Intel 14th/13th gen, latency/stability issues?
-
- KVRAF
- 3222 posts since 23 Dec, 2002
Although it is early days, Youtubers are starting to cover the microcode updates. If your CPU was damaged it won't fix it but if you CPU wasn't compromised it seems hopeful that the microcode / bios updates will present a solution with minimal performance hit; at least as measured by CPU intensive gaming benchmarks. The biggest issue for the those who have the processor is deciding if a random crash is degradation or something more benign? Do you RMA it or take a chance? How do you have confidence in the processor? You'd be foolish to buy an affected CPU on the used market. Your CPU is now devalued if you want to sell it for whatever purpose.
- KVRAF
- Topic Starter
- 4801 posts since 1 Aug, 2005 from Warszawa, Poland
Depends on the price. If you can get 14900k dirt cheap, even if it's degraded somewhat, you should be able to run it fine on raised voltage, you only need to figure a way to prevent it from further degradation: sync cores or limit VID. Assuming you have time to be bothered with it.
-
- KVRAF
- 3222 posts since 23 Dec, 2002
I suppose it depends on how the CPU is degraded. Some reports show overall instability even at significantly reduced voltage levels after the damage has taken place. Perhaps you could disable a core or two if that were the case. If the price was right and you are a tinkerer it might be worth the time. A disaster for some. An opportunity for someone else. I think there are fewer tinkerers out there though.
Zombie Queen wrote: Fri Aug 09, 2024 12:35 pmDepends on the price. If you can get 14900k dirt cheap, even if it's degraded somewhat, you should be able to run it fine on raised voltage, you only need to figure a way to prevent it from further degradation: sync cores or limit VID. Assuming you have time to be bothered with it.
- KVRAF
- Topic Starter
- 4801 posts since 1 Aug, 2005 from Warszawa, Poland
That's how it works, degraded CPU becomes unstable at lover voltage, if you rise the voltage it may run fine, until it degrades more. With higher voltage there's a better chance of further degradation, but if you can stop it from going dangerously high, it can work. Like prevent single core high boosts, possibly reduce frequency a bit. Something for tinkerers.Scotty wrote: Fri Aug 09, 2024 1:13 pm Some reports show overall instability even at reduced voltage levels.
-
- KVRAF
- 7886 posts since 24 Feb, 2003 from Earth, USA
What Kontakt instrument in question gets crackles? Something from Komplete? I haven't heavily trying to use Kontakt 7 yet on mine, but wanted to test on my i7 14700k.Zombie Queen wrote: Sun Jun 23, 2024 12:29 pm Finally official numbers for 14700k, took them a whole to write those.
Meanwhile I had my first bluescreen. W10 gets stuck on welcome screen out of the blue, what an abomination of a system.
And if I run a really heavy Kontakt instrument I get crackles. It feels less and less an upgrade.
Devon
Simple music philosophy - Those who can, make music. Those who can't, make excuses.
Read my VST reviews at Traxmusic!
Read my VST reviews at Traxmusic!
- KVRAF
- Topic Starter
- 4801 posts since 1 Aug, 2005 from Warszawa, Poland
One that is not released yet. And it turned out to be Bitwig vs Kontakt thing, I had to configure it differently, that is let Kontakt manage multithreading, or something like this. Crackles are gone as long as I run "high performance power plan" + Kontakt standalone is fine in any circumstances. Kontak 6 it is, I haven't upgraded.DevonB wrote: Fri Aug 09, 2024 7:13 pm What Kontakt instrument in question gets crackles? Something from Komplete?
Also I didn't get any crashes since. Had to tweak memory settings.
- KVRAF
- Topic Starter
- 4801 posts since 1 Aug, 2005 from Warszawa, Poland
So this is what Intel considers the safe limit.Intel wrote:The latest microcode update (0x129) will limit voltage requests above 1.55V
- KVRAF
- Topic Starter
- 4801 posts since 1 Aug, 2005 from Warszawa, Poland
First reports I've seen indicate that the update still allows high voltages, just for shorter spikes. And on the other hand it forcefully downclocks when idle... Thinking it may be looking like forced "balanced power plan". I may be a good idea not to rush into this update.
-
- KVRAF
- 7886 posts since 24 Feb, 2003 from Earth, USA
I applied it to two machines. No problems so far. Gigabyte Aorus Elite X AX on F6f.Zombie Queen wrote: Sat Aug 10, 2024 7:09 pm First reports I've seen indicate that the update still allows high voltages, just for shorter spikes. And on the other hand it forcefully downclocks when idle... Thinking it may be looking like forced "balanced power plan". I may be a good idea not to rush into this update.
Devon
Simple music philosophy - Those who can, make music. Those who can't, make excuses.
Read my VST reviews at Traxmusic!
Read my VST reviews at Traxmusic!
- KVRAF
- Topic Starter
- 4801 posts since 1 Aug, 2005 from Warszawa, Poland
- KVRAF
- Topic Starter
- 4801 posts since 1 Aug, 2005 from Warszawa, Poland
Here's todays note. Just saw an ad, local shops offer 14700k at half the price I paid 3 months ago.
- KVRian
- 1188 posts since 21 Aug, 2017 from Brasil
YOU DON'T HAVE TO TURN OFF CEP to undervolt intel 13/14th gen CPUs
-
- KVRAF
- 3222 posts since 23 Dec, 2002
Given the price reduction, I think your tinker's hypothesis has some serious traction.
Zombie Queen wrote: Wed Aug 14, 2024 2:43 pm Here's todays note. Just saw an ad, local shops offer 14700k at half the price I paid 3 months ago.
-
- KVRAF
- 4345 posts since 20 Feb, 2004
I have some questions for the PC experts related to this over-voltaging issue.
- Is the over-voltaging problem equivalent to what would happen if you overclocked your CPU?
- Does the problem with the over-voltaging stem entirely from the increased CPU temperatures this brings? So if you have adequate/excellent CPU cooling, the over-voltaging wouldn't be a problem?
- Is there a good way to test your CPU to see if it's already been damaged by the over-voltaging issue?
- Is the microcode update a motherboard update? I.e., is it something you'd flash onto your motherboard?
- Is the over-voltaging problem equivalent to what would happen if you overclocked your CPU?
- Does the problem with the over-voltaging stem entirely from the increased CPU temperatures this brings? So if you have adequate/excellent CPU cooling, the over-voltaging wouldn't be a problem?
- Is there a good way to test your CPU to see if it's already been damaged by the over-voltaging issue?
- Is the microcode update a motherboard update? I.e., is it something you'd flash onto your motherboard?
A well-behaved signature.
- KVRAF
- Topic Starter
- 4801 posts since 1 Aug, 2005 from Warszawa, Poland
Not sure I'm an expert, but I've been following this closely. here's my understanding...
If you have Gigabyte or ASUS board you should have option to manually put a voltage limit, even without the microcode update.
No. The problem is CPU requesting too much voltage and possibly voltage provider (motherboard) not able to keep up with different voltage requests resulting in voltage overshoots, which destroy CPU. Heat may accelerate the degradation, but these overshoots actually rather won't happen under full load, as everything will slow down to escape heat... more likely would happen if one core is doing something at full speed and rest is idle.JerGoertz wrote: Mon Aug 19, 2024 11:58 pm Does the problem with the over-voltaging stem entirely from the increased CPU temperatures this brings? So if you have adequate/excellent CPU cooling, the over-voltaging wouldn't be a problem?
Yes. However the voltage limit is still quite high, there's speculation than it is too high and it is middle ground, so the CPU won't die too fast. While lower limit would cripple performance of 14900 - read possible legal trouble.JerGoertz wrote: Mon Aug 19, 2024 11:58 pm Is the microcode update a motherboard update? I.e., is it something you'd flash onto your motherboard?
If you have Gigabyte or ASUS board you should have option to manually put a voltage limit, even without the microcode update.
