Question about overclocking Ryzen 7 1700X and its impact on performance.
Question about overclocking Ryzen 7 1700X and its impact on performance.
I raised my base speed from the default 3.4GHz to 3.8GHz. I also turned off core performance boost. No other adjustments were made. Voltages remained set to auto.
The OC seems stable. I've tested Prime95 multiple times without any crashes, including several short "Small FFTs" sessions and a couple of one-hour "Blend" runs. My CPU temperature stayed under 80°C during these tests.
Then I performed nearly identical tests in Battlefield V, once before the OC and once after. Before the OC, my average FPS was 126. After the OC, it dropped to 122!? That means I actually lost about 4 average FPS.
Please advise.
What is the speed of your memory modules and where are they installed? This information might be useful. It’s not critical, but it could matter.
Are your BIOS settings current with the latest stable version?
Even if your core temperatures stay under 80 degrees, your VRMs may still be throttling, which can lead to noticeable performance drops. I recommend using HWinfo (not HWmonitor or any other tool) and selecting the "Sensors only" option during startup. The Summary feature is likely ineffective.
I often think this could point to a power delivery issue, where overclocking pushes the PSU beyond its capacity to supply enough power to the GPU or CPU. However, this unit is quite capable; unless it’s an older model, that’s unlikely.
Memory modules operate at 3200MHz and are installed in slots #2 and #4. I possess the second latest BIOS, version 2202 from April this year. However, I notice there is a newer release dated May 15th? I have HWiNFO64 but can't locate any readings for "VRM". Could it be labeled differently? Thanks!
I repeated the OC test in Battlefield V once more. My average FPS increased to 123, which remains three less than without the OC. This shows my results are consistent but not very encouraging.
In the Heaven Benchmark, however, my average FPS improved by 2% with the OC. It’s a small gain, but at least it’s positive.
Did you check if your core frequencies remained steady during testing, ensuring the clocks stayed at their intended speed?
I haven't made much changes with Ryzens, so I might be mistaken. However, could it be that your temperatures are quite high, and XFR actually raises the frequency above 3.8 GHz on default settings? This would result in a higher score compared to manually setting it at 3.8 GHz?
I just completed the Heaven Benchmark and monitoring.
All core clocks remained within 3,791.6MHz to 3,793.0MHz.
CPU/thread usage fluctuated between 1.3% and 95.0%, with an average of 46.2%.
Total CPU usage ranged from 0.2% to 29.5%, averaging 6.9%.
Does everything look typical?
This isn't an effective method for comparison. Game performance can differ based on location and other factors. You'll see fewer frames on your overclocked BF1 bench due to the intensive tasks you're performing.
You should evaluate your CPU using tools like Cinebench R15.
Start with the baseline, then increase the settings and recheck.
Not sure about how Heaven evaluates differently each time or if framerate changes based on your perspective since it's always the same test. I admit if your system lacks automatic defragmentation and other cleanup features, a Windows process might be affecting performance, but the difference should be negligible—around 4%. While I understand overclocked values should be higher than stock ones, the actual numbers could vary. I suggest downloading Realbench, running the benchmark, and comparing the results. Make sure to disable automatic defragmentation and other unnecessary processes before testing to ensure accurate outcomes.
I’m planning to run RealBench soon—it’s the next step I’m taking. Thanks for the suggestion!
Update: I switched my core voltage from auto to 1.3375V. This should help save electricity. I noticed crashes in Prime95 when the voltage was below that value.
My testing plan now looks like this:
1) Prime95 "Small FFTs" for 15 minutes, AVX disabled. Completed for thermal checks; temps stayed low.
2) RealBench "Stress Test" for 8 hours to check system stability.
3) Prime95 "Blend" for 8 hours for stability.
4) Heaven Benchmark for my video card overclock.
5) Battlefield V test runs for real-world use.
I’ll share the results as they come in.
What are your thoughts on this approach?