Swap 13700k for 7800X3D

[Zach_01]

The iameatingjam person has a B760 board. I'm using a B660 temporarily. Under-voling isn't an option for may users.

I guess an option(s) in this situation is to to lower max multiplier (for single/low loads) and power limit (for all core loads).
Not let the boost go max and CPU requests for high voltage.

I think its one of the things Intel's microcode will do... to cut some boost in all situations.

 

[ShrimpBrime]

I guess an option(s) in this situation is to to lower max multiplier (for single/low loads) and power limit (for all core loads).
Not let the boost go max and CPU requests for high voltage.

I think its one of the things Intel's microcode will do... to cut some boost in all situations.

The problem people are having is a simple misunderstanding.

When most electronics have a MTBF at its max operating temp.

Like VRMs for example. Most will handle well over 100c. Some boards may throttle at 90c. Even if it ran at 90 or 100c, it would still have a MTBF of 100,000 hours.

Processors, also MTBF at its rated peak temp. 100c. Some boards allow 105c.

Running 100c shouldn't be an issue even for long periods of time. Please note I used the word "should", with not.

But I mean why buy 450$ processor and then cool it with 65$ AIO? Then expect top clocks and low temps???? This is the issue I see most. Just like the OP with a dark rock pro 4. People must have a hard time imagining what 253w really is an what it takes to move that kind of BTU around.

253w = 863BTU/hr.

 

[Zach_01]

Personal opinion
Zero blame to any user.
The CPU should have protection mechanisms to sustain its self even if it falls on the uneducated hands.
Called fool proofing.
And I'm not talking for manual OC, but auto boost.

Running at 100C shouldn't be an issue under specific conditions. What are those?
Dont know but certainly not with max current and boost at play. Its just common sense to me at least.
Thats what AMD is doing and thats what the TVB was suppose to do, but on many boards its a bunch of 0s instead of values that would decrease boost progressively as temp goes up and closer to the limit.
Its right on the name ThermalVelocityBoost.

And what about those PL1/PL2 limits...
Yeah, lets put them at 253W and 4096W. And the (used to be) 125W long duration power limit has become a distant past.
Who cares? We are winning most of the competition!
200... 300Amps! ...shove it in, it can take it
Well apparently it cant

Yes board vendors have their share on this but Intel was allowing it in the first place.

And its even worse than the above.
Desktop CPUs that are used on servers that need high single boost clocks are falling like flies. Those are nowhere near the power or temp limit.

Its simple. Intel wanted to stay in the "game" of performance and now its biting ass time, unfortunately for many users too that spend their money.

I used to be impressed by the robustness of Intel chips. Not anymore.
Intel made them a house of cards.

Apparently 12th gen was the upper limit for the specific fabrication process. Already overclocked to the roof.

 

[iameatingjam]

The iameatingjam person has a B760 board. I'm using a B660 temporarily. Under-voling isn't an option for may users.

Also OP said even small offsets were causing crashes.

Btw, I actually can undervolt but I gotta downgrade the microcode, its an option in my bios
we talked about this on reddit one time
You didn't believe me
I had to post screenshots lol

But yeah no I am even more hesitant to go back to old ucodes now.

 

[Vario]

Personal opinion
Zero blame to any user.
The CPU should have protection mechanisms to sustain its self even if it falls on the uneducated hands.
Called fool proofing.
And I'm not talking for manual OC, but auto boost.

Running at 100C shouldn't be an issue under specific conditions. What are those?
Dont know but certainly not with max current and boost at play. Its just common sense to me at least.
Thats what AMD is doing and thats what the TVB was suppose to do, but on many boards its a bunch of 0s instead of values that would decrease boost progressively as temp goes up and closer to the limit.
Its right on the name ThermalVelocityBoost.

And what about those PL1/PL2 limits...
Yeah, lets put them at 253W and 4096W. And the (used to be) 125W long duration power limit has become a distant past.
Who cares? We are winning most of the competition!
200... 300Amps! ...shove it in, it can take it
Well apparently it cant

Yes board vendors have their share on this but Intel was allowing it in the first place.

And its even worse than the above.
Desktop CPUs that are used on servers that need high single boost clocks are falling like flies. Those are nowhere near the power or temp limit.

Its simple. Intel wanted to stay in the "game" of performance and now its biting ass time, unfortunately for many users too that spend their money.

I used to be impressed by the robustness of Intel chips. Not anymore.
Intel made them a house of cards.

Apparently 12th gen was the upper limit for the specific fabrication process. Already overclocked to the roof.

That is not a correct analysis, the problem is microcode applying too much voltage and board vendors using aggressive default settings. The "Icarus flew too close to the sun" bullshit does not apply.

There's probably transient voltage spikes that are not detectable in monitoring software which periodically overvolt the snot out of the CPU and result in the degradation over time.

 

[ShrimpBrime]

Also OP said even small offsets were causing crashes.

Btw, I actually can undervolt but I gotta downgrade the microcode, its an option in my bios
we talked about this on reddit one time
You didn't believe me
I had to post screenshots lol

But yeah no I am even more hesitant to go back to old ucodes now.

With a loss of some performance.

And these processors don't seem to like an under-volt when running hot.

Because leakage happens from heat. When leakage is high, it costs more v-core to remain stable. The cooler they run, the less leakage, less v-core needed.

I remember. And it was like a year ago (+/-)I think you showed then screen shots. ME has had a few changes since then.

I can't change micro-code with 12400F. But with the 14700K I can. It doesn't give much under-volt anyways, yes stability will be an issue then.

 

[Bill_Bright]

Dark Rock pro 4 is just not an adequate cooler.

Sure it is. I note it received a TPU "Editor's Choice" award with comments like,

Exceptional performance,
upper echelons of available air-cooling options
Top tier build quality

be quiet! Dark Rock Pro 4 Review: Return to the Dark Side - Tom's Hardware "Great cooling ability"

be quiet! Dark Rock Pro 4 CPU Cooler Review: Silent & Powerful - Tech4Gamers - "The be quiet! Dark Rock Pro 4 is a fantastic CPU cooler built for pure performance." "Phenomenal cooling performance."

No, it is not the latest and greatest but it definitely is more than adequate for that CPU in a properly cooled case.

No CPU should require alternative (water) cooling solutions and to the best of my knowledge, neither Intel nor AMD require water cooling for any of their CPUs, particularly at stock clocks and voltages.

Still with all the drama going on with intel right now I don't blame the OP for being concerned

With or without any drama, 100°C IS concerning. There is a problem that needs to be addressed. We just don't know what the problems is. Any one suggesting this or that is just guessing.

@warriorunited - I think at this point, considering there is a 3-year warranty on that processor, it would not hurt to contact Intel support and see what they say. Present all the steps you have taken to resolve the issue. Point out you've tried different cases, different PSUs (if true), etc. Include your current case model and case cooling configuration, as well as your cooler model and mention your ambient (room) temps are controlled (assuming that is true too). Include motherboard model and BIOS version(s) you have used. Make sure to say you have all settings set to the out-of-the-box defaults.

BE NICE!

The worse they can do is say, no.

And you have Intel.

I have both. My main, this, computer, the one listed in my System Specs has Intel but 2 other computers in this house have AMD - by choice (all PCs here are self-built). I made it clear, with emphasis no less, I was not defending Intel but you decided not to include the context of my statement in your reply/quote.

Everyone has one of two vendors and therefore there is no unbiased opinions possible here

"No unbiased opinions possible"? Come on! That is just insulting nonsense! Now you are trying to tell us that no one here is capable of being unbiased!!! Wow! What? Are we all bumbling imbeciles in your mind? Incapable of being objective?

So, according to evernessince, if we own Brand A, there is no way we cannot be unbiased about Brand A. Perhaps you, evernesssice, have such biased tendencies but don't pretend to know, or speak for the rest of us here at TPU, or the rest of humanity, for that matter.

FOR SURE, there are some who are so set in their ways and beliefs, that they refuse (or are incapable) to accept any other possibility could ever exist. But to claim no unbiased opinions are even possible is just plain asinine.

 

[iameatingjam]

With a loss of some performance.

And these processors don't seem to like an under-volt when running hot.

Because leakage happens from heat. When leakage is high, it costs more v-core to remain stable. The cooler they run, the less leakage, less v-core needed.

I remember. And it was like a year ago (+/-)I think you showed then screen shots. ME has had a few changes since then.

I can't change micro-code with 12400F. But with the 14700K I can. It doesn't give much under-volt anyways, yes stability will be an issue then.

Well it gained performance for me, I proved it with screenshots.

https://www.reddit.com/r/overclocking/comments/18innhj/_/kdkxkta
The screenshots are still there too. They made you finally believe me in the end.

But I'm not touching 104 with a stick anymore. And not just because thats what my first 14700kf was on when it died. That gamer's nexus video released the other day was talking about some issues it had.

Besides... whats it all for in the end... benchmarks? Not like just underclocking a few hundred mhz is gonna effect general computing and 4k gaming.

 

[ShrimpBrime]

Sure it is. I note it received a TPU "Editor's Choice" award with comments like,


be quiet! Dark Rock Pro 4 Review: Return to the Dark Side - Tom's Hardware "Great cooling ability"

be quiet! Dark Rock Pro 4 CPU Cooler Review: Silent & Powerful - Tech4Gamers - "The be quiet! Dark Rock Pro 4 is a fantastic CPU cooler built for pure performance." "Phenomenal cooling performance."

No, it is not the latest and greatest but it definitely is more than adequate for that CPU in a properly cooled case.

No CPU should require alternative (water) cooling solutions and to the best of my knowledge, neither Intel nor AMD require water cooling for any of their CPUs, particularly at stock clocks and voltages.


With or without any drama, 100°C IS concerning. There is a problem that needs to be addressed. We just don't know what the problems is. Any one suggesting this or that is just guessing.

@warriorunited - I think at this point, considering there is a 3-year warranty on that processor, it would not hurt to contact Intel support and see what they say. Present all the steps you have taken to resolve the issue. Point out you've tried different cases, different PSUs (if true), etc. Include your current case model and case cooling configuration, as well as your cooler model and mention your ambient (room) temps are controlled (assuming that is true too). Include motherboard model and BIOS version(s) you have used. Make sure to say you have all settings set to the out-of-the-box defaults.

BE NICE!

The worse they can do is say, no.


I have both. My main, this, computer, the one listed in my System Specs has Intel but 2 other computers in this house have AMD - by choice (all PCs here are self-built). I made it clear, with emphasis no less, I was not defending Intel but you decided not to include the context of my statement in your reply/quote.


"No unbiased opinions possible"? Come on! That is just insulting nonsense! Now you are trying to tell us that no one here is capable of being unbiased!!! Wow! What? Are we all bumbling imbeciles in your mind? Incapable of being objective?

So, according to evernessince, if we own Brand A, there is no way we cannot be unbiased about Brand A. Perhaps you, evernesssice, have such biased tendencies but don't pretend to know, or speak for the rest of us here at TPU, or the rest of humanity, for that matter.

FOR SURE, there are some who are so set in their ways and beliefs, that they refuse (or are incapable) to accept any other possibility could ever exist. But to claim no unbiased opinions are even possible is just plain asinine.

It's not an adequate cooler for a 13700 or 14700K. I don't care who did a review.

I mean if a company paid me and gave me free hardware, then

Yes! Dark Rock Pro 4 is awesome and handles 300w easily and keep temps below 60c!!!

Nope, I know better. That's why they don't advertise wattage figures with most coolers.

 

[Bill_Bright]

I mean if a company paid me and gave me free hardware

Oh come on! So now TPU's reviews are all biased therefore everyone should discard their findings and just believe ShrimpBrime instead.

Yeah right.

 

[iameatingjam]

It's not an adequate cooler for a 13700 or 14700K. I don't care who did a review.

I mean if a company paid me and gave me free hardware, then

Yes! Dark Rock Pro 4 is awesome and handles 300w easily and keep temps below 60c!!!

Nope, I know better. That's why they don't advertise wattage figures with most coolers.

Wait, am I missing something? Wasn't OPs temps okay before a bios update? What if it flipped some switch, like CEP or one of the loadlines thats messing things up.

 

[ShrimpBrime]

Wait, am I missing something? Wasn't OPs temps okay before a bios update? What if it flipped some switch, like CEP or one of the loadlines thats messing things up.

He always had a cooling issue. I'm guessing. Before and after a bios update, the cpu still boosts 253w. Just how it is.

Cannot effectively cool 860 BTU/hr with most air coolers.

They also are not effective for cooling VRM package area either which could also cause throttling of the cpu.

Oh come on! So now TPU's reviews are all biased therefore everyone should discard their findings and just believe ShrimpBrime instead.

Yeah right.

I didn't ask people to believe me. That's you putting words in my mouth.

Belief is up to you and others.

1 review is not indicative of other user experiences. End user results.

 

[Bill_Bright]

I didn't ask people to believe me. That's you putting words in my mouth.

You are right. You didn't ask. But I didn't say you did either. So who's putting words into other's mouths?

That said, one has to wonder why anyone would publicly imply such deceptive tactics if they didn't want people to believe them?

 

[sneekypeet]

While I realize the power setting is set high, has anyone verified the cooler had its base sticker removed?

 

[Zach_01]

That is not a correct analysis, the problem is microcode applying too much voltage and board vendors using aggressive default settings. The "Icarus flew too close to the sun" bullshit does not apply.

There's probably transient voltage spikes that are not detectable in monitoring software which periodically overvolt the snot out of the CPU and result in the degradation over time.

Yeah sure its transients and not actual requests...
I would think this again after seeing someone using oscilloscope and doing actual measurements.

I guess you missed that on a previous post:

 

[AnotherReader]

Even in TechPowerUp's review of the 13700K, rendering in Blender resulted in a 102 C temperature at stock. Playing CyberPunk 2077, on the other hand, was much easier on the CPU: 73 C temperature at stock. Reducing the power limit should be the easiest way to reduce temperatures in heavily multithreaded applications. Computerbase found that reducing power draw to 142 W from the default 253 W increased render times in Blender by 11%. It's up to you to decide if that's too much.

 

[ShrimpBrime]

You are right. You didn't ask. But I didn't say you did either. So who's putting words into other's mouths?

That said, one has to wonder why anyone would publicly imply such deceptive tactics if they didn't want people to believe them?

Here's how you move 860 btu/hr.

370 grams of copper.

 

[Bill_Bright]

While I realize the power setting is set high, has anyone verified the cooler had its base sticker removed?

I wondered that too way back when I asked about idle temps in post #2. Not sure we got a direct answer but I did see some screenshots showing "min" temps in the mid 30s.

 

[Bobaganoosh]

I wondered that too way back when I asked about idle temps in post #2. Not sure we got a direct answer but I did see some screenshots showing "min" temps in the mid 30s.

They did say they've used it on multiple systems at this point and the sticker was removed years ago.

 

[warriorunited]

Wow, so many answers since yesterday and I don't even know where to start going to follow for now the main suggestion I took notes:

Ok, so first, please use HWInfo64 instead of HWMonitor so we have accurate readings.

Edit, NVM, just saw you switched already

You can find Load Line Calibration (LLC) in the Digi VRM tab. Should be above the v-core section.

I have two options here:

- CPU Input Voltage Load-Line Calibration - Values go from Level 1 to 3
- CPU Load-Line Calibration - Goes from level Auto to Level 7.

Wich of this options are you guys recommending to try Level 2?

Thanks!

 

[ShrimpBrime]

Wow, so many answers since yesterday and I don't even know where to start going to follow for now the main suggestion I took notes:



I have two options here:

- CPU Input Voltage Load-Line Calibration - Values go from Level 1 to 3
- CPU Load-Line Calibration - Goes from level Auto to Level 7.

Wich of this options are you guys recommending to try Level 2?

Thanks!

Cpu loadline calibration would be the one to try. Leave input voltage on auto, you're not doing FSB overclocking anyways.

 

[Solaris17]

@Bill_Bright @ShrimpBrime

your both pretty.

Now please just help OP or take it to PM. I don’t feel like sifting through the reports right now or this thread.

Thanks.

 

[ShrimpBrime]

Wow, so many answers since yesterday and I don't even know where to start going to follow for now the main suggestion I took notes:



I have two options here:

- CPU Input Voltage Load-Line Calibration - Values go from Level 1 to 3
- CPU Load-Line Calibration - Goes from level Auto to Level 7.

Wich of this options are you guys recommending to try Level 2?

Thanks!

Let us know if the suggestions helped!

Can always try lvl 1 as well.

 

[warriorunited]

Hello Everyone again!

So i tried today's new beta BIOS Microcode update just to see if something would change, and now it's even worse lol it reaches 100º way more faster like in 4/5 seconds, and the only change i did was activate XMP I and set my fans to turbo speed, nothing else, all the rest is in default/intel settings.

 

[AnotherReader]

Hello Everyone again!

So i tried today's new beta BIOS Microcode update just to see if something would change, and now it's even worse lol it reaches 100º way more faster like in 4/5 seconds, and the only change i did was activate XMP I and set my fans to turbo speed, nothing else, all the rest is in default/intel settings.

View attachment 358008

I think it's the default high power limit though temperatures are still higher than expected. In the recent 9700X review, the 13700K is reaching 86 C in Blender with a Noctua D15. What's the room temperature in the area where the desktop is located?