intel 13th and 14th gen xx900 and xx700 may have defective cores causing crashes in gaming

GoldenTiger said:
It was, if I recall correctly, water cooling, nothing exotic :).
Click to expand...

You might be right, my memory is very hazy, I just remember people criticizing it afterwards, thought it was some sort of custom refrigerator loop config in my head.


StoleMyOwnCar said:
Hope it doesn't put them too far under to keep competing with AMD
Click to expand...

Actually upon some further consideration I would like to retract this statement. Intel needs to fall behind for a gen or two and give AMD a near monopoly in the CPU space, so that AMD can funnel that into competing on the GPU high end (and corp AI space), and then start to give Nvidia a run for their money. Then they'll be sitting on their laurels in the CPU space long enough that Intel can come back and kick their face in with minimal effort, and everything will be right again after all is in equilibrium.

Yeah, I would make a good contemporary fiction writer.
 
Mazzspeed said:
...

Keep vcore and temps within acceptable limits and electromigration simply isn't a problem.
Click to expand...
A bit simplified, but this we can agree upon. Temps and current needs especially to be in check to avoid problems from electron migration. Outside edge cases, it should not be a problem during the lifetime of the CPU if run within acceptable limits. But my point still stands. You cannot state that there has been no degradation without specialized tools. That a system is stable, does not mean there is no degradation. System stability is not a proof of no degradation and the only thing you can state is that you have not observed any point of failures due to degradation (or you only played Solitare and not exposed the CPU to any scenario where you can observe this).

If you have an OC and over time you notice the CPU cannot handle the same OC anymore. To increase stability, you up voltage or you reduce clocks or maybe you set it back to stock. System is stable. Ok. That doesn´t mean that there has been no degradation of your CPU, even if you now run it stable at stock.

You can observe failures due to degradation that have already happened and are now causing problems, but you cannot say there is no degradation without specialized tools to do so. You can very well have a degradation that has not reached point of failure yet.
d3athf1sh said:
there are processors from the 70's that still work fine. these intel chips aren't lasting 3 months. that's how we can tell
Click to expand...
Yes. People are observing failures occurring more and more frequent. Some have changed CPU and problems went away, others have reduced clocks and/or used intel failsafe voltage to stabilize their systems. It seems to occur over time and start out stable. This smells like bad overclock @ stock from MB manufacturers. You can easier tell that there has been a degradation without specialized tools, then to say that there is no degradation which was my point to Mazzspeed.

That said, I am curious to Intels statements that are supposed to come this month about this. That the CPU have been degraded, does not necessarily mean you have a ticking timebomb. Even if you have encountered issues, they can be mitigated (voltage and clocks being the most obvious) and you can have a stable system for years/expected lifetime of the CPU, depending on the extent of the degradation. Intel have the tools to investigate this.
 
Starfalcon said:
It was an icewater chiller, not just water cooling.
Click to expand...
Yes exactly. It was definitely a low point for Intel. They got caught out for running absolutely exotic cooling with huge power requirements. So funny :ROFLMAO:
@GoldenTiger it is hilarious you have such bad memory in regards to this event that you go to levels of telling us it was basic water cooling.
 
Tamlin_WSGF said:
A bit simplified, but this we can agree upon. Temps and current needs especially to be in check to avoid problems from electron migration. Outside edge cases, it should not be a problem during the lifetime of the CPU if run within acceptable limits. But my point still stands. You cannot state that there has been no degradation without specialized tools.
Click to expand...
Well if that's the case, than you can't claim electromigration can be a problem during the lifespan of a CPU without specialized tools. So checkmate, my point still stands.

The 8700k in my sig has been running at between 5GHz with AVX offset and 4.8GHz without AVX offset since 2018, all cores synced when running gamemode under Linux - Stability is perfect to this day. Furthermore, once again, the 68030 that's likely older than you are still runs perfectly in an overclocked state to this day:

20240501_214916-jpg.651363

20240501_214751.jpg
 

Attachments

  • 20240501_214916.jpg
    20240501_214916.jpg
    325.8 KB · Views: 0
Mazzspeed said:
Well if that's the case, than you can't claim electromigration can be a problem during the lifespan of a CPU without specialized tools. So checkmate, my point still stands.
Click to expand...
Moving the goalpost much? My claim is that you cannot claim that you have no degradation or no extra degradation due to overclock without any specialized tools to observe that. And I dont claim electron migration can be a problem during the lifespan of a CPU without specialized tools. That is moving the goalpost.

You claimed no observable degradation. You need specialized tools to do such observation...

Saying your system is stable, does not prove no degradation.
 
I think they just mean functional degradation when they say observable. Arguably the terms are interchangeable for the average consumer.
 
Tamlin_WSGF said:
Moving the goalpost much? My claim is that you cannot claim that you have no degradation or no extra degradation due to overclock without any specialized tools to observe that. And I dont claim electron migration can be a problem during the lifespan of a CPU without specialized tools. That is moving the goalpost.

You claimed no observable degradation. You need specialized tools to do such observation...

Saying your system is stable, does not prove no degradation.
Click to expand...
Except I don't need specialized tools to observe the possibility of operational degradation regarding the CPU's in question, which was my point from the beginning. No goal posts were shifted whatsoever.

In the case of this very thread, Intel CPU's are showing observable degradation due to the fact they crash under operational workloads the CPU's in question should handle with perfect stability - No mystical specialized tools required. This very thread renders your perspective 100% moot. In the case I cannot observe operational degradation under any workload presented to the processor, as is the case regarding CPU's I have overclocked in the past, the unfounded belief that I require mystical specialized tools to measure degradation that in no way affects the operation of the CPU is moot.

It's like saying: "My engine handles 500HP at the 7500RPM rev limit it was built to sustain just fine", and then you come along and say: "But after 4000klms you have to run the engine at full load to 8000RPM just to be sure" - Even though the engine still runs fine, makes a perfect 500HP at the 7500RPM limit and is showing no observable faults...
 
Last edited:
Mazzspeed said:
Except I don't need specialized tools to observe the possibility of operational degradation regarding the CPU's in question, which was my point from the beginning. No goal posts were shifted whatsoever.

In the case of this very thread, Intel CPU's are showing observable degradation due to the fact they crash under operational workloads the CPU's in question should handle with perfect stability - No mystical specialized tools required. This very thread renders your perspective 100% moot.
Click to expand...
Yes you do if you are to make claim your CPU have no degradation. The others observe a point of failure and it points to a degradation to be the cause of that failure. That is a whole different chapter. Even here, the CPU have been stable for many and only unstable during certain tasks like UE5 shader compilation, not unstable all around. Even when degradation causes fault, it can be very hard to narrow it down to CPU in many cases, which is why finger has been pointed to everything from UE5 to Nvidia drivers.

Your statement about having NO observable degradation from OC require you to actually do proper observation of degradation to support that claim. Otherwise its only hot air. You dont have the tools for that and degradation is a slow death, meaning it can be perfectly stable for years until it shorts from degradation or become unstable.
Its not something you should use and can only act like some kind of invalidation of peoples issues or make people think that OC cannot cause degradation of CPUs and other chips for that matter.
 
Tamlin_WSGF said:
Yes you do if you are to make claim your CPU have no degradation. The others observe a point of failure and it points to a degradation to be the cause of that failure. That is a whole different chapter. Even here, the CPU have been stable for many and only unstable during certain tasks like UE5 shader compilation, not unstable all around. Even when degradation causes fault, it can be very hard to narrow it down to CPU in many cases, which is why finger has been pointed to everything from UE5 to Nvidia drivers.

Your statement about having NO observable degradation from OC require you to actually do proper observation of degradation to support that claim. Otherwise its only hot air. You dont have the tools for that and degradation is a slow death, meaning it can be perfectly stable for years until it shorts from degradation or become unstable.
Its not something you should use and can only act like some kind of invalidation of peoples issues or make people think that OC cannot cause degradation of CPUs and other chips for that matter.
Click to expand...
....
Mazzspeed said:
It's like saying: "My engine handles 500HP at the 7500RPM rev limit it was built to sustain just fine", and then you come along and say: "But after 4000klms you have to run the engine at full load to 8000RPM just to be sure" - Even though the engine still runs fine, makes a perfect 500HP at the 7500RPM limit and is showing no observable faults...
Click to expand...
My CPU's showed no observable degradation under any workload presented to the processors in question. There was nothing to 'narrow down' as the CPU's in question handled every workload presented to them in a faultless manner. Unlike the Intel CPU's highlighted in the topic of this thread.

The only individual invalidating anything here is yourself with talk of mystical CPU degradation dyno's.
 
Last edited:
Mazzspeed said:
....

My CPU's showed no observable degradation under any workload presented to the processors in question. There was nothing to 'narrow down' as the CPU's in question handled every workload presented to them in a faultless manner. Unlike the Intel CPU's highlighted in the topic of this thread.

The only individual invalidating anything here is yourself with talk of mystical CPU degradation dyno's.
Click to expand...
That is factually wrong. In best case, your CPU did not show you any errors in given tasks that you can suspect is caused by degradation. You did no observation of the degradation itself. Even more, besides being factual wrong, that statement does not contribute anything but muddy the issue. I am not trying to correct grammar, that is not my point, my point is that you did not observe any degradation because you dont have the nessesary tools to do so. You have infact no idea the state of the degradation inside your CPU at this stage. In an enthusiast forum, it should be made clear also that OC can reduce the lifespan of your CPU and in some cases it can fail or start to fail within the lifetime of your CPU. Even if you have a stable OC, there is no guaranty that it will remain stable. You might even believe its stable, until you run certain tasks like Cinemabench or UE5 shader compilation as discussed in this thread.

Do you find CPU degradation mystical? Or electromigration mystical? There is nothing mystical about it. Just displacement of atoms from running current through a conduit. Might lead to higher resistance over time and you need more voltage to overcome it or it might short. Might leak outside of insulation. Or nothing meaningful will happen during lifetime and usage of your CPU. Its not mystical, just common physics known industry wide.

You can feel free to claim you have OCd a CPU and have no errors that you believe is caused by your OC. But, please dont state that your CPU have no degradation from it without testing it. Might lead people to believe that this is not something that they might encounter when venturing into the world of overclocking.

What you state, is very different from what others do in this thread. People talk about errors that can be caused by a degradation in the CPU as a result from bad default MB settings. Similar to a bad overclock. If there is no point of failure, does not mean degradation of those settings have not already happened, only that it might not have reached point of failure yet.
That goes for you too. If your CPU have no point of failure yet, it does not mean that your CPU have not degraded more from your OC. You cannot make that claim.

Your statement that I first responded to was:
Mazzspeed said:
I've been overclocking since the Celeron 300a days, possibly earlier if you consider I have a Commodore Amiga running an overclocked 68030 25MHz @ 40MHz. I've never encountered any degradation as a result of overclocking in all the years I've been doing it.
Click to expand...
Again, how do you know? Did you measure degradation in any way and check if there were any deviation or just blowing hot air about a degradation you most likely have no tools to check for? Are you trying to claim that OC cannot cause more degradation? Or want a pat on your back because your chip had high enough quality to maintain OC so far, while others might have not been so lucky?
 
Tamlin_WSGF said:
That is factually wrong.
Click to expand...
No, no it's not.

CPU's are binned according to clock speed, quite often certain CPU's will be binned at a lower clock speed than the speeds they're realistically capable of to fill a popular slot in the market. Such CPU's can be pushed to higher clock speeds with no notable increase regarding any form of degradation assuming vcore and thermals are kept within Intel specification, and that specification can provide notable allowances - That's factually correct, that's called 'overclocking'. Granted, the latest Intel CPU's effectively overclock themselves to the bleeding edge, as is the case regarding 13th and 14th generation i7 and i9's - Meaning that these days, CPU overclocking is almost redundant. However to claim that those with a good 30 years of experience in overclocking earlier generation CPU's are just 'wearing out their CPU's faster via electromigration' is nothing more than complete rubbish - Especially when it's claimed that the only way to know for certain if you've done damage is to use so called specalized tools no one has access to.

Tamlin_WSGF said:
You can feel free to claim you have OCd a CPU and have no errors that you believe is caused by your OC. But, please dont state that your CPU have no degradation from it without testing it.
Click to expand...
...
Once again:
Mazzspeed said:
The only individual invalidating anything here is yourself with talk of mystical CPU degradation dyno's.
Click to expand...

So swinging back to the actual topic of this thread: When a CPU runs 1.6v vcore from factory, it's obvious that mobo manufacturers bypassing current limiters by default may result in CPU damage/instability in certain scenarios. Likewise, as someone that's been overclocking for the better part of 30 years without issue, Intel's rated vcore regarding the CPU's in question appears quite extreme - Quite possibly Intel are pushing their silicon the the edge in order to gain an advantage over the competition.

That's it. There's nothing more to discuss. /micdrop
 
Last edited:
Mazzspeed said:
CPU's will be binned at a lower clock speed than the speeds they're realistically capable of to fill a popular slot in the market.
Click to expand...
(Coming back to topic)

The suspicion is Intel has binned more aggressively here. Board partners are now increasing voltage & decreasing current. (This can impact all-core boost/top frequencies, effectively converting the chip to a lesser version down the stack)
 
Marees said:
(Coming back to topic)

The suspicion is Intel has binned more aggressively here. Board partners are now increasing voltage & decreasing current. (This can impact all-core boost/top frequencies, effectively converting the chip to a lesser version down the stack)
Click to expand...
As quoted here:

Mazzspeed said:
So swinging back to the actual topic of this thread: When a CPU runs 1.6v vcore from factory, it's obvious that mobo manufacturers bypassing current limiters by default may result in CPU damage in certain scenarios. Likewise, as someone that's been overclocking for the better part of 30 years without issue, Intel's rated vcore regarding the CPU's in question appears quite extreme - Quite possibly Intel are pushing their silicon the the edge in order to gain an advantage over the competition.
Click to expand...
I totally agree. However, I believe not all CPU's are showing any sign of lasting damage - In a number of cases simply enabling all limiters as per Intel specification seems to resolve any stability issues.
 
Marees said:
Early testing of Intel baseline bios by chiphell user:

Raptor Lake stability rate, measured by user kmdkai @ Chiphell

20% stable – 14900K
30% stable – 14900K + Reduced Loadline

40-50% stable – 13900K
50-70% stable – 13900K + Reduced Loadline
60-70% stable – 13900K on B660/B760

100% stable – 13700K

https://www-chiphell-com.translate....html?_x_tr_sl=auto&_x_tr_tl=en&_x_tr_hl=en-GB

https://x.com/3DCenter_org/status/1786026159270224295
Click to expand...
I still think something changed in the microcode which changed core utilization and current and power draw, for certain scenarios. Probably a little after 14th gen launch. Maybe when they added that feature which boosts game performance. Called APO.


Raptor Lake has been out for a long time and has been well liked. If only 50% of 13900k are stable after setting Intel stock values----then most of them were not stable at motherboard stock. Which doesn't make long term sense. As we would have been hearing about this a lot, for a long time. Intel stock values are a temporary mitigation.

Aftter a few more weeks of profiling, they will find it and alter microcode and release new bios. And we will be back to good ol' Raptor Lake.
 
Lakados said:
I am sorry man but the "Failsafe" and "Factory Default" options should not be out of the box OC'd to high hell on any component. You need a safe baseline to start anything from and if Factory Default isn't that then there have been a lot of failures in that product line before you even touched it.
Click to expand...
It's not different than overclocked RAM or motherboards that claim extremely high OC figures.
Everyone thinks it should work out the box but they don't. I wish it just worked also, but they don't lol. It's the name of the game when it comes to anything overclocked. Just because it's printed on the box doesn't mean it's not overclocked so treat it as such and know better. Marketing is marketing they always have crazy claims. It's up to you if you want to deal with overclocked components as a customer.
 
xDiVolatilX said:
It's not different than overclocked RAM or motherboards that claim extremely high OC figures.
Everyone thinks it should work out the box but they don't. I wish it just worked also, but they don't lol. It's the name of the game when it comes to anything overclocked. Just because it's printed on the box doesn't mean it's not overclocked so treat it as such and know better. Marketing is marketing they always have crazy claims. It's up to you if you want to deal with overclocked components as a customer.
Click to expand...
It's very differen't because overclocked RAM isn't overclocked out the gate, you load the auto profile for that CPU and it will start out at 5400 with some loose ass timing even if that RAM is rated for 7000, and you have to implement the timings to move it up to those speeds, sure it will provide you a base "this should work" profile that you can work from there but the base default is always the one the CPU was initially rated for.
 
xDiVolatilX said:
It's not different than overclocked RAM or motherboards that claim extremely high OC figures.
Everyone thinks it should work out the box but they don't. I wish it just worked also, but they don't lol. It's the name of the game when it comes to anything overclocked. Just because it's printed on the box doesn't mean it's not overclocked so treat it as such and know better. Marketing is marketing they always have crazy claims. It's up to you if you want to deal with overclocked components as a customer.
Click to expand...
....what are you talking about?

I'm not sure what is so hard to understand that parts should be stable out of the box regardless if they're running at the number on the box or not.
 
Lakados said:
It's very differen't because overclocked RAM isn't overclocked out the gate, you load the auto profile for that CPU and it will start out at 5400 with some loose ass timing even if that RAM is rated for 7000, and you have to implement the timings to move it up to those speeds, sure it will provide you a base "this should work" profile that you can work from there but the base default is always the one the CPU was initially rated for.
Click to expand...


Right but in the case of the Intel CPUs and In
tel motherboards the motherboards are not getting the setting correct in the first place running them unlocked with whatever settings and voltages they see fit. Every motherboard runs different voltages and tweaks it's not universal across any 2 boards so unlocked is meaningless because the board has built in settings for all the different CPUs.


NightReaver said:
....what are you talking about?

I'm not sure what is so hard to understand that parts should be stable out of the box regardless if they're running at the number on the box or not.
Click to expand...

You're delusional if you think overclocked CPUs and motherboards and RAM will work perfectly out the box. You can kick and scream all you want but until the manufacturers do something about it for baseline settings it's just the way it is right now and you have to manually tweak it. Not sure what you don't understand about that?
 
xDiVolatilX said:
Right but in the case of the Intel CPUs and In
tel motherboards the motherboards are not getting the setting correct in the first place running them unlocked with whatever settings and voltages they see fit. Every motherboard runs different voltages and tweaks it's not universal across any 2 boards so unlocked is meaningless because the board has built in settings for all the different CPUs.




You're delusional if you think overclocked CPUs and motherboards and RAM will work perfectly out the box. You can kick and scream all you want but until the manufacturers do something about it for baseline settings it's just the way it is right now and you have to manually tweak it. Not sure what you don't understand about that?
Click to expand...
Overclocked ram does come stable out of the box. At least the ones I've bought. You usually have to enable a profile to reach the stated overclocked speed.

Nobody is complaining that Intel allows you to go to such high values on their processors. It's the allowing of such potentially damaging settings as factory default that is the problem.
 
xDiVolatilX said:
You're delusional if you think overclocked CPUs and motherboards and RAM will work perfectly out the box. You can kick and scream all you want but until the manufacturers do something about it for baseline settings it's just the way it is right now and you have to manually tweak it. Not sure what you don't understand about that?
Click to expand...
Your 13900KS is an overclocked processor out of the box, constantly pushing itself to the bleeding edge of performance - You'd want to use those specialized tools to check for degradation. As for overclocked ram, it's called an XMP profile, and I've never seen it enabled 'out the box'.
 
Mazzspeed said:
As for overclocked ram, it's called an XMP profile, and I've never seen it enabled 'out the box'.
Click to expand...
This actually screwed me upgrading some laptop ram. The bios had no way for using xmp profiles.

Thankfully Kingston makes laptop ram that IS overclocked out of the box for this specific reason.

Only time I've ever seen it.
 
NightReaver said:
Overclocked ram does come stable out of the box. At least the ones I've bought. You usually have to enable a profile to reach the stated overclocked speed.

Nobody is complaining that Intel allows you to go to such high values on their processors. It's the allowing of such potentially damaging settings as factory default that is the problem.
Click to expand...

Absolutely not. RAM has the most instability even with xmp enabled out of the box. A lot of profiles work, then again a lot of profiles don't. A majority of the complaints about RAM is that is doesn't provide perfect stability at XMP settings. Leaving it default is fine but that's not exactly the same as what is being discussed here. RAM has it's own settings.

Mazzspeed said:
Your 13900KS is an overclocked processor out of the box, constantly pushing itself to the bleeding edge of performance - You'd want to use those specialized tools to check for degradation. As for overclocked ram, it's called an XMP profile, and I've never seen it enabled 'out the box'.
Click to expand...

Except that even my overclocked out the box 13900KS on my high end board with high end RAM 2 different sets I've tried is not stable out of the box. I've had to manually tweak it get it running fully stable in OCCT Linepack and Prime. Not to mention a slew of games.

Yes again XMP has it's own settings. That's different and that's fine. Doesn't relate to CPU and boards and their relationship with each other.
The board has it's own values and settings that control each CPU and it varies greatly from board to board and chip to chip. I've never had a CPU fail on me in the 25 years I've been building my own rigs for a hobby so I am not concerned about it. Every single one of my CPUs have been overclocked mildly. One this is for sure even in my early days I have never left it out the box settings and have always tweaked the settings in the motherboard, that was most of the fun for me having the control in my hands.

This isn't a PlayStation, in my opinion it is meant to be tuned just like a high performance car. It doesn't work out the box you absolutely need to tune it. To think it runs flawless out the box is silly. It's custom built. Anything custom built it's expected.
 
xDiVolatilX said:
Absolutely not. RAM has the most instability even with xmp enabled out of the box. A lot of profiles work, then again a lot of profiles don't. A majority of the complaints about RAM is that is doesn't provide perfect stability at XMP settings. Leaving it default is fine but that's not exactly the same as what is being discussed here. RAM has it's own settings.
Click to expand...
Yes, fast/tight profiles are not guaranteed to work depending on your IMC as well as a number of other factors - Such is the nature of XMP profiles, not too sure just what you're trying to argue here.

xDiVolatilX said:
Except that even my overclocked out the box 13900KS on my high end board with high end RAM 2 different sets I've tried is not stable out of the box. I've had to manually tweak it get it running fully stable in OCCT Linepack and Prime. Not to mention a slew of games.
Click to expand...
See above. Linpack and Prime are not enough, you need specialized tools to check for stability/degradation. /s
 
I've already typed my argument in response to yours. I'm not sure what you're asking me now? I guess I can say the same thing, see above Lol
 
xDiVolatilX said:
I've already typed my argument in response to yours. I'm not sure what you're asking me now? I guess I can say the same thing, see above Lol
Click to expand...

Because reading this post, you're a hypocrite:

xDiVolatilX said:
I've never had a CPU fail on me in the 25 years I've been building my own rigs for a hobby so I am not concerned about it. Every single one of my CPUs have been overclocked mildly. One this is for sure even in my early days I have never left it out the box settings and have always tweaked the settings in the motherboard, that was most of the fun for me having the control in my hands.
Click to expand...
 
News is getting worse on this. A tester in China bought up 100+ chips to test. 50% of 13900k are stable in auto profile. Worse numbers for 14th gen.

https://wccftech.com/only-5-out-of-...rofile-intel-board-partners-stability-issues/


  • Intel Core i9-13900K "AUTO -253W" - 40/50% (4/5 out of 10 units stable)
  • Intel Core i9-13900K "Reduced Loadline" - 50-60% (5/6 out of 10 units stable)
  • Intel Core i9-13900K "B760/B660 Board" - 60-70% (6/7 out of 10 units stable)
  • Intel Core i9-14900K "AUTO - 253W" - 20% (2 out of 10 units stable)
  • Intel Core i9-14900K "Reduced Loadline" - ~30% (3 out of 10 units stable)
  • Intel Core i9-14900K "B760/B660 Board" - 40% (4 out of 10 units stable)
Click to expand...
 
Assuming the i9 buyers are not too numerous compared to the sub $400 cpu crowd. I wonder how many RMA requests Intel is getting over this. Even 10 or 20% of halo buying customers wanting a replacement would be incredibly bad.
 
I think we need to return to a baseline of being able to expect CPU manufacturers and mobo makers to have "safe-ish" defaults for Auto and "tested OC" settings. These are not meant to be the highest, most granular tweaks possible so the idea that someone buying the $800+ motherboard and expecting to get something better out of it than the $200 without harming their CPU is completely reasonable. Likewise, CPU manufacturers keeping the stock settings so low as a way to claim its everone else's fault isn't viable either - the same is true for memory for that matter. It wan't that long ago you could pick up a literal 1GHZ OC all core turbo, perfectly safely within reasonable temps and with minor voltage tweaks if any.

While today's CPUs seem to have less "room" left, it should be relatively easy to ensure that there are sane yet performant options even for high performance CPUs on a highest end mobo. Mobo manufacturers, especially those offering advanced features and trying to set themselves apart from others with high end items, enhanced VRMs and cooling etc. OCing getting easier is generally a good thing, but I think the issue with having the official spec for a CPU, RAM, or even GPU etc.. being something that's so far beneath what the chip can do should be framed as a "failsafe, if you need to rule things out put it back to this", not as a common usage spec expected because otherwise it gets people used to looking past the defaults as a norm when manufacturers put them far from what the platform can realistically accept. This also goes for mobo manufacturers who, as long as CPU manufactures will be overly conservative with parameters, should try to offer their own verified fallback/normal/high performance settings that, while ideally they'd be able to go back to Intel or AMD and say "Look, we've done this work and we really think these voltages and other settings are safe under these parameters of cooling, the VRMs we're using on this board etc... so please verify we didn't miss anything", but if nothing else can come up with their own settings they're willing and comfortable supporting themselves in a "We tested it enough we are willing to take RMAs if our customers require it".

If a user is clearly making use of advance granular options or tried to run LN2 mode on air because "I thought it would make it faster", otherwise dropped all the guardrails even though there is a "Disabling this should only be for advanced users who know what they're doing. We're serious, this is not something even most overclockers should need. Press 1 to abandon all hope ye who enter this menu" mesage, that's one thing. However if someone turned on the MONSTER MONDO OC XD DYNAMIC PROFILE 1 setting that came with the mobo,it should be relatively safe at least provided sufficient cooling and doesnt' do things like disable C-states or otherwise mess with things.

With Intel specifically I think the design for the 12, 13, and certainly 14 series higher end i7 and i9 K procs are known to be massive "housefire" levels of power required which is its own issue. However that is separate from (though certainly excerbated by) mobo settings and variations there. There have always been CPUs that thanks to their designs have some potential issues (ie back in the Bulldozer/Excavator days for AMD etc) but clearly there's more going on here but I am worried that there will be too much finger pointing and not enough working together to sort things out properly.
 
Last edited:
TheHig said:
Assuming the i9 buyers are not too numerous compared to the sub $400 cpu crowd. I wonder how many RMA requests Intel is getting over this. Even 10 or 20% of halo buying customers wanting a replacement would be incredibly bad.
Click to expand...
And returns. I was in the return window on my 14th gen and let the retailer deal with Intel or their distributor. How many will be resold through those channels?

Still waiting for Intel. Trust is pretty much gone at this point but May just started.

If you want one, go old school and plan on buying a few to get the silicon lottery roll you expect.
 
wareyore said:
And returns. I was in the return window on my 14th gen and let the retailer deal with Intel or their distributor. How many will be resold through those channels?

Still waiting for Intel. Trust is pretty much gone at this point but May just started.

If you want one, go old school and plan on buying a few to get the silicon lottery roll you expect.
Click to expand...
yeah they'll prob just modify the internal chip id's and sell them as lower bins to somebody like Dell for non oc'able prebuilts
 
You must log in or register to reply here.
Back
Top