• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

Intel Statement on 13th and 14th Gen Core Instability: Faulty Microcode Causes Excessive Voltages, Fix Out Soon

Nvidia Bumpgate ~2008. Permanently ruined their relationship with Apple. Nvidia blamed TSMC for it.
Was that the faulty GeForce 8-series GPU? I thought that one was faulty packaging. But yes, it's close.
 
So amd overvolted their 3d chips in order to compete turning them into handgrenades and then Intel overvolted theirs resulting in crashes. Lovely market, competition is great

It was not amd who overvolted the cpus but the motherboard manufacturers with their bioses.

X3D cpus are voltage limited so overvolting them serve no purpose, especially if they are already winning.
 
Was that the faulty GeForce 8-series GPU? I thought that one was faulty packaging. But yes, it's close.
The Nvidia Bumpgate controversy refers to a significant issue that emerged around 2008 involving defective Nvidia graphics processing units (GPUs). These defects were primarily associated with the materials and manufacturing processes used in the GPUs, leading to widespread failures in laptops and other devices. Here are the key aspects of the controversy:

### 1. **Defective GPUs**
The problem was linked to a specific packaging material used in the GPUs. The solder bumps, which connect the GPU die to the package substrate, were prone to cracking under heat and stress. This was exacerbated by thermal cycling (repeated heating and cooling) which is common in laptop usage.

### 2. **Affected Devices**
The defective GPUs were used in a range of laptops from major manufacturers, including Dell, HP, Apple, and others. Users began reporting issues such as graphical artifacts, system crashes, and complete device failures.

### 3. **Nvidia’s Response**
Initially, Nvidia was slow to acknowledge the extent of the problem. However, as the issue became more widespread, Nvidia took some responsibility and set aside a substantial sum of money (around $200 million) to cover warranty claims and repairs.

### 4. **Legal and Financial Repercussions**
The controversy led to several class-action lawsuits against Nvidia and the affected laptop manufacturers. These lawsuits alleged that both Nvidia and the manufacturers knew about the defects but continued to sell the affected products without adequately informing consumers.

### 5. **Settlement and Compensation**
Eventually, Nvidia and the manufacturers reached settlements with affected consumers. These settlements typically involved extending warranties, offering free repairs, and in some cases, providing replacement devices.

### 6. **Impact on Reputation**
The Bumpgate controversy significantly impacted Nvidia's reputation. It highlighted the challenges of maintaining quality control in complex semiconductor manufacturing processes and underscored the importance of transparent communication with consumers when defects are discovered.

### 7. **Technical Specifics**
The issue was traced back to the high-lead solder bumps used in the GPUs, which were less reliable under the thermal stresses of laptop operation compared to more resilient materials like eutectic solder.

The Bumpgate controversy remains a notable example of the complexities and risks involved in high-tech manufacturing, as well as the potential consequences when these risks are not managed effectively.
 
So it was sufficiently well-known that ChatGPT could give a canny summary. Great.

Interesting how history repeats itself.
 
So many questions, how long did intel know about this? Is this really the only factor? What other CPU's might be or are impacted by this? Even if totally true, its going to be hard to convince the public otherwise. The performance/watt of Arrow Lake would have to be 2X that of Zen5 X3D, and Zen5 would have to be a hot unstable mess before considering it, and the likelihood of that happening is near zero. GG intel, are you sure you aren't in their rear view mirror now?
 
Nvidia Bumpgate ~2008. Permanently ruined their relationship with Apple. Nvidia blamed TSMC for it.

That was actually around ~2006, though? Bumpgate affected G80 and its smaller subvariants, particularly in mobile where temperatures were higher and thermal expansion/contraction cycles punished the interposer more. By 2008 Nvidia was already moving to TSMC 55 nm process with GT200B (GTX 260 216, 275, 285, 295), from the original 65 nm GT200 (260 with 192 cores, 280). I believe G92 was not affected (9800 GTX, GTS 250)

Was that the faulty GeForce 8-series GPU? I thought that one was faulty packaging. But yes, it's close.

Yes.
 
So amd overvolted their 3d chips in order to compete turning them into handgrenades and then Intel overvolted theirs resulting in crashes. Lovely market, competition is great
What did AMD overvolt to compete with?
The culprit was mostly just board makers trying to out do each other, AMD responsible for not gate keeping enough.
FYI the issue with VSOC has nothing to do with Vcore or much to do with performance at all.
The memory settings AMD used in their benchmarks nor just about every review (DDR5 6000) did not require high Vsoc.
Some boards wanted to claim DDR5 6600 support and cranked up the Vsoc hoping for the best. You put the same ram and same CPUs in different boards can result in vastly different Vsoc.

The stock Vsoc of Raphael is only around 1V. The 1.4V+ that Asus fed into the IOD resulted in well over double the SOC power.
The new limit of 1.3V Vsoc still gives you around a 200% margin over stock, that seems pretty reasonable.
1721727564507.png

ASRock%20X670E%20Taichi%20AGESA%20Update%20Power%20No%20EXPO.png

1721727425894.png


The stock V/F curve of 13/14th gen on the other hand is programe by Intel at the factory, and this is killing chips running at stock without any OC.
 
Last edited:
It was not amd who overvolted the cpus but the motherboard manufacturers with their bioses.

X3D cpus are voltage limited so overvolting them serve no purpose, especially if they are already winning.

What did AMD overvolt to compete with?
FYI the issue with VSOC has nothing to do with Vcore or much to do with performance at all.
The memory settings AMD used in their benchmarks nor just about every review (DDR5 6000) did not require high Vsoc.
Some boards wanted to claim DDR5 6600 support and cranked up the Vsoc hoping for the best. You put the same ram and same CPUs in different boards can result in vastly different Vsoc.

The stock Vsoc of Raphael is only around 1V. The 1.4V+ that Asus fed into the IOD resulted in well over double the SOC power.
The new limit of 1.3V Vsoc still gives you around a 200% margin over stock, that is pretty reasonable.

The V/F curve of 13/14th gen on the other hand is programe by Intel at the factory, and this is killing chips running at stock without any OC.
So it seems I didn't remember incorrectly. From what I do remember it was a board issue (well, manufacturers running things out of specs for that teeeny little bit of bragging power). AMD did take the blame on themselves, though, and replaced the CPUs.

As for that voltage issue Intel now seems to blame, from what I can tell, they don't as much talk about the voltage going into the chip, but how said power is distributed internally. Which apparently is in a way resulting in these defects.
 
I have always deemed Intel's YOLO behavior with boosts since 12th gen a questionable one. I'm no silicon engineer, but when ones literally behind the tech mention that tolerances are being tested, who am i to doubt anything.
When I run intel for personal use, its being done with conservative V/F values. I'm not going to gain any significant performance by torture testing these chips anyway.
The sad part of this whole story to me is the fact most users will never know about this issue, and also aren't tech savvy enough to get into voltage tuning to prevent any longer term damages.

This, despite a possible microcode update, stands quite exposed to a regulatory problem Intel will have to face in court.
 
So it seems I didn't remember incorrectly. From what I do remember it was a board issue (well, manufacturers running things out of specs for that teeeny little bit of bragging power). AMD did take the blame on themselves, though, and replaced the CPUs.

As for that voltage issue Intel now seems to blame, from what I can tell, they don't as much talk about the voltage going into the chip, but how said power is distributed internally. Which apparently is in a way resulting in these defects.
Yes as said, while AMD had the responibility of keeping an eye on what the boards were doing, overall it was initiated by the board partners.
Also FYI anandtech tested different AGESA and bios versions on Asrock boards, and it seems they kept out of the Vsoc (VDDCR SOC) insanity.
AGESA 1.0.0.5c to 1.0.0.7 Firmware Testing: Temps, Voltages, Currents, and Power - Voltage Lockdown: Investigating AMD's Recent AM5 AGESA Updates on ASRock's X670E Taichi (anandtech.com)

ASRock%20X670E%20Taichi%20AGESA%20Update%20Voltages%20EXPO%20Enabled.png
 
Last edited:
I believe G92 was not affected (9800 GTX, GTS 250)
I remember a 9800 GT that i put in the oven a couple of times to fix it. So it defenetly had the same solder issue as my G80 based 8800 GTS 320.
 
Please can we keep discussions about G80/G92 bumpgate, and AMD CPU voltage, as minimal as possible to prevent this thread from being derailed and thus locked by mods?

Thank you.
 
Please can we keep discussions about G80/G92 bumpgate, and AMD CPU voltage, as minimal as possible to prevent this thread from being derailed and thus locked by mods?

Thank you.
I suspect that is the goal of some of the comments with "whataboutism".
Get the discussion derailed and hopefully shield Intel from further discussion/critisim.
 
So amd overvolted their 3d chips in order to compete turning them into handgrenades and then Intel overvolted theirs resulting in crashes. Lovely market, competition is great
Please get this right, AMD did not release overvolted or faulty CPUS. Certain motherboard companies pumped a little to much SOC voltage to it which resulted in the CPUS not working. To put the blame on AMD is a bit desperate to divert blame away from Intel to say the least.

What is not at fault is that Intel have a major design problem and have stayed radio silent until called out on it while knowing there is a fault.

To put it as simply as these are crashing is also belittling the problem. These are crashing and degrading.

You want to see the extent of the problem look here.

1721728138360.png

1721728154215.png

1721728176412.png

1721728193468.png
 
I suspect that is the goal of some of the comments of whataboutism.
Get the discussion derailed and hopefully shield Intel from further discussion/critisim.
Seeing how that sprung from a semi-rhetorical question for previous, similar, occurrences, guess I should have known better. Sorry about how that turned out.

Still interesting how things are repeating themselves.
 
Intel fucked up big time.
 
hm... I know some guys running i9 12900k at 1,56V for years now no issue, in my humble opinion 13 gen architectures changes caused CPUs to no longer tolerate high voltages.
 
So amd overvolted their 3d chips in order to compete turning them into handgrenades and then Intel overvolted theirs resulting in crashes. Lovely market, competition is great
This is pure bothersiderism bull that no one should listen to. Intel boasted that their chips were extremely stable versus the competition because reasons and blamed everyone (motherboard manufacturers especially) but themselves when their products turned out to be anything but stable.

Add statements from the CEO such as ‘geostability’ when it come to manufacturing and Intel looks like a scam company. Theranos esque comes to mind.


Intel should be boycotted and denied government funding until they spin off their fabs and forced to concentrate on actually making stable high quality products.
 
hm... I know some guys running i9 12900k at 1,56V for years now no issue, in my humble opinion 13 gen architectures changes caused CPUs to no longer tolerate high voltages.
The vast ocean of AVX512 mostly idle silicon acting like some sort of capacitance buffer?
 
hm... I know some guys running i9 12900k at 1,56V for years now no issue, in my humble opinion 13 gen architectures changes caused CPUs to no longer tolerate high voltages.
No he doesn't. He is lying to you. No way he is running 1.56 for more than 5 minutes.

This is pure bothersiderism bull that no one should listen to. Intel boasted that their chips were extremely stable versus the competition because reasons and blamed everyone (motherboard manufacturers especially) but themselves when their products turned out to be anything but stable.

Add statements from the CEO such as ‘geostability’ when it come to manufacturing and Intel looks like a scam company. Theranos esque comes to mind.


Intel should be boycotted and denied government funding until they spin off their fabs and forced to concentrate on actually making stable high quality products.
Can you stop talking about amds handgrenades and focus on the topic? Thanks.
 
Given this is a microcode fix, it will be delivered as a Windows update - no new BIOS version or end-user action needed. That's at least a relief for Intel and its board partners.
 
Back
Top