Some recent Intel microprocessors are crashing systems – and the problem appears related to the chips’ firmware and clock rates settings.
RAD, a maker of game development tools for Epic Games’ Unreal Engine, on Wednesday published a note that describes decompression failures with its Oodle Data software and crashes in games built with Unreal.
The coding biz, a subsidiary of Epic, said it believes the errors stem from a hardware glitch affecting mainly Intel Core i9 13900K and 14900K processors, and possibly 13700, 14700, and related silicon too.
Some small percentage of processors go out of their functional range of clock rate and power draw under high load, and execute instructions incorrectly
“Only a small fraction of those processors will exhibit this behavior,” RAD said. “The problem seems to be caused by a combination of BIOS settings and the high clock rates and power usage of these processors, leading to system instability and unpredictable behavior under heavy load.”
The game tools biz says the issue is not a software problem. Rather, it claims, “Due to what seem to be overly optimistic BIOS settings, some small percentage of processors go out of their functional range of clock rate and power draw under high load, and execute instructions incorrectly.”
RAD says this is more apparent when Oodle Data decompression is being used, possibly due to the extra data integrity checks needed compared to other applications. However, it’s not just those running RAD’s decompression software who may experience this issue. The game tools biz claims standard benchmark tests and stress tests will also elicit chip errors.
Crashes have also been reported in applications including RealBench, CineBench, Prime95, Handbrake, and Visual Studio, according to RAD. And in Intel’s own forums, posts like “What is wrong with the 14900k?”, “i9-13900K: very frequent crashes (Windows 11) with apps, games and benches”, and “CPU not running as expected” suggests others have encountered the problem. Similar reports can be found on forums like Reddit.
Gelsinger splits Intel in two to advance foundry vision
Arm targets AI performance with latest Neoverse Compute Subsystems
India won’t become a semiconductor superpower anytime soon, says think tank
Sam Altman’s chip ambitions may be loonier than feared
In an update to his 2016 post predicting an increase in CPU bugs, Dan Luu, who worked previously as a software/hardware engineer for Microsoft, Google and X, took note of the RAD report and said that while the bug may follow from system configuration settings, Intel knows their CPUs are shipping with these settings.
“If they wanted their CPUs to not crash due to this issue, they could have and should have enforced these settings as well as some others,” he said. “Instead, they left this up to the BIOS settings, and here we are.”
Luu echoed several system tweaks suggested by RAD that may mitigate crashes, like setting “SVID behavior” to “Intel fail safe,” and reducing the power settings for “Long duration power limit,” and “Short duration power limit”.
According to Luu, Intel used to do better at chip verification than AMD, but over time the two have become more comparable in terms of the level of serious chip bugs discovered. He observes that Intel seems unlikely to improve its verification story amid intense competition with AMD, Arm, and Nvidia.
Intel did not immediately respond to a request for comment though reportedly the chip biz is aware some of its Core 13th and 14th Gen desktop processors have had issues with certain workloads. ®
>>> Read full article>>>
Copyright for syndicated content belongs to the linked Source : The Register – https://go.theregister.com/feed/www.theregister.com/2024/02/23/intel_core_13th_crash/