Arrow Lake instability issues?? RUH ROH!!

[ShrimpBrime]

Like everyone else, been looking for some news of the new Arrow Lake chips and in my news feed without needing to search found an article about Intel having instability issues early in the game.

Things look dire for Intel as Arrow Lake reportedly has instability issues

Stop me if you've heard this one before.

www.xda-developers.com

Not they haven't had enough problems with stability issues on Raptor Lake parts, supposedly fixed now, things look bleak for Intel.

For now, it seems speculation only while the processors are just released.

 

[sneekypeet]

I get it, but it didn't stop buyers if x99 or 299, and they both had major RAM issues at launch. Things will get better.

 

[unwind-protect]

Not much in the way of data there.

Failure of aggressive DRM or anti-cheat with a new CPU is much more likely the malicious software's failure.

 

[Ruru]

I wouldn't say that they have degradation issues like Raptor, probably just some early adopter stuff.

Still a hella meh launch though.

 

[ShrimpBrime]

Looks like the E-cores clock ok on LN2, I'm seeing some 6.6ghz out there. But so far, looks like only extreme overclocking (265K I'm looking at)_ going on instead of realistic ambient benchmarks so far.

Wprime
Ultra 265K (LN2) 1024m 6.9ghz/6.6Ghz scores 30.555s
14700K Ambient water (mine) 5.8Ghz/4.7Ghz was 30.949s

CBR2
Ultra 265K (LN2) 6.9Ghz/6.6Ghz - Max so far 20045pts
14700K (mine-DryIce no degradation) 6.16Ghz/4.95Ghz 15701pts

I would do more comparisons if the frequencies where closer, but I don't see awesome sauce here.

 

[dgianstefani]

When pairing Windows 24H2 with Arrow Lake, performance will be terrible—we've seen games running at 50% the FPS vs 23H2. One solution is to turn off Thread Director or disable the "Balanced" power profile, which is why we decided to use 23H2 for the time being. Last but not least, there are some driver issues and bluescreens when both a dGPU and iGPU are active at the same time.

Just launch firmware/Windows issues.

 

[Ruru]

Looks like the E-cores clock ok on LN2, I'm seeing some 6.6ghz out there. But so far, looks like only extreme overclocking (265K I'm looking at)_ going on instead of realistic ambient benchmarks so far.

Wprime
Ultra 265K (LN2) 1024m 6.9ghz/6.6Ghz scores 30.555s
14700K Ambient water (mine) 5.8Ghz/4.7Ghz was 30.949s

CBR2
Ultra 265K (LN2) 6.9Ghz/6.6Ghz - Max so far 20045pts
14700K (mine-DryIce no degradation) 6.16Ghz/4.95Ghz 15701pts

I would do more comparisons if the frequencies where closer, but I don't see awesome sauce here.

Wonder what arch they can be compared when going clock-to-clock? The E's I mean.

 

[unwind-protect]

They gave up Hyperthreading for this?

 

[Ruru]

They gave up Hyperthreading for this?

It's been reported for being a security issue, but couldn't that be fixed?

No idea that does that issue affect AMD's SMT, I remember reading about that just being with HT.

 

[unwind-protect]

It's been reported for being a security issue, but couldn't that be fixed?

No idea that does that issue affect AMD's SMT, I remember reading about that just being with HT.

Well, you can fix known security issues, but there were just so many new ones popping up with HT all the time. Maybe they wanted to plug that hole once and forever.

 

[chrcoluk]

Its not only that but some existing mitigations can brutally kill performance so e.g. if you leave HT enabled, there will be more cache and predictor flushes going on for security reasons. I can understand why they plugged that hole.

 

[Ruru]

Well, you can fix known security issues, but there were just so many new ones popping up with HT all the time. Maybe they wanted to plug that hole once and forever.

Just wondering that are those just some "bubblegum fixed" BIOS update-things like with 1st rev Phenom's TLB issues which were fixed with a new revisions.

To be honest, I don't know THAT much about hardware-related bugs and shit.

 

[ShrimpBrime]

Just launch firmware/Windows issues.

The quote you made was for 23H2, I'm even on 24H2 already cause it's pushed onto the consumer. I can only postpone the updates.

Management engine Firmware and Software will have lots of updates just like last gen. But doubtful will raise any performance, typically bug fixes degrade performance. Only time will tell I suppose.

 

[sLowEnd]

Sucks that this kind of thing made it through testing (if there even was any testing for popular anti cheats), but I expect it'll be fixed pretty quick

 

[RJARRRPCGP]

Looks like the E-cores clock ok on LN2, I'm seeing some 6.6ghz out there. But so far, looks like only extreme overclocking (265K I'm looking at)_ going on instead of realistic ambient benchmarks so far.

Wprime
Ultra 265K (LN2) 1024m 6.9ghz/6.6Ghz scores 30.555s
14700K Ambient water (mine) 5.8Ghz/4.7Ghz was 30.949s

The Wprime benchmark, reminds me of Pentium4, LOL!

 

[ShrimpBrime]

The Wprime benchmark, reminds me of Pentium4, LOL!

Right, no reason to change the benchmark, it has accurate consistency to reference CPU clock performance.

No different than you using "Prime"95 which uses large prime numbers also since P4 ""LOL!""

 

[RJARRRPCGP]

It wasn't a much better result at a much higher frequency!

 

[ShrimpBrime]

It wasn't a much better result at a much higher frequency!

Might be some skills involved from the lower frequency guy's submission too. WPrime isn't everyone's go to benchmark, there is some skill involved getting good scores from benchmarks(any). Not exactly always about the frequency, but it sure does help. But only 7Ghz on LN2 means P-cores won't/don't OC well on ambient setups.

 

[RJARRRPCGP]

This does remind me of Northwood, at least, where Northwood wasn't bad, but would still need a higher core frequency and likely would do better with new software than with old software, especially pre-2000 software, I suspect.

Are these new chips simply a variation of that, where performance will suck on many pre-2020s software?

 

[ShrimpBrime]

This does remind me of Northwood, at least, where Northwood wasn't bad, but would still need a higher core frequency and likely would do better with new software than with old software, especially pre-2000 software, I suspect.

Are these new chips simply a variation of that, where performance will suck on many pre-2020s software?

That is a good question. I think in many ways it's got great performance for certain things, but I think gaming is a huge focus. And if games aren't stable, which is widely used software, then yes. Could be a sucky variation of processors.

I find it funny that Intel shaved 3w and people say it's the most efficient cpu they've had yet. No, not in the top tier market at all. In fact it's horrible at 250w. These are the second highest wattage chips at the top end ever released(desktop). The work done for the wattage isn't that good for bragging rights, the competition clobbers them at only 170w.

Gotta say that 95% of my software is pre-2020 for sure. I love a lot of older 1080P DX-9 games. And luckily, the P-cores without HT should really shine here. But will it require turning off e-cores or waiting for firmware and software updates until it's good for most programs?? No idea. But I would hope to purchase something, install it and it work without the need for all that.

 

[RJARRRPCGP]

In fact it's horrible at 250w. These are the second highest wattage chips at the top end ever released(desktop). The work done for the wattage isn't that good for bragging rights, the competition clobbers them at only 170w.

Gotta say that 95% of my software is pre-2020 for sure. I love a lot of older 1080P DX-9 games. And luckily, the P-cores without HT should really shine here. But will it require turning off e-cores or waiting for firmware and software updates until it's good for most programs?? No idea. But I would hope to purchase something, install it and it work without the need for all that.

200-something watts, is like an OC'ed first gen! (Bloomfield)

Also, I hope that DirectX 9-based game performance, doesn't go down a cliff!

 

[ShrimpBrime]

200-something watts, is like an OC'ed first gen! (Bloomfield)

Also, I hope that DirectX 9-based game performance, doesn't go down a cliff!

It should go up with HT gone actually. Most of them older games don't scale past 4 cores maybe. Like 3DMark06 doesn't scale past 6 threads for example.

 

[MikeSnow]

Not they haven't had enough problems with stability issues on Raptor Lake parts, supposedly fixed now, things look bleak for Intel.

Yeah, I'm not so sure about those fixes. After my old 13900K required lower and lower maximum P-core frequencies to be stable, I got a replacement from Intel at the end of July. The replacement was stable for about two months, then this month it started showing stability issues again, browser tabs crashing, apps crashing. And today even a blue screen, immediately followed by another crash during the memory dump triggered by that blue screen, at 15%.

I applied all the Asus BIOS updates religiously, as they came out, did no overclocking (except XMP), so, unless I was extremely unlucky, it seems Raptor Lake is fundamentally flawed and unfixable.

At this point, after almost 2 years of having continuous issues with Raptor Lake, enough is enough. I will probably switch to the 9800 X3D as soon as it is released. Meanwhile, I'll have to start cutting the CPU frequency again.

 

[ShrimpBrime]

Yeah, I'm not so sure about those fixes. After my old 13900K required lower and lower maximum P-core frequencies to be stable, I got a replacement from Intel at the end of July. The replacement was stable for about two months, then this month it started showing stability issues again, browser tabs crashing, apps crashing. And today even a blue screen, immediately followed by another crash during the memory dump triggered by that blue screen, at 15%.

I applied all the Asus BIOS updates religiously, as they came out, did no overclocking (except XMP), so, unless I was extremely unlucky, it seems Raptor Lake is fundamentally flawed and unfixable.

At this point, after almost 2 years of having continuous issues with Raptor Lake, enough is enough. I will probably switch to the 9800 X3D as soon as it is released. Meanwhile, I'll have to start cutting the CPU frequency again.

I'm sorry to hear that, I have the opposite experience actually. Can't seem to kill my 14700K. It's a power house. Has been on DryIce across 3 boards. Have statically run 1.6+ v-core. Has scored top 100 and top 50 in a few benchmarks with it for reslut.

Your issue, I'd have to ask if the system specs are accurate. But 2x2x16 mix matched memory my not be favorable performance and stability.

I've noticed higher Cache frequency and SA voltage too low will also cause issues. Done be afraid of 1.3v SA. It'll run warmer, but should help stability.

 

[chrcoluk]

Yeah, I'm not so sure about those fixes. After my old 13900K required lower and lower maximum P-core frequencies to be stable, I got a replacement from Intel at the end of July. The replacement was stable for about two months, then this month it started showing stability issues again, browser tabs crashing, apps crashing. And today even a blue screen, immediately followed by another crash during the memory dump triggered by that blue screen, at 15%.

I applied all the Asus BIOS updates religiously, as they came out, did no overclocking (except XMP), so, unless I was extremely unlucky, it seems Raptor Lake is fundamentally flawed and unfixable.

At this point, after almost 2 years of having continuous issues with Raptor Lake, enough is enough. I will probably switch to the 9800 X3D as soon as it is released. Meanwhile, I'll have to start cutting the CPU frequency again.

Your story seems quite extraordinary and is probably the only case I have read about where a replacement has had the same problem post bios fixes.

Since you already decided to move of the platform, before another RMA, consider offering to sell it to someone like GN, who are known to buy failed parts to get content out of it, and it would also act as validation the chip is the problem.

Personally my 13700k has probably been one of my most stable builds, I had no trouble under volting, havent get any sign of instability at all, and only reduced the under volt out of caution rather than as a reaction to anything. It solved all my i/o corruption issues I was having on my 9900k setup (that ended up been a dodgy ram timing though, that I didnt pick up on until I started playing around with post upgrade).