F5F Stay Refreshed Hardware Desktop Exploring unstable performance testing for Intel CPUs.

Exploring unstable performance testing for Intel CPUs.

Exploring unstable performance testing for Intel CPUs.

J
JEFF_JEFFERSON
Senior Member
627
12-01-2023, 04:28 AM
#1
I've been using a 14700K since launch. My CPU has remained stable with BIOS updates and only the 253W limit on my pl1 and 2. I'm trying to see if anything changed. I installed the latest NVIDIA drivers and everything seems normal. Any other suggestions?
J
JEFF_JEFFERSON
12-01-2023, 04:28 AM #1

I've been using a 14700K since launch. My CPU has remained stable with BIOS updates and only the 253W limit on my pl1 and 2. I'm trying to see if anything changed. I installed the latest NVIDIA drivers and everything seems normal. Any other suggestions?

Z
ZombieBoss7
Member
77
12-08-2023, 06:36 AM
#2
Run UE5 projects as if the engine is relatively light, even though certain steps can cause performance issues on weaker CPUs.
Z
ZombieBoss7
12-08-2023, 06:36 AM #2

Run UE5 projects as if the engine is relatively light, even though certain steps can cause performance issues on weaker CPUs.

P
Prawnflakes
Member
204
12-08-2023, 08:23 AM
#3
The situation doesn't involve any "affected" condition. It was mainly a rapid decline in performance over time. All electronic parts and semiconductors lose effectiveness gradually. Ten years back, pushing your processor beyond its limits and then trying to push it again today would likely result in failure due to wear. Still, it can operate within its intended capabilities without breaking the micro-architecture. The recent crashes linked to Intel stemmed from a swift deterioration caused by excessive voltage demands. It seems only certain processors were impacted, possibly due to manufacturing variations or oxidation problems. I believe these chips are built to endure heavy use, but testing showed that high voltage requests triggered microcode errors. Many users didn't encounter issues because of three factors: 1) Variety in silicon quality – some chips handled the stress better; 2) Light workloads last longer; 3) Using default power settings reduces risk. Adjusting voltages and limiting power can help, and setting a core voltage cap is especially effective. Running under balanced power profiles works well, though Windows' approach isn't ideal. If everything runs smoothly, your processor should be fine.
P
Prawnflakes
12-08-2023, 08:23 AM #3

The situation doesn't involve any "affected" condition. It was mainly a rapid decline in performance over time. All electronic parts and semiconductors lose effectiveness gradually. Ten years back, pushing your processor beyond its limits and then trying to push it again today would likely result in failure due to wear. Still, it can operate within its intended capabilities without breaking the micro-architecture. The recent crashes linked to Intel stemmed from a swift deterioration caused by excessive voltage demands. It seems only certain processors were impacted, possibly due to manufacturing variations or oxidation problems. I believe these chips are built to endure heavy use, but testing showed that high voltage requests triggered microcode errors. Many users didn't encounter issues because of three factors: 1) Variety in silicon quality – some chips handled the stress better; 2) Light workloads last longer; 3) Using default power settings reduces risk. Adjusting voltages and limiting power can help, and setting a core voltage cap is especially effective. Running under balanced power profiles works well, though Windows' approach isn't ideal. If everything runs smoothly, your processor should be fine.

B
bmw355games
Junior Member
47
12-10-2023, 06:15 AM
#4
Observe closely what is happening.
B
bmw355games
12-10-2023, 06:15 AM #4

Observe closely what is happening.

C
crosser146
Member
63
12-13-2023, 11:37 AM
#5
Extreme overclocking setup was diverse for CPU stress testing at 95% load. You can try a burn test if you want, just like in Jays 2 cents. For real-world use, run rendering tasks, install Tlou, and let shaders compile for about half an hour.
C
crosser146
12-13-2023, 11:37 AM #5

Extreme overclocking setup was diverse for CPU stress testing at 95% load. You can try a burn test if you want, just like in Jays 2 cents. For real-world use, run rendering tasks, install Tlou, and let shaders compile for about half an hour.

L
LissieBear
Member
163
12-13-2023, 06:45 PM
#6
I tried the Jayz video with CB r15/ R23, Prime95 small ffts, OCCT, Black Myth wukong benchmark tool, and Intel Burn Test. Based on my experience, primt95 didn’t assist much in diagnosing issues. The most useful tools for checking stability were OCCT, Wukong benchmark, Intel Burn Test, and launching a messy Minecraft setup. I ran Wukong at full settings, cinematic quality, full RT, top DLSS, and monitored for crashes or BSODs. If no soft crash occurred, my CPU settings seemed fine. In OCCT’s CPU stability tester, I used Normal Mode, Variable Load, Auto Instructions, cycling through all cores for two minutes each. During the Intel Burn Test, I followed Jay’s instructions—100 cycles at normal load. If everything was stable, I expected a game to run smoothly, but starting Minecraft caused a memory access violation and BSOD shortly afterward. My Intel 14900K ran at 1.39v VCORE, with -2 P core turbo ratio offset, and memory was underclocked to DDR5-5800 CL36.
L
LissieBear
12-13-2023, 06:45 PM #6

I tried the Jayz video with CB r15/ R23, Prime95 small ffts, OCCT, Black Myth wukong benchmark tool, and Intel Burn Test. Based on my experience, primt95 didn’t assist much in diagnosing issues. The most useful tools for checking stability were OCCT, Wukong benchmark, Intel Burn Test, and launching a messy Minecraft setup. I ran Wukong at full settings, cinematic quality, full RT, top DLSS, and monitored for crashes or BSODs. If no soft crash occurred, my CPU settings seemed fine. In OCCT’s CPU stability tester, I used Normal Mode, Variable Load, Auto Instructions, cycling through all cores for two minutes each. During the Intel Burn Test, I followed Jay’s instructions—100 cycles at normal load. If everything was stable, I expected a game to run smoothly, but starting Minecraft caused a memory access violation and BSOD shortly afterward. My Intel 14900K ran at 1.39v VCORE, with -2 P core turbo ratio offset, and memory was underclocked to DDR5-5800 CL36.