"Something" tells me Intel has a good chance of pulling market availability back to "sooner" status :)

That picture literally listed some PR BS they are going to do.

Expect 'Intel Leaks" everyday starts on 15th Sep :)

CrackongThat picture literally listed some PR BS they are going to do.

Expect 'Intel Leaks" everyday starts on 15th Sep :)

Let's just hope Raptor Lake isn't a hose job such as AMD's 7600X @ $300 USD.

The price of the 7950x is a preemptive move by AMD, I hope Intel has enough stock of 13th leading to pricing war.

Info "leaks" just on the day of AMD's presentation. Intel NDA lifts just on the day AMD chips hit the shelfs.

They sure :love: playing their games, lol.

MarsM4NInfo "leaks" just on the day of AMD's presentation. Intel NDA lifts just on the day AMD chips hit the shelfs.

They sure :love: playing their games, lol.

Compare the shear amount of AMD leaks and Intel 'leaks'.
I am pretty sure everyone and his dog knows where all the intel 'leaks' came from.

Otherwise Intel would have the worse NDA practices on the planet.

Can't wait to see how they overheat.

GoldenXCan't wait to see how they overheat.

AMD you mean? I mean it seems they all need undervolting/tweaking to keep them from running too hot at stock speed....

P4-630AMD you mean? I mean it seems they all need undervolting/tweaking to keep them from running too hot at stock speed....

Intel is overheating my boy ;)

usinameIntel is overheating my boy ;)

Don't even know what load and what cooler used there...

Well my i7 12700K Air cooled benchmarked on a warm summer day: 74C max.... ;) Oh and no tweaked voltage or whatever....

P4-630Don't even know what load and what cooler used there...

Well my i7 12700K Air cooled benchmarked on a warm summer day: 74C max.... ;) Oh and no tweaked voltage or whatever....

My 12900k hits 76c on a u12a. Whats your cooler?

usinameIntel is overheating my boy ;)

Seems like user error.

Most leaks show little to no IPC increase for Raptor Lake. That will be the focus when it launches and hopefully prove the leaks wrong.

fevgatosMy 12900k hits 76c on a u12a. Whats your cooler?


Seems like user error.

I doubt that @W1zzard made those in an important review like that ;)
Anyway from the article that graph is from: We use a Noctua NH-U14S and measured the CPU temperature while running Blender.

ncrsI doubt that @W1zzard made those in an important review like that ;)
Anyway from the article that graph is from: We use a Noctua NH-U14S and measured the CPU temperature while running Blender.

Oh he did plenty regarding the alderlake reviews. His power limited numbers are all flawed for example.

fevgatosOh he did plenty regarding the alderlake reviews. His power limited numbers are all flawed for example.

That's interesting, care to elaborate?

fevgatosWhats your cooler?

U12A ;)

MarsM4NInfo "leaks" just on the day of AMD's presentation. Intel NDA lifts just on the day AMD chips hit the shelfs.

They sure :love: playing their games, lol.

Like always, they can play their discount and rebate card :D

ncrsThat's interesting, care to elaborate?

Well, a 12900k at 125w power limit scores 24 to 24.5k in cbr23. In tpus testing it scores 18k. Which doesnt make any sense, since that's how much a 12600k scores at same wattage, something that is literally impossible. The same pretty much applies to the rest of his testing, the power limited numbers are all off, some of them even by up to 70%

usinameIntel is overheating my boy ;)

ADL runs surprisingly cool if you disable HT, and you will easily afford to do so on RPL with all the e-cores to do throughput tasks.
Very yummy, honestly!

With HT enabled - yeah, those P-Cores are ovens unless you run a reasonable voltage/frequency, heehee

Why would you disable HT? Well, you know - the usual SKIZZO omgomg le input latency xddd run 1t memoery low rtl itx board whoah whao hw ahohhhh thingymabob.
Also the heat thing, I guess.

fevgatosWell, a 12900k at 125w power limit scores 24 to 24.5k in cbr23. In tpus testing it scores 18k. Which doesnt make any sense, since that's how much a 12600k scores at same wattage, something that is literally impossible. The same pretty much applies to the rest of his testing, the power limited numbers are all off, some of them even by up to 70%

I am assuming you mean this test:

Since 12600K's default configuration is PL1=125W, PL2=228W and turbo power of 150W it makes sense to me that a 12900K limited to 125W on all is only slightly faster while having 2 more P-cores, and the relatively small PL2 time with 241W limit also increases the score. Or am I reading this wrong?

ncrsI am assuming you mean this test:

Since 12600K's default configuration is PL1=125W, PL2=228W and turbo power of 150W it makes sense to me that a 12900K limited to 125W on all is only slightly faster while having 2 more P-cores, and the relatively small PL2 time with 241W limit also increases the score. Or am I reading this wrong?

You are reading it wrong. On the last page there is power draw during test. The 12600k consumed the same amount as tje 12900k at 125w. Also the 12900k has 2 p cores and 4 e cores more..

All that is irrelevant though, cause both me and other reviewers have actually tested the 12900k (or even the 12700) and got way different results. For example, techspot tested a 12700 non k at 65w and it scored 16k. Club365 and igorslab tested the 12900k at 125w and it basically scored 24k in cbr23 and was also faster than the 5900x in blender. Ive personally tested 3 12900ks on 4 different mobos, all of them score between 23.5 and 24.5k.

Tpus results are without a doubt wrong. And the further down you go in the power limit numbers the more wrong they are. At 75w for example it should have scored around 18k, he has it at 11k, lol.

fevgatosYou are reading it wrong. On the last page there is power draw during test. The 12600k consumed the same amount as tje 12900k at 125w. Also the 12900k has 2 p cores and 4 e cores more..

All that is irrelevant though, cause both me and other reviewers have actually tested the 12900k (or even the 12700) and got way different results. For example, techspot tested a 12700 non k at 65w and it scored 16k. Club365 and igorslab tested the 12900k at 125w and it basically scored 24k in cbr23 and was also faster than the 5900x in blender. Ive personally tested 3 12900ks on 4 different mobos, all of them score between 23.5 and 24.5k.

Tpus results are without a doubt wrong. And the further down you go in the power limit numbers the more wrong they are. At 75w for example it should have scored around 18k, he has it at 11k, lol.

Poorly configured DC loadline might be the cause? That's the only thing I can think of that would explain the results as I'm pretty sure the PLs are based on power package so if that number's wrong you'll throttle too early or too late.

Not sure why you'd need to touch LLCs for testing power limits though.

fevgatosYou are reading it wrong. On the last page there is power draw during test. The 12600k consumed the same amount as tje 12900k at 125w. Also the 12900k has 2 p cores and 4 e cores more..

All that is irrelevant though, cause both me and other reviewers have actually tested the 12900k (or even the 12700) and got way different results. For example, techspot tested a 12700 non k at 65w and it scored 16k. Club365 and igorslab tested the 12900k at 125w and it basically scored 24k in cbr23 and was also faster than the 5900x in blender. Ive personally tested 3 12900ks on 4 different mobos, all of them score between 23.5 and 24.5k.

Tpus results are without a doubt wrong. And the further down you go in the power limit numbers the more wrong they are. At 75w for example it should have scored around 18k, he has it at 11k, lol.

Thanks for the explanation.
Were those test you reference done as early on as TPU's? Is an un-optimized BIOS or microcode a possible culprit?

napataPoorly configured DC loadline might be the cause? That's the only thing I can think of that would explain the results as I'm pretty sure the PLs are based on power package so if that number's wrong you'll throttle too early or too late.

Not sure why you'd need to touch LLCs for testing power limits though.

Yea, either dc ac lls that cause this or fixed voltage

ncrsThanks for the explanation.
Were those test you reference done as early on as TPU's? Is an un-optimized BIOS or microcode a possible culprit?

No they werent as early except igors lab, but it is highly unlikely its caused by bios. Most likely wrongly configured lls or fixed voltage

fevgatosWell, a 12900k at 125w power limit scores 24 to 24.5k in cbr23. In tpus testing it scores 18k. Which doesnt make any sense, since that's how much a 12600k scores at same wattage, something that is literally impossible. The same pretty much applies to the rest of his testing, the power limited numbers are all off, some of them even by up to 70%

You have to consider the designs as a whole. Less E cores in the design will mean better efficiency by extension relative to P cores. Round 1 goes to the 12600K on efficiency. You have to consider turbo frequencies and base frequencies of each chip, but it's important to also look those variables for both E cores and P cores of each CPU in question. The 12900K boosts higher on E cores and P cores, but with lower base clocks on E cores and P cores relative to the 12600K. That's not really favorable either since higher frequency is going to be better in a power limit test comparing peak multi-thread CPU utilization. Round 2 therefore goes to the 12600K once again. The 12900K is better for performance scores to make it look like it's punching above it's weight more than it really is if efficiency is important to you which with rising electrical costs and environmental aspects should be a bit, but isn't to everyone in any case it's adequately fair criticism even if it doesn't apply to you personally.

I know you think W1zzard was wrong because the score isn't what you want, but perhaps trust the guys testing methodologies a bit he's ran a few benchmarks here and there. I don't believe he was intentionally trying to make the+ 12900K look inferior to the 12600K if efficiency is important to someone, but guess what the results are pretty cut and dry he just put random numbers in a hat to pull out and say this was the score for each. He's a W1zzard not a M4ggician. I don't know why you think adding more E cores will automatically improve efficiency when they were designed for MT uplift and have lower IPC than P cores. Either way it's a hot take you'd think it so.