Behind-the-Scenes: Intel Report
Added 2024-07-30 01:35:40 +0000 UTCHey everyone - (and probably Intel PR!)
We have been working on a report diving into more of the Intel failures. The story has been covered to death on the technicals and possible causes, so what we are working on currently has more to do with Intel Corporation's actions historically to these types of situations. I thought I'd give a preview for Patreon supporters.
Currently, we are going through as much of Intel's history as the team can remember to try and better understand if it has any form of 'playbook' or trusted course of action it takes for its colossal failures. Knowing this and plotting trends would be helpful for trying to better understand where it might go next with its current instability and oxidation issues.
Some examples include Intel's Atom issues, its old 1990s-era FDIV bug (that, IIRC, resulted in a recall), and its handling of the old "toothpaste" controversies during its era of CPUs with bad TIM.
We are looking at dozens of articles, reports, statements, reviews, and material about each of these and trying to map Intel's responses. It's remarkable how well some of them match today's trickle-feed of official information.
Regarding the current issue, I am still talking to FA labs. That Intel has confirmed oxidation issues at the vias on some chips (13th Gen, at least) certainly makes the job easier but, to quote a commenter, "it's like finding a needle in a billion haystacks." Maybe knowing this is a confirmed issue reduces that to "finding a needle in 900 million haystacks" instead. It'd be great educational material, so we're still pursuing that.
In the interim, we are monitoring the microcode situation closely. I've had a few extreme overclockers contact me with information about the new microcode, but we're not at a stage where we can publish yet. Requires vetting.
The bigger question that we've been kicking around internally remains the impending Zen 5 CPU launches: With Intel's microcode almost certainly dropping after these reviews, we currently will take the following approach -
We can't recommend any 13th or 14th Gen CPUs until Intel gives confidence to buyers, including with this microcode update
We also can't test what isn't available, so for the review, we'll test with the current revisions of everything (as of last week), but maintain our stance of "we are not recommending Intel 13th & 14th Gen at this time"
The data will still be on the charts
Intel has otherwise been impressively useless at making concrete commitments for its customers. It's like an ASUS situation, except with about 100,000 more layers of middle management and PR-by-committee, so at times, it is at least a little less tonedeaf than the ASUS situation.
A little.
We'll keep you all posted. In the least, it's been super educational to work on this content alongside some experts in the industry. I've enjoyed having a project to work with Wendell on. I'm also hoping to get with Gordon for a video sometime in the next month or two, but we'll see on timing.
-Steve
Comments
If you wan't to cover a somewhat breaking case of Intel dumbfuckery their new WiFi 7 M.2 module BE200 (even the ones not using CNVIO) are locked to Intel platforms. According to posts on reddit (so not confirmed to be an actual quote from Intel) it is claimed that their new modules are dependent on Intel vPro technology not present on AMD systems. However given how hard AMD is currently beating them in the consumer segment I would not be the least bit suprised if this is just Intel firmware checks that simply shuts down the card if it's seeing an AMD system. https://www.reddit.com/r/Amd/comments/1h280mc/does_intel_be200_work_on_am5_boards_with_the_new/ PS: Things get even spicier with some users reporting the card working on AM2 so there goes the Intel vPro explanation. Unless AMD somehow licensed this tech for AM2 which I can't see any reason for.
Anders tn
2025-01-28 10:24:01 +0000 UTCOh jeez you'd think these a ANUSES would want to do everything possible to keep a good rep but it's looking like the head in the sand approach only this one could turn out to raw dawg em hard and cost em a shit ton of their long term customers I'd say wake the fuck up INTEL as for the Gpu division i wrote a letter awhile back explaining the NGREEDIA SITUATION and from what I just saw of the 580 running on 8 of 16 pci-e Lanes hellz yeah they're in line to kick NGREEDIAs balls off especially if they release a card that's 3/4-1/2 the price and out performs NGREEDIAs Flagship 😁☺️👃🤞👍 omg let's pray for that 1☺️😃😄
Ernest G. Harvey Jr
2024-12-27 06:52:00 +0000 UTCThere is an interesting blog post over at Puget Systems who build a variety of different PCs for various applications: https://www.pugetsystems.com/blog/2024/08/02/puget-systems-perspective-on-intel-cpu-instability-issues/ Their experience appears to be quite different from what has been reported.
Alan Goldhammer
2024-08-07 01:35:51 +0000 UTCGreat GN team, will be interesting to see how Intel handles this. So far not impressed.
Leviathanprym
2024-08-06 02:10:34 +0000 UTCI could just about save a friend from buying a prebuilt 13th gen 💪 In today's world of zero responsibility (and getting away with it), I completely get why Intel is behaving the way it does - still, it stays scummy af to work that way. All for the short-term shareholder value... we have really peaked with our current system -_- As an EU citizen, it'd be very interesting to see the legal side of this. In Germany, we have a Mängelanspruch - roughly translated to "deficiency entitlement" and this is a LAW! It goes pretty far as well actually. First you can decide what action is allowed to be given to you. Yes, THIS WAY AROUND! ;) You pick between repair and replacement. The seller can deny both, or follow your choice. The seller can NOT simply vader it's way out of this (can we establish this saying?? 😁😂), he HAS to follow your choice. Repair is allowed to fail 3 times (because computers are complicated, duh, dinosaur logic). Replacement is of course fine, BUT the whole Mängel-datecounter is reset to zero in that case!! (for the defective part only) If no repair/replacement is fulfilled or successful, you can request at discount or rescind from the deal. To be exact: Usually the deal is with a seller like MindFactory, caseking etc., just as with your MicroCenters etc. As long as the defect is reported within the warranty period it is valid, even if verification etc. might take longer, or effects of the suspected defect only turn up later. Cases like this specific one will of course be difficult to associate and you will still be praying for your seller to play ball with you and not simply pull up a wall. https://www.verbraucherzentrale.de/wissen/vertraege-reklamation/kundenrechte/alles-zu-gewaehrleistung-und-schadenersatz-5057 https://www.verbraucherzentrale.de/umtauschcheck-54413 Sooo.... I don't see how Intel can weasel their way out of this over here. I suspect EU laws will be pretty similar, but I am not familiar with them. Thanks already for investigating this!! 😎
n8c
2024-08-01 22:25:53 +0000 UTCAfter seeing how hard Steve goes after scummy corporate companies intel must be shaking in their boots right now 😂
ZX Josh
2024-08-01 10:53:48 +0000 UTCI wish I had something more constructive to say, but all I can say is your recent coverage has been knocking it out of the park! Scripts and jokes recently have been fantastic. I don't look forward to much in life, but I do look forward to GamersNexus videos! (feel free to put that on a mug lol) Anyway, hope the entire team (especially you, Steve!) are getting enough rest and sleep!!!
Richard Williams
2024-08-01 00:31:45 +0000 UTCI want to say Intel had to do a recall on the Pentium 3 when they tried to push it past 1GHz and the silicon just wasn't capable. Might be one to add to your list for analysis.
skyhawk
2024-07-31 00:35:42 +0000 UTCBack to you, Steve.
MGsubbie
2024-07-31 00:17:18 +0000 UTCMoore's Law Is Dead says a lot of things, 90% of it turns out to be bullshit. I don't see how Intel would contact them and not far bigger, more reliable channels.
MGsubbie
2024-07-31 00:16:26 +0000 UTCSomebody had a 13700k on the way. I told them to return it and the motherboard and go with a 7800X3D insteaed. Thanks for all the hard work, Steve and the rest of the staff.
MGsubbie
2024-07-31 00:14:16 +0000 UTCIt sounds like you need to start building company profiles for easy reference. Similar how newspapers have person profiles already assembled for if something hits the fan when someone dies. They have something that can go to with all the key talking points already assembled. Clearly you can't do everyone but having a top 30 players in the software and top 30 in the hardware sector with a front to back profile might be worth the time and effort? No doubt in your copious free time. .....Yes that was a joke.
Kellic
2024-07-30 19:44:20 +0000 UTCBeen looking forward to this video for weeks! Can't wait. I trust you guys the most so I actually hold off on having set opinions until I hear GN's take on it. Also as a kid we had a HP Vectra at home with a 66mhz Pentium lol. It totally borked itself from that FDIV bug. I was pretty young at the time but I still remember it because I was addicted to Command and Conquer and was mad I couldn't play it any more! It was something to do with continued usage and these minor errors adding up and adding up until it was a major issue, and apparently C&C was really prone to this kind of issue.
Chairman Meow
2024-07-30 17:59:59 +0000 UTCYou guys are the best, keep it coming..
Thorgen_Ironside
2024-07-30 17:30:10 +0000 UTCI've worked with many people over the last 20+ years building custom PCs. I've found that most don't understand the individual components let alone root cause analysis. Many blame the "computer", "hard drive", "windows", or " 'app name' " when asked them to describe their issue. This works in Intel's favor because the vast majority wouldn't know to even suspect the CPU as the problem or how to determine it. This solution may work well for the less than 1% in the PC tech space, and have the knowledge to identify and address it appropriately. General users buy a computer and never upgrade the BIOS or any components. They may or may not contact the system OEM if they have occasional stability problem and will likely wait until the issue worsens. If this happens to be after the standard warranty period of 1 year. It likely won't be covered :(
Cameron
2024-07-30 17:19:00 +0000 UTCAdding a couple of pull-up resistors and enamelled wires will not address the problem this time.
Yugo
2024-07-30 11:41:43 +0000 UTCWhat if Intel would grant all buyers after the microcode hotfix new 3 years of guarantee. So people with advanced degeneration get a chance for exchange and Intel does not require for recall of all products. Would be kind a fair deal dont you think? If you dont have degeneration or less, fine your cpu can survive the guarantee if not you get a replacement. Nobody can determin the state of degeneration right now. So would be unfair to those people where the CPU dies after micro code and still in guarantee to be exchanged while others might die right after 3 years because the fix came to late.
n3cron
2024-07-30 11:00:36 +0000 UTCHaving bought an i5-13600K 2 weeks after its release date, 2 things I'm interested in: * is my CPU affected by the oxidation issue? * how can we be sure a CPU has degradation issues? When we have random BSODs? For the oxidation issue: is that static? I mean: will the oxidation still continue to get worse over time? Or is that a thing that happened during the fabrication, but it will stay as is? GN has already mentioned Intel should have some form where we can enter the identifiers on the CPU's lid to know whether it was built prior or after the oxidation issue was addressed. Would indeed be good for piece of mind. But that perhaps opens a can of worms for Intel, if customers start demanding replacements... I run my CPU pretty much stock, apart from enabling XMP and an undervolt of 75mV. I do not think my CPU is affected by degradation (yet). I do have an occasional software crash. As anyone, I guess... Sh*t happens... :) I don't think the CPU is to blame, probably just a software/driver bug. Nonetheless, as degradation might become an issue over time, perhaps a few tips what to look out for would be nice?
Nico Weytens
2024-07-30 10:49:24 +0000 UTCWill you use the "Intel Default Settings" with regards to TDP when comparing Zen 5 to Intel 13/14th gen? PL1 = 125W and PL2 = 253W for K-SKUs and PL1 = PL2 = 253W for KS-SKUs. Since these are the settings that are the common baseline among all Motherboard vendors and probably what Intel will stand by in terms of warranty support (despite previously claiming that "uncapped TDP was not overclocking and doesn't void warranty"). https://www.msi.com/blog/improving-stability-of-13th-14th-gen-intel-core-processors-with-intel-default-settings
Andreas Stennert
2024-07-30 09:30:06 +0000 UTCThe 14900k I got from igors binned batch has never seen automatic voltages, but even with that it has been acting up. Also asus apex board might be an issue but I have nothing to test against. I'd just really like to know what is going on.
Mia Mieps
2024-07-30 07:44:15 +0000 UTCThis does really feel like the tipping point personally. They have been losing confidence in their product for the past few generations with things like the bad efficiency and high temps, but that can all be fixed with a great product release. Can't fix confidence in support with anything but a lot of time.
Doug Mullaly
2024-07-30 05:00:04 +0000 UTCWould really love to see an entry on the graphs for an i9 with the "Wendell Safe Specs" just to see how truly terrible running a config that should keep your high-end 13th/14th gen chip from ever experiencing this over-voltage failure would be. 53x multiplier and DDR5-4200!
Schattenjäger
2024-07-30 04:52:53 +0000 UTCPC Magazine / John Dvorak used to use Intel as a punching bag. He was about as close to "gadfly" as you got in those days.
Brad
2024-07-30 04:43:52 +0000 UTCI think asianometry has done a few YouTube videos about intel's old past. It might give you a springboard on what to look for.
Nicholas Buckner
2024-07-30 04:32:47 +0000 UTCWould be nice to see Gordon again. :)
Mr.Frost
2024-07-30 04:04:29 +0000 UTCMoore's Law is Dead said he is having Intel employees contact him to find out what is going on. That speaks volumes about the lack of communication even within Intel. The oxidization issue according to Moore's Law is Dead occurred in their Handler AZ fab between March and June 2023. That was swept under the run until now. This may be a case study in business schools on how not to handle damage control in a large behemoth of a corporation.
OldFart
2024-07-30 04:03:26 +0000 UTCThanks a ton for the update and your diligence on this issue. I’m abroad but will try and get you a date code from my abysmally bad 13900k next week if possible - I’m honestly half tempted to not try and RMA it, and see if it’s still an abject failure post microcode at this point hahaha
David Moffitt
2024-07-30 02:23:09 +0000 UTCDefinitely agreed on that. Terrible handling. They can recover, but they need to start on that right away and make a lot of make-good gestures.
GamersNexus
2024-07-30 02:07:01 +0000 UTCAgreed!
GamersNexus
2024-07-30 02:06:35 +0000 UTCWorking on this as well. We don't have anything yet, though.
GamersNexus
2024-07-30 02:06:23 +0000 UTCGreat question. A lot of the "recent" history is easy. The stuff from the early 90s has been difficult. A lot of the newspapers and magazines have fortunately been preserved in some capacity, but it has been much harder to dig up!
GamersNexus
2024-07-30 02:06:03 +0000 UTCI have built countless Intel systems over several decades for friends and family and the way Intel has handled the oxidization utter lack of what happened to these CPUs and the microcode issues has left a sour taste in my mouth. I had ordered the parts for a simple i5-13600K non-gaming built. All those parts are going back and I built an AMD system instead. All the goodwill Intel had built up over the years has vanished. Way to go Intel!
OldFart
2024-07-30 01:55:29 +0000 UTCThanks for the update. I'm on 12th gen so it feels like I barely dodged a doozy of a situation. Who would have thought Alder Lake would still be the best Intel option in 2025?!
FrenziedManbeast
2024-07-30 01:44:06 +0000 UTCIt'd be good to know if there's more evidence of laptops also being affected. I get that isn't the general focus of the channel but it'd be good knowledge to have.
B. Fish
2024-07-30 01:43:52 +0000 UTCIs the historical information difficult to find? I feel that older stuff can be hard to find on the internet; whether it was mostly printed or all of the old publishers and sites are out of business.
Cody Konior
2024-07-30 01:42:06 +0000 UTCthanks steve!
Fdert
2024-07-30 01:39:53 +0000 UTC
