Question Alder Lake - Official Thread

Page 106 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
25,380
14,344
136
  • Like
Reactions: Drazick

Exist50

Platinum Member
Aug 18, 2016
2,445
3,043
136
Guys, Mark has had cancer. Give him a break and let him be.
So because he's had cancer, we should give him a pass on lying about a CPU just because he (by his own admission) hates the company in question? Again, let's be perfectly clear that this isn't just a matter of different priorities and preferences. His most recent example showed quite clearly that he doesn't care what the data is from, just that it shows AMD winning. And frankly, do we have any reason to believe his stance on power consumption isn't similar?



Your post is nothing but being inflammatory and attacking another member.


esquared
Anandtech Forum Director
 
Last edited by a moderator:
  • Like
Reactions: SMU_Pony

Tech Junky

Diamond Member
Jan 27, 2022
3,269
1,085
106
If you're not contributing to the solution of providing datapoints you're arguing about the data provide by bogus benchmarks that aren't measuring the data correctly in the first place. Getting an accurate and full scope of the HW is still a bit of a mess as most programs used in the past don't have the proper capabilities for hybrid arc's at this point.

Bashing either AMD / Intel doesn't prove a point. They both have their pro's / cons when it comes to how they perform.

Being able to bifurcate PCI lanes on AMD is a perk of you're going to use a riser to mount additional drives in a x16 slot
Intel has niche applications when it comes to things like Thunderbolt they price other companies out of with licensing royalties,

TB4 / USB4 get rid of the costs associated with licensing and open the doors to other options.

When it comes to power consumption it's a bit futile since systems are all designed with different components and it's not going to be apples to apples no matter how closely you try to match things. When you change CPU's it's all out the window.

The first PC I ever built was an AMD and now I run Intel because it's cheaper for similar specs. A dream machine with unlimited budget though would be one of those AMD beasts with 100+ threads just to play with it and see what's possible. I would never be able to max it out though with traditional use cases and it's completely overkill outside of a production environment.

Right now I'm looking at a move from 9750H to 12700H on the laptop side just to play around with the new tech in a different format to see how they compare between desktop / laptop iterations. Also, the boost from 9th to 12th will be nice. The only disappointment though is that no Laptop ADL will support NVME Gen 5 unlike the desktop side which has an open slot waiting for something fun to play with.
 
  • Like
Reactions: lightmanek

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
25,380
14,344
136
So because he's had cancer, we should give him a pass on lying about a CPU just because he (by his own admission) hates the company in question? Again, let's be perfectly clear that this isn't just a matter of different priorities and preferences. His most recent example showed quite clearly that he doesn't care what the data is from, just that it shows AMD winning. And frankly, do we have any reason to believe his stance on power consumption isn't similar?
I have not lied anywhere. Sow me where. The fact that I may have missed something in reading, and then pointed it out myself later certainly is not lying.
 
  • Like
Reactions: Drazick

Exist50

Platinum Member
Aug 18, 2016
2,445
3,043
136
I have not lied anywhere. Sow me where. The fact that I may have missed something in reading, and then pointed it out myself later certainly is not lying.
For example, claiming that AMD wins "almost everything" at 142W (because apparently we're just supposed to know that "everything" means "embarrassingly parallel compute tasks"), claiming that Alder Lake takes 240W+ in gaming, and your recent false claims about the benchmarks in that Tom's article.

And since you were still doubling down on those results, are you going to admit they're wrong, or will you insist that Golden Cove is worse than than Skylake? Because it's blatantly obvious to everyone that you pulled the worst result you could find for Alder Lake, without even reading the context or applying an iota of common sense.

Honestly, it gets tiresome to point out your own tactics. You have to realize what you're doing, and thus are just making a mockery of discussion here. If you just want to talk about how much you love your 5950x, why not leave it in the AMD thread?
 

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
25,380
14,344
136
For example, claiming that AMD wins "almost everything" at 142W (because apparently we're just supposed to know that "everything" means "embarrassingly parallel compute tasks"), claiming that Alder Lake takes 240W+ in gaming, and your recent false claims about the benchmarks in that Tom's article.

And since you were still doubling down on those results, are you going to admit they're wrong, or will you insist that Golden Cove is worse than than Skylake? Because it's blatantly obvious to everyone that you pulled the worst result you could find for Alder Lake, without even reading the context or applying an iota of common sense.

Honestly, it gets tiresome to point out your own tactics. You have to realize what you're doing, and thus are just making a mockery of discussion here. If you just want to talk about how much you love your 5950x, why not leave it in the AMD thread?
Again, you lie. I never said Alder lake takes 240w+ in gaming. And as for the other, it was win 10 vs win 11, and I pointed that out. You need to stop accusing me, and start READING.
 
  • Like
Reactions: Drazick

Exist50

Platinum Member
Aug 18, 2016
2,445
3,043
136
Again, you lie. I never said Alder lake takes 240w+ in gaming.
In response to someone looking for a CPU for gaming, you told them not to buy Alder Lake because it consumes 241W.

And as for the other, it was win 10 vs win 11, and I pointed that out.
1) You only pointed that out after quoting the result as representative.

2) You directly claimed that the W11 suite was using different (and somehow lesser) benchmarks, which was your stated justification for using the W10 data. That is objectively false. You clearly still refuse to acknowledge that.

3) You continued (and continue) to double down on those numbers after several people have pointed out how ridiculous they are.

And that's on top of a long legacy of this sort of behavior in this and related threads. Pretending it's mere coincidence just rings laughably hollow.
 

epsilon84

Golden Member
Aug 29, 2010
1,142
927
136
To be honest, let's not pretend like that's a huge win for the 12900K. It's squaring off against a CPU that was a year old when the 12900K came out, and not winning everything. I think Mark does exaggerate the situation a bit, though I will also point out that most of those reviews do not focus on fixed power limits. @142W the 5950X probably does win a majority of MT benchmarks over the 12900K.
Where did I pretend that was a 'huge win' for the 12900K? Mark exagerrates the situation a bit? LOL thats like the understatement of the century.

Let's reiterate what his claims were:

Either way its an act of desperation. The 5950x wins in almost everything except gaming at 142 watts. This monster does at least 241 watts or more and still looses most other than gaming(well, the 12900k+4%). And the 5800X3D will take care of that. So it makes it useless, and expensive. The P core is strong, and for single threaded apps, it is great. But consider this, if you are a gamer ? the 5800X3D will be the thing to get, and if you are a productivity person, the 5950x will be the thing to get. So that makes the single P core use case pretty niche.

So winning 'almost everything' equates to 45% of benchmarks? That's against a 12900K no less, it's probably closer to a 60/40 split with the 12900KS, not that I really care what the exact %s are. My point is that is a gross exaggeration on marks part claiming the 5950X is almost a clean sweep win in non gaming benchmarks. Calling it 'useless' is by extension calling the 12900K useless as well, since they are basically the same CPU minus the binning. Expensive? Sure. But useless?

Imagine me going to the 5800X3D thread and calling it 'expensive and useless' outside of gaming because its $450 and would probably lose to a $350 12700K in almost anything except gaming benchmarks.

I have no issue with factual debates based on reliable data points, but I'll continue to call out blatant bias and unsubstantiated claims, especially in this thread. I have no idea why AMD is mentioned so much here, frankly, apart from a clear agenda from a certain someone who likes to insist that Alder Lake is only good for tasks that scale up to 8C/16T and is 'useless' beyond that.




This is nothing but thread derailing.


esquared
Anandtech Forum Director
 
Last edited by a moderator:
  • Like
Reactions: SMU_Pony

Exist50

Platinum Member
Aug 18, 2016
2,445
3,043
136
How is it different? Dude... Seriously.
Well, he almost immediately proceeded to prove my point by insisting that the 12900k ~= 11900K in MT workloads. If that doesn't convince you that he has an agenda, then what would?




Continuing to derail the thread.


esquared
Anandtech Forum Director
 
Last edited by a moderator:
  • Like
Reactions: SMU_Pony

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
25,380
14,344
136
Well, he almost immediately proceeded to prove my point by insisting that the 12900k ~= 11900K in MT workloads. If that doesn't convince you that he has an agenda, then what would?
Insisting ? I posted a link to tomshardware, and when it was pointed out there was an inconsistancy, I investigated, and found the reason and reported it. Troll.
 
  • Like
Reactions: Drazick

Exist50

Platinum Member
Aug 18, 2016
2,445
3,043
136
I posted a link to tomshardware, and when it was pointed out there was an inconsistancy, I investigated, and found the reason and reported it.
For now the third time, you've called those results "not erroneous", and still haven't admitted to lying when you claimed the W10 results used a different and better benchmark suite than the W11. There was nothing in the article to suggest that, yet you claimed it anyway as an excuse for cherry picking the W10 results. It's honestly sad how you still cannot admit to being wrong about that, even if you avoid acknowledging that you did so deliberately.

And pretty funny to call me a troll for pointing out your incessant attempts to derail this thread and mislead people, up to and including lying to make your preference seem justified.
 

esquared

Forum Director & Omnipotent Overlord
Forum Director
Oct 8, 2000
23,504
4,717
146
The topic for this thread is:
Alder Lake - Official Thread

Posting inflammatory posts and harassing members and anything else not related to the
topic at hand is going to stop. Now.


esquared
Anandtech Forum Director
 
Last edited:

JoeRambo

Golden Member
Jun 13, 2013
1,814
2,105
136
PSA: Intel has released a new version of their XTU utility, and it seems it is actually mostly working this time. I think it matches 1:1 what is selected in BIOS now, where previuosly variuos voltages and other items were showing up incorrect.

I have thought it was due to my MSI Edge D4 Bios being shoddy, but it turns out XTU was actually buggy itself.
 

DrMrLordX

Lifer
Apr 27, 2000
21,528
10,712
136
Where did I pretend that was a 'huge win' for the 12900K?

Maybe I understated my point then. It's barely a win for the 12900k.

Let's reiterate what his claims were:

How about you look at the performance of the 5950X vs 12900k in the same power envelope and then get back to us on what his claims were.

Anyway I think these points are repetitive and bordering on the useless since both AMD and Intel are poised to move on to other products. Alder Lake, for the brief window where it will remain relevant, will represent a compromise between the 5800X3D and 5950X, with better gaming performance than the 5950X and clearly better MT performance than the 5800X3D. The downside being that, sometimes, you will need serious cooling to help the thing reach its preprogrammed power targets. Or you can just lower those targets and lose some MT performance.
 

JoeRambo

Golden Member
Jun 13, 2013
1,814
2,105
136
My view of 5950x vs 12900K ( as owner of both ) and coming from enthusiast point of view, you know where you actually take time to tune memory, undervolt and underclock CPUs, find efficiency points, make sensible tradeoffs, undo damage inflicted by marketing departments etc

1) my 5950X is UNDISPUTED king of very efficient multi threaded performance. One needs to disable AMDs retarded boost "algorithms" and clock it with static voltage and at static frequency. But 16C of 4.4Ghz ZEN3 performance at incredible efficiency, great thermals, linear curve of power use over whole load range ( as in not letting retards from AMD marketing to chew 50W at 3 cores worth of load ). Golden stuff.
There is no way i can tune, configure or massage 12900K into so wide perf capacity at so great efficiency. E-cores are slow, hybrid scheduling is bad, 8C of GC even at efficient voltage/clock would provide fraction of capacity of 5950x in what it does.

now fine print of (1) is, that i am not using that system at all, it chugs work alone, accessed remotely. That brings us to reason why in (2)

2) my 12900K is UNDISPUTED king in performance and efficiency of every day use that does not require very efficient multi threaded performance. So basically everything normal people who do not render or run DC on their computers. And i mean everything, gaming, browsing web, office work, using whatever program they are using like some IDE or graphic design or CAD stuff.
fine print is obviously you have to prepare it by disabling E-Cores, undoing turbo boost to the point where it looses efficiency, using static clock and volts for cores and uncore etc. For me that point is 5Ghz / 4.5Ghz uncore / 3800C15 D4 ram

The problem with backing up my words is that benchmarks for smoothness are hard to come by. For example ADL might have 15% advantage in GB5 ST, but large part of it is irrelevant junk like encryption etc. CB23 might show big gains in rendering and advantage might be 30%+ in web benchmarks like Speedometer 2.0,. but that also represents just mostly performance of JavaScript engine.
Things that involve my work stuff, like loading projects in IDE, compiling, building etc feels way smoother on ADL. It is hard to quantify, as even if build takes 30% less on ADL, it might be due to I/O differences or some arcane settings. But what is timed, shows at least 25% advantage for ADL.
Gaming is also a forte of ADL. Games like Anno1800 never felt smooth in late game on Z3 and ADL is literally chewing through Last Epoch that is full of level loads every several minutes. Sure both keep 60FPS locked on 3090, but ADL is loading way faster. Is there a way to do scientific comparison? Probably not and i can't be bothered to move my 3090 again.

There is no way i can tune 5950x into beating 12900K in these workloads, i might gain 10% more clock at cost of MT efficiency, but then again 12900K can be tuned to 5.4+ghz low threaded, keeping the gap the same.
 

epsilon84

Golden Member
Aug 29, 2010
1,142
927
136
How about you look at the performance of the 5950X vs 12900k in the same power envelope and then get back to us on what his claims were.

Anyway I think these points are repetitive and bordering on the useless since both AMD and Intel are poised to move on to other products. Alder Lake, for the brief window where it will remain relevant, will represent a compromise between the 5800X3D and 5950X, with better gaming performance than the 5950X and clearly better MT performance than the 5800X3D. The downside being that, sometimes, you will need serious cooling to help the thing reach its preprogrammed power targets. Or you can just lower those targets and lose some MT performance.
Why is there a need to move the goalposts all of a sudden and match power usage? By design the 12900K TDP is 241W, though yes you can artificially impose a 125W TDP in the BIOS and you are correct, the 5950X would then win virtually every MT benchmark, though gaming and ST performance would remain largely unaffected at 125W.

Will AMD fans also ask for the 3090 Ti be capped at the 300W instead of 450W so their 6900XT wins more gaming benchmarks? You can see where I'm going with this, right?

If you go back to one of my earlier posts, I already said that I didn't care who came out on top in gaming (would be a few % either way) between the 5800X3D and 12900K/KS because in a few months they will be irrevlant and superceded anyway by faster and cheaper next gen chips, so I agree with you there. I think both are overpriced halo products, albeit in different categories, but they will inevitably be compared anyway in the all important 1080P gaming tests ;)

Btw, I somewhat disagree with this:
The downside being that, sometimes, you will need serious cooling to help the thing reach its preprogrammed power targets.

Unless we are classing a 280mm AIO as 'serious cooling' now? Because thats what I'm using on my 12900K and it is more than enough to cool it. You can realistically get away with using a 240mm AIO or a strong HSF, especially if you know how to do a simple 50mV undervolt in the BIOS which brings the power usage down significantly for no perfomance penalty. But I'll concede that for a stock 12900K, pegged at 241W, a larger AIO (280mm or 360mm) is preferred to keep temps in check.
 
Last edited:

DrMrLordX

Lifer
Apr 27, 2000
21,528
10,712
136
Why is there a need to move the goalposts all of a sudden and match power usage?

It was not sudden. It's been repeated multiple times, though people have a tendency to jump all over the guy who keeps bringing it up.

Unless we are classing a 280mm AIO as 'serious cooling' now?

Yes. There are many cases that aren't compatible with AiOs at all, or have awkward mounts. At least the prices are coming down, though I question the reliability of units that can be sold for $65.

In any case, an AiO represents the most powerful cooling you can buy short of custom water (or more exotic), and even then not all AiOs are going to handle 241W. Anything running 150W or less can still be tamed by a wide variety of tower HSFs.
 
Last edited:
  • Like
Reactions: Markfw

epsilon84

Golden Member
Aug 29, 2010
1,142
927
136
It was not sudden. It's been repeated multiple times, though people have a tendency to jump all over the guy who keeps bringing it up.

Yes. There are many cases that aren't compatible with AiOs at all, or have awkward mounts. At least the prices are coming down, though I question the reliability of units that can be sold for $65.

In any case, an AiO represents the most powerful cooling you can buy short of custom water (or more exotic), and even then not all AiOs are going to handle 241W. Anything running 150W or less can still be tamed by a wide variety of tower HSFs.
That's because as soon as you 'equalise' the TDP you are artificially handicapping the 12900K. Not sure how much more simple I can make this.

Btw, if AIOs aren't your thing, there are plenty of air coolers that can sufficiently cool a 12900K, so I really don't know where you are going with this argument:


FWIW, as a 12900K owner, I'm not suggesting that it doesn't run hot, especially when left at stock with no undervolting applied etc. But with a bit of common sense with regards to AIO or HSF selection, its really not as hard as you make it out to be.
 
Last edited:

Zucker2k

Golden Member
Feb 15, 2006
1,810
1,159
136
Well, we do know mobo makers are overvolting boards in order to win performance comparisons as well, and that also needs to be taken into consideration. Igor's is the only website looking into these things.

Link

And apparently, even the power reporting of the Ryzen platform is inaccurate. Wall numbers are the better representative power numbers. This, according to Chips and Cheese.
 
  • Like
Reactions: Mopetar

DrMrLordX

Lifer
Apr 27, 2000
21,528
10,712
136
That's because as soon as you 'equalise' the TDP you are artificially handicapping the 12900K. Not sure how much more simple I can make this.

Yes, but why is that an invalid commentary? Whose fault is it that the 12900k requires up to 241W to follow its default power spec, or more if the motherboard manufacturer decides to meddle with P1/P2 values?

Btw, if AIOs aren't your thing, there are plenty of air coolers that can sufficiently cool a 12900K, so I really don't know where you are going with this argument:

They CAN, depending on what you do with one. If all you do is game, a 12900k may not cross 100W very often. If you game and stream in high resolutions, however, you will start to hit higher power usage. Please don't tell me you think $25 tower coolers with stock fans can handle ~240W, because even some AiOs can't deal elegantly with those loads. I was cranking out over 200W with my 1800x and an NH-D15 with custom fans (3000rpm Noctua industrialPPC) barely handled the load.

Intel has been schlepping 200W+ CPUs for awhile now, and we all ought to know what kind of cooling is or is not required to prevent throttling.
 

epsilon84

Golden Member
Aug 29, 2010
1,142
927
136
Yes, but why is that an invalid commentary? Whose fault is it that the 12900k requires up to 241W to follow its default power spec, or more if the motherboard manufacturer decides to meddle with P1/P2 values?

Because the moment you start changing parameters on a CPU and out of its stock configuration, you're just moving the goal posts to suit your agenda, and its not really an apples to apples stock vs stock comparison anymore.

Again, I re-iterate my point earlier, but on the GPU side: Do AMD fans ask for a 450W 3090 Ti to be power limited to 300W so to make it 'fair' against a 300W 6900XT? See how absurd that argument becomes? Yes, you *can* artificially cap a 3090 Ti to 300W, in which case it may well perform worse than a stock 6900XT. But good luck finding a reviewer to do that.

The same argument applies to the 12900K. The 5950X is more efficient under an all core workload, we all know that. The 12900K takes ~100W more to do roughly the same amount of work, but the advantages you get in return are stronger ST and gaming performance, and under those scenarios power efficiency isn't that much worse than Zen 3.

They CAN, depending on what you do with one. If all you do is game, a 12900k may not cross 100W very often. If you game and stream in high resolutions, however, you will start to hit higher power usage. Please don't tell me you think $25 tower coolers with stock fans can handle ~240W, because even some AiOs can't deal elegantly with those loads. I was cranking out over 200W with my 1800x and an NH-D15 with custom fans (3000rpm Noctua industrialPPC) barely handled the load.

Intel has been schlepping 200W+ CPUs for awhile now, and we all ought to know what kind of cooling is or is not required to prevent throttling.

A lot of large tower style coolers are rated for 250W usage, some are even below $50.

Have a look at the video I linked earlier, even the $40 Arctic Freezer 34 was enough to cool the 12900K without thermal throttling under full load, albeit with temps in the 90s. Not ideal if you want to constantly have the CPU under full load, but perfectly acceptable for someone who mainly games, for example.

The $80 Noctua NH-U14s is able to keep temps in the 80s, which is preferable for those who push the CPU harder. The NH-U14s isn't even the best Noctua cooler, so I'm sure an NH-D15 would fare slightly better still.
 

lobz

Platinum Member
Feb 10, 2017
2,057
2,856
136
My view of 5950x vs 12900K ( as owner of both ) and coming from enthusiast point of view, you know where you actually take time to tune memory, undervolt and underclock CPUs, find efficiency points, make sensible tradeoffs, undo damage inflicted by marketing departments etc

1) my 5950X is UNDISPUTED king of very efficient multi threaded performance. One needs to disable AMDs retarded boost "algorithms" and clock it with static voltage and at static frequency. But 16C of 4.4Ghz ZEN3 performance at incredible efficiency, great thermals, linear curve of power use over whole load range ( as in not letting retards from AMD marketing to chew 50W at 3 cores worth of load ). Golden stuff.
There is no way i can tune, configure or massage 12900K into so wide perf capacity at so great efficiency. E-cores are slow, hybrid scheduling is bad, 8C of GC even at efficient voltage/clock would provide fraction of capacity of 5950x in what it does.

now fine print of (1) is, that i am not using that system at all, it chugs work alone, accessed remotely. That brings us to reason why in (2)

2) my 12900K is UNDISPUTED king in performance and efficiency of every day use that does not require very efficient multi threaded performance. So basically everything normal people who do not render or run DC on their computers. And i mean everything, gaming, browsing web, office work, using whatever program they are using like some IDE or graphic design or CAD stuff.
fine print is obviously you have to prepare it by disabling E-Cores, undoing turbo boost to the point where it looses efficiency, using static clock and volts for cores and uncore etc. For me that point is 5Ghz / 4.5Ghz uncore / 3800C15 D4 ram

The problem with backing up my words is that benchmarks for smoothness are hard to come by. For example ADL might have 15% advantage in GB5 ST, but large part of it is irrelevant junk like encryption etc. CB23 might show big gains in rendering and advantage might be 30%+ in web benchmarks like Speedometer 2.0,. but that also represents just mostly performance of JavaScript engine.
Things that involve my work stuff, like loading projects in IDE, compiling, building etc feels way smoother on ADL. It is hard to quantify, as even if build takes 30% less on ADL, it might be due to I/O differences or some arcane settings. But what is timed, shows at least 25% advantage for ADL.
Gaming is also a forte of ADL. Games like Anno1800 never felt smooth in late game on Z3 and ADL is literally chewing through Last Epoch that is full of level loads every several minutes. Sure both keep 60FPS locked on 3090, but ADL is loading way faster. Is there a way to do scientific comparison? Probably not and i can't be bothered to move my 3090 again.

There is no way i can tune 5950x into beating 12900K in these workloads, i might gain 10% more clock at cost of MT efficiency, but then again 12900K can be tuned to 5.4+ghz low threaded, keeping the gap the same.
I think that is absolutely how things logically should be, so we can only be glad when we can verify that they actually are 😂
Intel's next-gen vs AMDs latest but somewhat aged gen.
Can't wait to see what happens soon when AMD jumps 1.5 gens (I mean as a whole: uarch, node, platform etc.) and Intel follows them arriving at their 1.3 gen, battle of the titans 😂
 
  • Like
Reactions: lightmanek

coercitiv

Diamond Member
Jan 24, 2014
6,108
11,515
136
Because the moment you start changing parameters on a CPU and out of its stock configuration, you're just moving the goal posts to suit your agenda
The stock configuration for the 12900K is 125W TDP, Intel says so very clerly in their 12th Gen Datasheet:
1649145049573.png

They're also giving recommendations for Tau timers that limit time spent by the CPU @ PL2=241 in order for the CPU to actually behave like a 125W TDP CPU in the long-run:
1649145209113.png

So if your intent is to advocate that the 12900K be judged at the peak of its performance then by all means, don't let me interrupt. But please stop promoting the idea that stock config for 12900K is 241W TDP, it's even more ludicrous to say that considering the 12900KS is defined by Intel as a 150W TDP SKU.