Game patch notes recommend downclocking high-end Intel CPUs for the sake of stability

Daniel Sims

Posts: 1,376   +43
Staff
PSA: Most patch notes for PC games inform players about the developer's progress in fixing bugs. However, the situation with Outpost: Infinity Siege appears to be so bad that a recent update mainly consisted of steps players should take to avoid crashes, including underclocking certain Intel CPUs. Admittedly, the problem isn't entirely with the developer in this case.

The developer of Outpost: Infinity Siege suggested that players using Intel i9-13900K or 14900K CPUs should downclock them from 5.5GHz to around 5GHz to avoid crashes. The tower defense first-person shooter hybrid game is the latest to suffer frequent crashes on these processors.

Developer Team Ranger launched Infinity Siege earlier this week and has issued four patches in the few days since to address crashes and other problems that are likely dragging down its Steam rating. Although the patch notes highlight many issues the studio has addressed on its end, it also gave users suggestions for minimizing crashes.

Aside from underclocking the CPU using Intel XTU, Team Ranger advised players to change power settings to energy-saving mode, verify game file integrity, and use the latest graphics drivers. However, Infinity Siege may have simply launched at the wrong time.

Since February, users with 13th and 14th-generation Intel i7 and i9 processors have reported frequent crashes in many games and other applications. The problem disproportionately but not exclusively affects Unreal Engine titles (like Infinity Siege). Examples include Remnant 2, Nightingale, CineBench, and Handbrake. Crashes have also been reported during the shader compilation processes in Star Wars: Jedi Survivor, Lies of P, and Horizon: Forbidden West. Furthermore, Team Ranger suggests that GeForce RTX 4090 owners should undervolt, but others haven't mentioned related problems stemming from GPUs.

The developers of the popular Bink video codec and Oodle data compressor blame flawed default BIOS settings for making the affected processors too power-hungry while under load. The available evidence suggests the ball for fixing the problem lies entirely in Intel's court.

The issue should be a priority for Intel, as most users likely aren't experienced with overclocking or underclocking, but those affected could take this as an opportunity to learn in the meantime. Although undervolting using Intel XTU can help, doing so through the BIOS instead ensures that the system remembers the new settings. Disengaging overclocking functions from motherboard vendors is a given.

Permalink to story.

 
Yep, 13th Gen and 14th Gen are garbage at default settings. If you really use the cpu in productivity workloads or games really push its capabilities to the max, you'll find they are completely unstable.
We had several pc's at work with these processors and with all of them we had to set power limits to get them to stop corrupting files or crashing applications, which reduces performance by about 10%.The errors are really subtle, we had only one full blown bsod, but many unexplainable app crashes and data corruption. At first you think there's something wrong with the applications, or maybe system memory since it doesn't happen all the time, but after data corruption and the problems getting worse over time and having crashing stress test applications, it got clear it can only be the cpu. It was really terrible and cost us a lot of time and frustration and even data loss.

Asus motherboard defaults are the worst, but even other vendors do not come stable out of the box.
It's a nightmare. I don't know what happened but Intel is delivering extremely hot power hungry processors that are unstable in demanding work loads. Never thought we'd be at this point but I'm now avoiding Intel when I can until they get their act together.
 
Last edited:
For this CPU (13900K) some sites even gave Editors choice. Luckily Techspot only gave 75/100. Something done right.

This is another example of benchmarking problems. Review sites say "there are no real problems with high temps/power consumption because our benchmarks do not reveal them". Basically anything you don't see on benchmarks do not exist.

As usual, benchmarking practices do not allow reviewers to use their brains. And things like this "no-one couldn't know about" happen.
 
Back