Seems like memory problems occur only during C++ compilation?
Seems like memory problems occur only during C++ compilation?
Hi. I recently assembled a new PC, aiming for top-of-the-line performance so it could serve me well for years. The components include: Motherboard: ASUS PRIME Z790-P CPU: Intel Core i9-13900K FPS: GPU: Gigabyte 4090 Memory: G.SKILL 64GB Ripjaws S5 DDR5 6000MHz CL30 KIT F5-6000J3040G32GX2-RS5K PSU: Seasonic Prime PX 1600W Storage: GIGABYTE 2TB AORUS NVMe GEN4 SSD M.2 PCIe M.2 2280 plus an Arctic AIO cooling system. This is my personal machine, and I spend a lot of time gaming and working on projects. I began focusing intensely in August for a project deadline and noticed strange behavior. Randomly, Visual Studio—especially during rebuilds—will report memory issues or throw up odd errors, like complaints about random characters in header files that don’t belong to the Windows SDK. Compiling multiple times sometimes helps, so I managed to meet my deadline but then dug deeper. The same problems appear with other systems (Intel NUC, laptop, work PC) that I’ve used over the past 13 years. A colleague faced similar issues with a new PC until adjusting memory timings resolved it. Notable points about the problem: * Running Visual Studio as a 64-bit app helps a bit, but the errors become less frequent and often disappear after a full rebuild. * The compiler randomly reports three types of errors: heap out-of-memory, internal compilation warnings, or strange complaints about code that should be fine, suggesting possible memory corruption. I attempted several fixes—new Windows install, BIOS update, memory test, disabling XMP, turning off efficiency cores, shuffling sticks, changing frequencies, testing individual modules, and even replacing the RAM. However, my exact RAM configuration isn’t listed on the motherboard’s support list anymore, so I ended up getting a replacement kit today: Kingston Fury 64GB Beast DDR5 5600MHz CL40 KIT. Despite these efforts, the issue persists. Interestingly, the machine performs flawlessly for gaming and development tasks. It only fails as a build server when memory problems arise. Could you share some advice on more thorough testing methods to pinpoint the cause? Thanks.
You might consider disabling XMP while maintaining DRAM voltage and CPU turbo settings to address instability at high clocks. Have you checked the SSD? When compiler errors appear, do you see any details in the event viewer? Are there shared characteristics between the two configurations? Is Windows security enabled (such as virtualisation or HVCI) that could affect Visual Studio compatibility?
This appears to be a straightforward repair request. The issue seems resolved by turning off the turbo boost, suggesting it might be an RMA scenario.
You could try increasing the stock voltages a bit with an offset and see if that helps, with these CPUs I'm not completely convinced how much they're tested to ensure 100% stability with the stock profiles and presumably there's still some variance to account for. If it passes Intel's diagnostics ( here ), then I'd assume it would not be considered broken, unless it reproducibly fails something like Prime 95.
It passed the Intel diagnostics and most other tests I ran. It seems compiling code is especially challenging. After learning about the turbo boost issue, several users reported similar problems with that CPU and motherboard setup, so I headed back to the shop for an RMA. Thanks for guiding me through the next steps—I was stuck for weeks.
I ran additional tests and noticed the CPU temperature jumps sharply right when I start a build. It’s unclear if it’s due to the AIO failing to keep up or a genuine CPU problem, but this is definitely concerning. Also, performing several rebuilds in just about two minutes triggered the first BSOD I’ve encountered, especially with such high values like 113—definitely something to investigate.
Did you turn on the turbo settings for this test? It seems the voltage rise might be the cause, as a 13900KF includes extra boost methods: thermal velocity boost and adaptive boost, plus a turbo boost limit of 3.0. You might want to disable them individually or together to see if it reduces overboosting. I’m not sure how an AIO would handle only one core getting hotter—perhaps you can adjust its settings in the BIOS to change its sensitivity and response speed.
It's been four weeks since the PC was returned and today the supplier confirmed the problem. I'm receiving a full refund for the CPU. It's nice knowing that the 14th generation has just launched, and since the original price covers the 14900kf, the concern about upgrading from the 13th generation is no longer relevant. Thanks to everyone for your support. I hope I can get my PC back in a week.