Re: Hogwarts Legacy Shader Crash At Launch For 13th Gen/13900K

VaultBoi · ‎01-02-2025

Since game release last year, I've had ZERO technical launch issues launching/playing Hogwarts Legacy. Which is the only game in my Steam library designed in UE5 to date. So this issue is HL specific as I'm able to play every other game in my Steam library to date with zero VRAM issues.

A recent Nvidia RTX 4090 driver update also seemed to exacerbate whatever technical issues the game potentially had with my 13900K CPU. The result seems to be a hardware p-Core/voltage issue that has adversely blocked the ability of Hogwarts Legacy to successfully launch on Steam. Note I have NEVER overclocked (and never intend to) overclock my CPU since buying it a year ago. My dxdiag is here:

So this issue appears to be a UE 5 software v. Intel 13th/14th gen v. Nvidia RTX 4090 hardware issue. After all the finger pointing and researching the web, it seems the root cause is the auto setting in the BIOS for the p-Core (and possibly other efficient core settings). Tom's Hardware gives a summary on why this may be the issue:

https://www.tomshardware.com/pc-components/cpus/nvidia-blames-intel-for-gpu-vram-errors-tells-geforce-gamers-experiencing-13th-or-14th-gen-cpu-instability-to-contact-intel-support

Was redirected by Steam games forum to your XTU utility here

https://www.intel.com/content/www/us/en/download/17881/intel-extreme-tuning-utility-intel-xtu.html

I installed and ran the XTU Setup_7.143.2.14.exe and on launching the app, I got the following XTU app UI window:

Intel XTU App.png

My question is given my CPU & GPU setup, what is the best setting to get Hogwarts Legacy shaders to successfully complete so I can start the game again?

I have seen a lot of feedback of lowering the pCore value to 54 or 53. But I don't want to change any of the core ratios that could adversely impact my CPU and other critical hardware components. Again this tweak is only to get Hogwarts Legacy to launch as other 4090/13th gen gamers apparently have succeeded in doing on Steam/Reddit.

Some gamers even recommended a permanent fix of changing these values directly in the BIOS. Instead of using the XTU app utility. Which of these options is best?

pressed_for_time · ‎01-02-2025

The underlying cause here is that Hogwarts Legacy is written using Unreal Engine 5. Game authors using UE5 use an internal Shader/Resource compression utility called Oodle. Oodle offers the highest levels of compression, and consequently when a game using it is run it needs the highest levels of decompression or shader compilation. This puts a huge load on the CPU and storage such as SSDs. A feature of Oodle is that it enforces a zero tolerance policy for errors and will exit games with an error message which is perceived as a crash.

And it can be sensitive to the Core frequency of Intel 13th/14th gen processors which is why reducing core speeds to 54x or 53x can stop games crashing.

Some users find that the error code generated by Oodle changes with each attempt to run the game. So by continually running the game Oodle will eventually sucessfully decompress/compile all the game resources required and the game will then run. An update to the game will trigger this process all over again.

Updating your BIOS to the latest release and installing the latest drivers from the support site for your motherboard may help.

VaultBoi · ‎01-03-2025

THANK YOU so much for taking the time to read my rant and explain the issue! I really thought there was something wrong with my hardware and couldn't understand why that was since my PC is only 1 year old. That being said, what you said here:

"And it can be sensitive to the Core frequency of Intel 13th/14th gen processors which is why reducing core speeds to 54x or 53x can stop games crashing."

Next to updating the BIOS, is the Intel Extreme Tuning Utility (in the screenshot below) the solution to this issue? First I'll update the BIOS and test/relaunch the game.

But if the CTD happens again (likely because Murphy's Law says it must) then would adjusting the Performance Core Ration from 55x to 54x fix the problem? And what do the Efficient Core Ratio and Voltage Offset do? What values should these be tweaked to?

I'm concerned making these minor tweaks potentially have catastrophic effects on CPU voltage e.g. wreak havoc with the amount of voltage and/or shorten the lifetime of the 139000K CPU and/or other components in the system.

pressed_for_time · ‎01-04-2025

XTU can be more convenient than the BIOS for adjusting P-core speeds, you could use 53x to start with and if that works try 54x. The efficient core ratios are not a factor in this issue and you are observing the standard limits in any case. XTU and/or the BIOS will take care of the voltages. The latest BIOS will include Intel Default Settings which are a set of inter-related settings and I would not advocate manually altering any of these.

Whether you use the BIOS or XTU to adjust P-core frequency within the normal range it won't have any negative impact on voltage values and certainly none on the long term operation of the 13900K. Intel has extended the warranty on boxed 13900K CPUs to 5 years from the date of purchase as extra reassurance for those who have any concerns about the lifetime of their 13th/14th gen K/KF/KS CPUS.

PC1997 · ‎01-04-2025

Hello,

There is excellent advice posted in this thread and I agree. But it is too late for your CPU and you need to replace it. You MUST run the latest BIOS with actual correct Intel default settings (PL2 = 253 W, PL1 = 125 W & 307 amps) from day one or your new cpu WILL degrade, even with the latest microcode, if your motherboard is allowed to pull too many watts/amps through your CPU.

This is clear as night and day - non debatable, common knowledge, written in stone - that Raptor Lake will NOT tolerate mad electricity levels of current motherboard manufacturers have been pushing in the recent past (and current, depending on when you last updated your BIOS). And the problem is not what people think - 100°C max operating temperature is NOT the problem. It's at what voltage and how many amps / heat density is the issue. Or in other words, you can direct die cool the CPU and be well under 100c and still degrade the chip because of too much power draw! Obviously reading your post this is not you! But I needed to add this so people will read it and hopefully sticks in thier skulls....ALL of us!

To reiterate: nothing you can do short of replacing your CPU will permanently fix the problem. It is possible to tinker with and find out how much additional voltage... (vmin shift) yada yada yada that your chip needs to stabilize it, and with the correct bios aforementioned settings, prevent future degradation but existing damage will last for the life of that CPU.