GPU (now used as a test-bed, previously RX 5700 long term, then GT 730, now RX 5300)
Athlon 200GE using iGPU
That somehow had been a problem for a very long time since day one, but somehow was never returned or replaced under warranty.
And, I'll point out, it's a post that is completely irrelevant to this thread's topic. Just like your previous comment about GPUs.
And, your absolutely false statement of AMD video cards being "objectively worse"