You're missing context such as being behind in process tech. You can't design your way out of node with half the density and lower perf/watt than your competitors while keeping up in performance.
It's not as if they're pushing to insanely high clocks relative to their competitors. When RPL first launched it was competing against Zen 4 that clocked up to 5.7ghz, the 7950X fmax is actually 5.85ghz under the right conditions - this is very similar to the original RPL that had TVB boost clocks of 5.8ghz. For the lower end SKUs, they clocked lower than their AMD equivalent (The 7600X/7700X has higher clocks than the 13600K, the 7900X has higher clocks than the 13700K). My point being that it's not as if the original RPL had some unprecedented boost clocks.
Edit: A contributing factor to this fiasco could be that RPL didn't have the typical product cycle. It was cobbled together in a hurry once it became apparent that MTL wouldn't be ready for Fall of 2022. The development time for the RPL lineup was considerably shorter than it was for ADL.
Last edited:
Yes, it may. But then again it seems to vanish - for some time - after a Clear CMOS even after loading my old BIOS settings afterwards. And I am willing to put all the blame for possible PCIe based BIOS corruption on Intel XTU's watchdog hard crashing the PC "as a failsafe if anything happens during tuning".
I will do another TFD tutorial run without skipping the movies now.
Last edited:
Watch a degraded 13900k crash running intel extreme tester in a couple of seconds. Time coded to 2m18s.
That's a very transients heavy and thus demanding test, basically as if you were switching Prime95 SFFT AVX2 on and off all the time within milliseconds. I like that and will add it to my test arsenal in the future to accompany the other transients heavy tests I like to use on top of the usual "constantly high load" suspects (P95, OCCT, Yc etc).
Last edited:
I played another round through the The First Descandant tutorial, skipping none of the cut-scenes. No crash.
Before the Clear CMOS it was crashing the system via slow-death blue screen (taking a minute) every time before the end of the tutorial (tried 4 times).
Before the Clear CMOS it was crashing the system via slow-death blue screen (taking a minute) every time before the end of the tutorial (tried 4 times).
I looked more closely. This is run with a zero AVX offset, so you are watching a considerably AVX overclocked 13900KS crash in a very demanding test. The (BIOS on "Auto") stock AVX offset on my 13900K is -5. He is running a KS at 5.6 GHz with no offset, that would be 56 vs. 57-5=52 aka 400 MHz higher than stock.
In fact I just had to change my AVX offset from 1 to 2, because 1 didn't pass the XTU test without blue screen, despite running stable in every other test I threw at it (including ones that likely no one else does). This runs my 13900K at 54/54/55/55/54/54/55/55 for all-core load then.
Last edited:
This is so bad if true
www.techspot.com
Intel laptop CPUs are crashing, too, as the company remains silent
Shortly after blasting Intel for the high failure rate of the company's recent CPUs, Alderon Games claims laptops running on Raptor Lake chips are no exception. The...
www.techspot.com
poke01
Diamond Member
- Mar 8, 2022
- 3,434
- 4,714
- 106
great, Intel craps the bed and ms craps the bed today. PC industry in shambles.This is so bad if true
Intel laptop CPUs are crashing, too, as the company remains silent
Shortly after blasting Intel for the high failure rate of the company's recent CPUs, Alderon Games claims laptops running on Raptor Lake chips are no exception. The...
- Aug 22, 2001
- 31,216
- 29,849
- 146
Techspot reporting is slooooow on the uptake. I posted Matt's feed almost a week ago - https://forums.anandtech.com/thread...engine-games-and-others.2617728/post-41250996
Also, Frame Chasers is human garbage, a conman, and will not be getting a single second of view time from me.
Also, Frame Chasers is human garbage, a conman, and will not be getting a single second of view time from me.
- Jun 10, 2004
- 14,573
- 5,971
- 136
New GN video with some more leads:
"until Intel puts out a better statement, we simply do not feel comfortable recommending 13th and 14th Gen CPUs while this matter is pending"
"until Intel puts out a better statement, we simply do not feel comfortable recommending 13th and 14th Gen CPUs while this matter is pending"
Boom!
I asked one of Tom's writers (on Twitter) why they, as a consumer site don't take this stand, and recommend consumers to stay away from these processors. He said (at the time) they would not, but I had a feeling they have considered it.
And now, after GN statement, the dam may break.
- Jun 10, 2004
- 14,573
- 5,971
- 136
One major OEM: 10-25% affected. 8 million sample size.
Another American SI:
Another American SI:
AdamK47
Lifer
- Oct 9, 1999
- 15,653
- 3,518
- 136
CrowdStrike must have run the build that went to prod on an Intel CPU.
RnR_au
Platinum Member
- Jun 6, 2021
- 2,413
- 5,754
- 106
Tinfoil time... I wonder if Intel is keeping quiet until after the AMD 9000 series have been reviewed. Reviewers have to run the Intel cpu's at stock even if the cpu's have a known issue at that performance level. At best the reviewers can put big fat question marks next to the Intel results.
GTracing
Senior member
- Aug 6, 2021
- 478
- 1,111
- 106
To be fair, most of Tom's writers probably don't understand CPUs well enough to give good commentary. Their latest article about the issue was the same "raptor lake mobile is affected as well" article that the everyone and their mom put out yesterday. It's clear that Tom's is copying everyone else because Alderon made the Reddit comment days ago. The article includes this nugget of wisdom.
Very insightful, Tom's, thank you. It's clear that they're just trying to hit a word count.
And if you look at their CPU articles, you'll see 1) a buying guide full of affiliate links, 2) a piece about igor_kavinski's Cinebench scores that he posted here, 3) AMD's ancient 3Dnow instructions are removed from LLVM, 4) another best CPU article full of affiliate links, 5) the raptor lake mobile is affected article.
- Aug 22, 2001
- 31,216
- 29,849
- 146
TLDW: Sending one or 2 CPUs off to a failure analysis lab. It costs 5 figures each. If the oxidation of the vias/fabrication issue is happening, the lab should confirm it. Confirmation was given that yes, the i5 K SKUs for 13th are not immune. 14th gen data hasn't even been provided yet. It is expected to be ugly.
As posted above- a strong do not buy for any 13 or 14 series until further notice from GN.
S.I.s are confirming x53-x55 multi and dog slow DDR speeds like 4800, needed to get the CPUs stable.
One OEM alone, (from the hints I think it is Dell) has up to 8 million CPUs involved, with 6 million or so sus, and between 10-25% of those defective. That is just one OEM.
Some are already ramping up Ryzen SKU as they wait on Intel for guidance.
Intel is definitely in the process of doing a replacement program for the big customers. Unlike my initial guess, this is almost certainly going to blow the Xbox RROD losses away, even in adjusted dollars.
Some of those big customers are furious with the companies that sold them the systems with the CPUs. This is the factor that puts enough pressure on Intel to practically guarantee public statements and actions soon.
If the answer is downclock the CPU significantly and XMP is ignored, you can bet the farm false advertising, as many of us have posited, is going to be real talk. Like us, GN suspects the legal implications are what is holding up the response.
Laying off tens of thousands of employees. Plus this zinger -
As posted above- a strong do not buy for any 13 or 14 series until further notice from GN.
S.I.s are confirming x53-x55 multi and dog slow DDR speeds like 4800, needed to get the CPUs stable.
One OEM alone, (from the hints I think it is Dell) has up to 8 million CPUs involved, with 6 million or so sus, and between 10-25% of those defective. That is just one OEM.
Some are already ramping up Ryzen SKU as they wait on Intel for guidance.
Intel is definitely in the process of doing a replacement program for the big customers. Unlike my initial guess, this is almost certainly going to blow the Xbox RROD losses away, even in adjusted dollars.
Some of those big customers are furious with the companies that sold them the systems with the CPUs. This is the factor that puts enough pressure on Intel to practically guarantee public statements and actions soon.
If the answer is downclock the CPU significantly and XMP is ignored, you can bet the farm false advertising, as many of us have posited, is going to be real talk. Like us, GN suspects the legal implications are what is holding up the response.
Laying off tens of thousands of employees. Plus this zinger -
Not exactly the best motivators for your workers. Might not have anything to do with this , then again...
I watched the video and came away with a totally different take away than GN’s conclusion.
All of his quotes seem to point in the direction that Intel knows what the issue is. They’re telling their partners directly that one of the Intel 7 fabs screwed up and created this premature oxidation issue. This leads me to wonder why not put out a statement identifying the serial numbers effected and issue a recall?
All of his quotes seem to point in the direction that Intel knows what the issue is. They’re telling their partners directly that one of the Intel 7 fabs screwed up and created this premature oxidation issue. This leads me to wonder why not put out a statement identifying the serial numbers effected and issue a recall?
Panino Manino
Golden Member
- Jan 28, 2017
- 1,067
- 1,293
- 136
If only they did, would actually prevent what happened yesterday.
GTracing
Senior member
- Aug 6, 2021
- 478
- 1,111
- 106
If you're referring to the quote where a partner mentioned the oxidation, I assumed that was because Gamers Nexus mentioned oxidation first.
Saylick
Diamond Member
- Sep 10, 2012
- 3,883
- 9,021
- 136
Wintel has not the been hit this hard in a loooong time.
Thunder 57
Diamond Member
- Aug 19, 2007
- 3,493
- 5,792
- 136
Let me guess, Paul Alcorn?
That would make me LoL. Twice that I can think of that Dell got bitten in the ass for being basically an outlet for Intel only.
poke01
Diamond Member
- Mar 8, 2022
- 3,434
- 4,714
- 106
Intel fabs not being good, colour me shocked....
Hey, now at least Arrow lake will be good at least the TSMC variants.
Hey, now at least Arrow lake will be good at least the TSMC variants.
✅
There would certainly be Karma if Dell, which is very strong in US commercial desktop (Intel only) suffered from this more than any other OEM.
NTMBK
Lifer
- Nov 14, 2011
- 10,400
- 5,636
- 136
They should be benchmarking against the 12900K, because that's the last generation that actually functions correctly.
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 24K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes Discussion Threads
- Started by Tigerick
- Replies: 20K
-
-
-
Facebook
Twitter