High-performance 5900x featuring exceptional assassin 120 temps and rapid clock speed.
High-performance 5900x featuring exceptional assassin 120 temps and rapid clock speed.
You recently upgraded from a 5600x to a 5900x. After selling your old unit, you spent $50 on a new CPU. You wanted more cores or threads for better performance, which was a major improvement. The only cooler I own is a Wraith Prism. In Cinebench R23 with multi-core enabled, I’m seeing an average of 21,104, scoring 85 on a 10-minute test at speeds between 4.05 and 4.1GHz. Gaming results are similar. I don’t like air coolers; the Peerless Assassin 120 looks like the best value. If I need an air cooler, I’ll consider it but prefer not to. Are anyone else using a PA120? If so, what temperatures, clock speeds, and scores do they report with a 5900x? I’m also curious about other cooler options for my Be Quiet DX500 with the 6×140mm silent wings fans.
your temperatures are normal during a 10 minute run, yet your 5900x reaches 4.5 on all core. I also test the 5900x on an x570 Aorus master. there’s no overclocking and XMP is turned off. my Cinebench scores stand at 22841. the cpu occasionally hits 82c to 83c.
The temperature rises and the clock speeds decrease until reaching 4.05ghz. Since there’s no overclocking, I’m using the built-in settings. Which cooler are you running it with?
You're experiencing quite an unusual behavior with the boost algorithms. They seem to interpret frequency data in unexpected ways—sometimes showing one frequency but actually behaving differently. I've noticed cases where 23,000 points are logged for a 4.4GHz signal, and other times 18,000 points for a 4.8GHz reading (with real-time priority ensuring background tasks don't disrupt the measurement). This varies based on BIOS settings and AGESA configuration. What really matters is the final score; the reported frequency seems less important. To the OP, I haven’t used a PA120 with my 5900X, though I did test it on a Scythe Fuma 2 for a week. Temperatures were acceptable, mid-80s, and PBO cinebench scores were around 22,500 (about ±500 points). The PA120 is a close match to the cooler, so performance should be better than the stock unit—definitely worth trying.
Focus less on the clock speed during R23 and check the score. I've noticed it operates between 4.4GHz and 4.65GHz at full stock, but the scores differ by about 100 points due to PBO glitches reporting different speeds. Adjusting the right PBO variables and BIOS version can push it up to around 5GHz, though it won't reach the expected performance. The score should fall roughly between 22,000 and 23,000 based on Windows cleanliness. Fine-tuning PBO settings or the curve optimizer might help exceed 23,000 slightly.
The issue with Ryzen processors is their clock stretching behavior—performance drops even when clocks stay constant. The most effective fix is applying a negative voltage offset. To undervolt safely, use the curve optimizer instead of manually setting negative offsets.
Here’s a revised version of your text:
I understand, I didn’t realize that. The results I’ve gathered so far are as follows:
- Wraith Prism PBO enabled, with a 10-minute R23 run at 4.05ghz and an 85C MC score of 21104 using Pa120 voltage on auto.
- PBO off, same 10-minute R23 run at 4.1ghz and 62C MC score of 20805 with Pa120 voltage on auto.
- PBO on, +200 pBO limit set to 10-minute R23 run at 4.3ghz and 74C MC score of 21695.
- PBO off, +200 pBO limit set to 10-minute R23 run at 4.35ghz and 69C MC score of 22076.