F5F Stay Refreshed Hardware Desktop Issues with 14900ks in Intel configurations.

Issues with 14900ks in Intel configurations.

Issues with 14900ks in Intel configurations.

Pages (2): 1 2 Next
S
Spiderb0y_
Junior Member
12
10-21-2023, 11:17 AM
#1
I am currently experiencing instability after 14900ks, even when using Intel's suggested configurations. The problem appears to occur when both hyperthreads on the same core are pushed to high clock speeds. This issue is especially noticeable with integer tasks, particularly during code compilation or file compression. Under multi-core conditions, power constraints prevent the cores from reaching sufficient speeds to trigger the problem—testing Cinebench on a single core doesn’t cause it because two threads are needed on the same core, and multi-core testing avoids it due to power limits. It’s unclear if Cinebench generates the appropriate workload for the CPU. Setting the power limit to 253W as recommended by Intel seems insufficient to resolve the issue. I’m curious if this affects all 14900ks or even higher models. My test uses Gentoo Linux, compiling with gcc and running tasks with both threads on the same preferred p-core. So far, it consistently fails within seconds on both machines I’ve tried. Recreating this scenario on Windows is challenging because Thread Director limits each core to one hyperthread before switching, which restricts performance due to power restrictions. You might be able to reproduce the fault by performing a large compilation or decompression with CPU affinity set to both threads on the same core. I’d appreciate any insights from users who own 14900ks or later, and whether this issue impacts other 14th-generation CPUs.
S
Spiderb0y_
10-21-2023, 11:17 AM #1

I am currently experiencing instability after 14900ks, even when using Intel's suggested configurations. The problem appears to occur when both hyperthreads on the same core are pushed to high clock speeds. This issue is especially noticeable with integer tasks, particularly during code compilation or file compression. Under multi-core conditions, power constraints prevent the cores from reaching sufficient speeds to trigger the problem—testing Cinebench on a single core doesn’t cause it because two threads are needed on the same core, and multi-core testing avoids it due to power limits. It’s unclear if Cinebench generates the appropriate workload for the CPU. Setting the power limit to 253W as recommended by Intel seems insufficient to resolve the issue. I’m curious if this affects all 14900ks or even higher models. My test uses Gentoo Linux, compiling with gcc and running tasks with both threads on the same preferred p-core. So far, it consistently fails within seconds on both machines I’ve tried. Recreating this scenario on Windows is challenging because Thread Director limits each core to one hyperthread before switching, which restricts performance due to power restrictions. You might be able to reproduce the fault by performing a large compilation or decompression with CPU affinity set to both threads on the same core. I’d appreciate any insights from users who own 14900ks or later, and whether this issue impacts other 14th-generation CPUs.

B
beasthockey13
Junior Member
40
10-24-2023, 06:30 AM
#2
This topic connects to an investigation about a game crash linked to the Intel 13900K and Fortnite using Unreal Engine.
B
beasthockey13
10-24-2023, 06:30 AM #2

This topic connects to an investigation about a game crash linked to the Intel 13900K and Fortnite using Unreal Engine.

H
hahahalo21
Junior Member
46
10-29-2023, 12:55 PM
#3
It might be connected. Adjusting the thread affinity to share the same preferred p-core appears to cause consistent failures, allowing you to quickly assess if a CPU has the issue. The data also indicates that adhering to Intel power limits doesn<|pad|>, as mentioned earlier, doesn’t eliminate the problem—it only lessens how often it happens. With a straightforward, repeatable test, you should be able to determine whether all 14900ks share the same issue or if it affects only a small portion.
H
hahahalo21
10-29-2023, 12:55 PM #3

It might be connected. Adjusting the thread affinity to share the same preferred p-core appears to cause consistent failures, allowing you to quickly assess if a CPU has the issue. The data also indicates that adhering to Intel power limits doesn<|pad|>, as mentioned earlier, doesn’t eliminate the problem—it only lessens how often it happens. With a straightforward, repeatable test, you should be able to determine whether all 14900ks share the same issue or if it affects only a small portion.

A
Avichi
Member
73
10-31-2023, 12:21 AM
#4
I discovered this video regarding your subject, and it stands out as a helpful explanation for your issue.
A
Avichi
10-31-2023, 12:21 AM #4

I discovered this video regarding your subject, and it stands out as a helpful explanation for your issue.

Z
Zynxx
Member
61
10-31-2023, 01:04 AM
#5
I already have PL1=253W PL2=253W and ICCMAX=307A, so I used a lower current limit than shown in the video. Applying power limits aids under all core loads since that 253W is divided among 8 p-cores. But if you run a two-thread task on one p-core, most of that 253W becomes available to just that core handling two threads. In short, limiting power doesn<|pad|> can help but doesn’t solve the problem on its own—it just cuts down failure chances. Solutions that work: - turning off hyper-threading, though for multi-core tasks like this it’s like removing half the chip—less ideal. - reducing the core multiplier for all cores, which sacrifices single-thread speed. It looks like hyper-threading and high boost speeds on p-cores are a CPU concern, not a motherboard one, especially with the two 14900K models I own. I’d like to know if this applies to all 14900K models.
Z
Zynxx
10-31-2023, 01:04 AM #5

I already have PL1=253W PL2=253W and ICCMAX=307A, so I used a lower current limit than shown in the video. Applying power limits aids under all core loads since that 253W is divided among 8 p-cores. But if you run a two-thread task on one p-core, most of that 253W becomes available to just that core handling two threads. In short, limiting power doesn<|pad|> can help but doesn’t solve the problem on its own—it just cuts down failure chances. Solutions that work: - turning off hyper-threading, though for multi-core tasks like this it’s like removing half the chip—less ideal. - reducing the core multiplier for all cores, which sacrifices single-thread speed. It looks like hyper-threading and high boost speeds on p-cores are a CPU concern, not a motherboard one, especially with the two 14900K models I own. I’d like to know if this applies to all 14900K models.

D
DantehIsGay
Posting Freak
902
11-09-2023, 08:38 AM
#6
Mention his current earnings around 14900ks, noting it's an older video. You should likely receive a response.
D
DantehIsGay
11-09-2023, 08:38 AM #6

Mention his current earnings around 14900ks, noting it's an older video. You should likely receive a response.

C
ChappieGames1
Member
123
11-14-2023, 05:36 AM
#7
C
ChappieGames1
11-14-2023, 05:36 AM #7

B
banshee45
Senior Member
726
11-14-2023, 09:15 AM
#8
BTW, PL1 delivers 150w on ks and 125w on k. Not a 253w mix-up—no article copied from 3dguru.
B
banshee45
11-14-2023, 09:15 AM #8

BTW, PL1 delivers 150w on ks and 125w on k. Not a 253w mix-up—no article copied from 3dguru.

W
Wither01
Member
201
11-14-2023, 05:03 PM
#9
The power restrictions mentioned seem to be misleading. There isn't a fixed limit per core, and performance drops when multiple threads are active on the same core. Testing with 150W limits still caused issues, especially when hyper-threading is on. The main concern appears to be when a core handles heavy workloads while others sit idle. Reducing power limits below certain thresholds might fix the problem, but it would significantly hurt multi-core efficiency. It’s better to cap overall core frequency rather than per-core limits.
W
Wither01
11-14-2023, 05:03 PM #9

The power restrictions mentioned seem to be misleading. There isn't a fixed limit per core, and performance drops when multiple threads are active on the same core. Testing with 150W limits still caused issues, especially when hyper-threading is on. The main concern appears to be when a core handles heavy workloads while others sit idle. Reducing power limits below certain thresholds might fix the problem, but it would significantly hurt multi-core efficiency. It’s better to cap overall core frequency rather than per-core limits.

L
Lxxn2002
Member
240
11-16-2023, 04:55 AM
#10
@Robchil Here's a link to Intel's actual specification for the 14900ks, rather than relying on third party sources: https://www.intel.com/content/www/us/en/...tions.html Quoting from this the maximum turbo power is 253W [The maximum sustained (>1s) power dissipation of the processor as limited by current and/or temperature controls. Instantaneous power may exceed Maximum Turbo Power for short durations (<=10ms). Note: Maximum Turbo Power is configurable by system vendor and can be system specific] Note: 150W is the TDP, that is with certain intel defined settings the CPU is guaranteed not to put out more than this amount of power, it's nothing to do with power limits.
L
Lxxn2002
11-16-2023, 04:55 AM #10

@Robchil Here's a link to Intel's actual specification for the 14900ks, rather than relying on third party sources: https://www.intel.com/content/www/us/en/...tions.html Quoting from this the maximum turbo power is 253W [The maximum sustained (>1s) power dissipation of the processor as limited by current and/or temperature controls. Instantaneous power may exceed Maximum Turbo Power for short durations (<=10ms). Note: Maximum Turbo Power is configurable by system vendor and can be system specific] Note: 150W is the TDP, that is with certain intel defined settings the CPU is guaranteed not to put out more than this amount of power, it's nothing to do with power limits.

Pages (2): 1 2 Next