Question Alder Lake - Official Thread

[Hulk]

With the release of Alder Lake less than a week away and the "Lakes" thread having turned into a nightmare to navigate I thought it might be a good time to start a discussion thread solely for Alder Lake.

 

[DooKey]

Has anyone here experienced bent sockets on their Z690 boards yet?

No

 

[IntelUser2000]

Not sure about that, at least when it comes to i7-1165G7. It was 65 degrees without HT but with HT, it was hitting 95 degrees in the TPU ThrottleStop stress test.

Thermals can be very different from power consumption. HT provides 30% gain with about the same power use. It's challenging to achieve 1:1 ratio of performance improvement for power increase. Plus it requires next to no increase in die area. On a core level maybe it's 3-5% but on the chip level it's going to be in the 1-2% range.

When it comes to thermals though, you can have a scenario where the device can use just 5% more power than the maximum capability of the heatsink and the temperatures will start to increase quite dramatically.

It also increases heat density as the same die size is using 30% more power.

This is different from caches where it's very low in terms of heat density.

 

[jpiniero]

Intel Core i9-12900K Alder Lake Tested at Power Limits between 50 W and 241 W

We test Intel's Core i9-12900K at various TDP levels all the way down to 50 W to determine how much efficiency is really in the new Alder Lake core, and how these power limits affect performance. Competing with the efficiency of AMD's Zen 3 Ryzen lineup is just two settings changes away.

www.techpowerup.com

Benchmarks at various power levels of the 12900K. All of the titles they chose there wasn't much of a difference between 241/241 and 125/125, but it can drop from there. 50/50 is super slow in a bunch of games.

 

[Hulk]

SMT gains are an indirect measurement of how much of the CPU is sitting idle. It is great to be able to recover that performance for applications that can benefit. But it is also a measure of how overengineered the chips are and not all applications benefit from it. In an ideal world, both Intel and AMD would both be simply 30% faster and not need SMT to recover that lost performance.

When Alder Lake went wider and deeper on many of its components, that meant that there is more excess capacity. More SMT performance is to be expected.

Also keep in mind that as IntelUser2000 informed me in the Raptor Lake thread that while there is more capacity when going wider and deeper there is also generally better OoO, branch prediction, etc.. with the newer parts, which increases ST performance but would decrease HT. So it's not always a definite that HT/SMT performance with increase with more resources. There are other factors at work.

 

[tomatosummit]

Intel Core i9-12900K Alder Lake Tested at Power Limits between 50 W and 241 W

We test Intel's Core i9-12900K at various TDP levels all the way down to 50 W to determine how much efficiency is really in the new Alder Lake core, and how these power limits affect performance. Competing with the efficiency of AMD's Zen 3 Ryzen lineup is just two settings changes away.

www.techpowerup.com

Benchmarks at various power levels of the 12900K. All of the titles they chose there wasn't much of a difference between 241/241 and 125/125, but it can drop from there. 50/50 is super slow in a bunch of games.

The 50w results just look broken all over the place, including the rendering results.
Some of the application results even drop down to near the 11400 and 10400 levels. Although nothing in the article suggests if those cpus are running at 65w or fully boosted.
Hopefully that's not representative of the coming mobile dies although the 8+8die is bring used in the h55 cpu.

 

[Ed1]

Some gaming performance per watt.

 

[Ed1]

The 50w results just look broken all over the place, including the rendering results.
Some of the application results even drop down to near the 11400 and 10400 levels. Although nothing in the article suggests if those cpus are running at 65w or fully boosted.
Hopefully that's not representative of the coming mobile dies although the 8+8die is bring used in the h55 cpu.

something is wrong with his Wprime scores, I ran 1024M one and I get 79.911 sec with 12600k.
The program default to only 4 core count in its settings, not sure if that was from my old 3570k or not as this second drive is from old system.
I also ran it with only the 4 core count and got 223.058 sec so that doesn't match his results either.

Edit: also ran Super PI 1.5mod, 32M and got 351.399 sec

 

[TESKATLIPOKA]

Intel Core i9-12900K Alder Lake Tested at Power Limits between 50 W and 241 W

We test Intel's Core i9-12900K at various TDP levels all the way down to 50 W to determine how much efficiency is really in the new Alder Lake core, and how these power limits affect performance. Competing with the efficiency of AMD's Zen 3 Ryzen lineup is just two settings changes away.

www.techpowerup.com

Benchmarks at various power levels of the 12900K. All of the titles they chose there wasn't much of a difference between 241/241 and 125/125, but it can drop from there. 50/50 is super slow in a bunch of games.

I found something very interesting.
i9-12900K limited to 50W for both PL1 and PL2 provides only 8872 points in CB R23 and the power consumption is 120W.
If you check Intel Core i9-12900K E-Cores Only Performance Review then you will find out that i9-12900K with only E-cores enabled provides 10366 points in CB R23 and consumes 118W.
You gain ~17% higher score and power consumption is actually 2W lower.

The other interesting thing is the actual clockspeed of this i9-12900K limited to 50W.
If i9-12900K's P-cores work at 4.9GHz and E-cores work at 3.7GHz with PL1=PL2=241W and generate 27780 points in CB R23 then PL1=PL2=50W and score of 8872(or ~32% of Full performance at 241W) would mean P-cores are working at measly ~1568Mhz and E-cores at 1184MHz.

 

[dullard]

I found something very interesting.
i9-12900K limited to 50W for both PL1 and PL2 provides only 8872 points in CB R23 and the power consumption is 120W.
If you check Intel Core i9-12900K E-Cores Only Performance Review then you will find out that i9-12900K with only E-cores enabled provides 10366 points in CB R23 and consumes 118W.
You gain ~17% higher score and power consumption is actually 2W lower.

I don't think that we should be very surprised that Alder Lake E cores are better than the P cores when power is limited. That is basically the whole point behind the hybrid CPU design and why mobile chips don't have 8 P-cores.

 

[TESKATLIPOKA]

I don't think that we should be very surprised that Alder Lake E cores are better than the P cores when power is limited. That is basically the whole point behind the hybrid CPU design and why mobile chips don't have 8 P-cores.

Just E-cores are better than a combination of P+E cores not just in efficiency but also in performance at 50W power limit based on TPU reviews.
Is that really not surprising?
Mobile ADL has 2-6 less P-cores, but also lower power limit and the leaked score for 6C+8c Alder Lake-P Core i7-12700H(TDP: 35-45W) was supposedly 18500 points in CB R23, Notebookcheck. Not sure what was Its actual package power.
You should also check the rest of my post about clockspeed.
50W for 16 cores(8C+8c) is 2.75-3W per core on average. Those clocks I calculated are low for that power consumption in my opinion, If not for P-cores then for E-cores for sure.
I have to wonder If the voltage didn't actually stay at much higher values than needed and because of that was the clockspeed and performance low.

 

[tomatosummit]

Just E-cores are better than a combination of P+E cores not just in efficiency but also in performance at 50W power limit based on TPU reviews.
Is that really not surprising?
Mobile ADL has 2-6 less P-cores, but also lower power limit and the leaked score for 6C+8c Alder Lake-P Core i7-12700H(TDP: 35-45W) was supposedly 18500 points in CB R23, Notebookcheck. Not sure what was Its actual package power.
You should also check the rest of my post about clockspeed.
50W for 16 cores(8C+8c) is 2.75-3W per core on average. Those clocks I calculated are low for that power consumption in my opinion, If not for P-cores then for E-cores for sure.
I have to wonder If the voltage didn't actually stay at much higher values than needed and because of that was the clockspeed and performance low.

You can't make a particulary accurate estimation of the clock speeds as the different cores will have different power to clock scaling.
Although I think the motherboard isn't playing well at 50w. There are going to be low wattage variants of this die (12900T and the H55 cpus) and the performance drop of here looks like an aberration. If this is really how it performs at 50w then I worry for the 6+8 mobile cpus, 8800 in r23 is well below tiger-h and cezanne-h.

 

[jpiniero]

You can't make a particulary accurate estimation of the clock speeds as the different cores will have different power to clock scaling.
Although I think the motherboard isn't playing well at 50w. There are going to be low wattage variants of this die (12900T and the H55 cpus) and the performance drop of here looks like an aberration. If this is really how it performs at 50w then I worry for the 6+8 mobile cpus, 8800 in r23 is well below tiger-h and cezanne-h.

I wonder how much undervolting room is there.

 

[dullard]

Just E-cores are better than a combination of P+E cores not just in efficiency but also in performance at 50W power limit based on TPU reviews.
Is that really not surprising?
Mobile ADL has 2-6 less P-cores, but also lower power limit and the leaked score for 6C+8c Alder Lake-P Core i7-12700H(TDP: 35-45W) was supposedly 18500 points in CB R23, Notebookcheck. Not sure what was Its actual package power.
You should also check the rest of my post about clockspeed.
50W for 16 cores(8C+8c) is 2.75-3W per core on average. Those clocks I calculated are low for that power consumption in my opinion, If not for P-cores then for E-cores for sure.
I have to wonder If the voltage didn't actually stay at much higher values than needed and because of that was the clockspeed and performance low.

Intel has always stated that the E cores were for multithreading and the P cores were power hungry but best at single threading. So, no it is not at all surprising when you take a 125+ W chip, give it only 50 W, have it's P cores suck up the little power you gave it, and you are left with an underperforming chip in a multi-threaded application. At the very least, wait for the non-K versions (especially the mobile or even the T versions) to see how they work at lower powers.

I suspect that at 50 W, the E cores aren't getting 3 W each. At least in the higher power situations, the P cores are taking about 4.5x the power as the little cores. If that ratio stays true when you run down to 50 W, you are looking closer to 5 W each for the P cores and just over 1 W each for the E cores. That might help you get more accurate estimates than assuming 3 W each to all 16 cores.

As for the rest of your post, I don't know what the voltage/frequency curve is supposed to be for the 12900K chip. So, I can't really comment. The dual voltage on Raptor Lake should address much of this issue, the P cores and E cores can get their optimum voltage each rather than being forced into a compromise voltage like Alder Lake has.

 

[Hulk]

Okay I'm going to admit it. I think Ian was right. The thread director isn't keeping enough compute on Handbrake when I'm transcoding something. For example, right now as I type this it's encoding at 16fps, but when I switch apps and put Handbrake on top it increases to 44fps. Now I could understand keeping Handbrake on the E's if I was doing something that required the P's, like gaming or something, but I'm typing a post in Chrome! No reading to have 8 P's standing at the ready doing nothing.

And yes, I realize I can get this going correctly with process lasso but I shouldn't have to. If you're simply typing in Word or something there is no reason the bulk of the processing power can't go to the background Handbrake encoding automatically. Maybe if there were 16 E's or something this behavior wouldn't be so bad but right now it's kind of dumb.

 

[Markfw]

Okay I'm going to admit it. I think Ian was right. The thread director isn't keeping enough compute on Handbrake when I'm transcoding something. For example, right now as I type this it's encoding at 16fps, but when I switch apps and put Handbrake on top it increases to 44fps. Now I could understand keeping Handbrake on the E's if I was doing something that required the P's, like gaming or something, but I'm typing a post in Chrome! No reading to have 8 P's standing at the ready doing nothing.

And yes, I realize I can get this going correctly with process lasso but I shouldn't have to. If you're simply typing in Word or something there is no reason the bulk of the processing power can't go to the background Handbrake encoding automatically. Maybe if there were 16 E's or something this behavior wouldn't be so bad but right now it's kind of dumb.

This is why I did not get an alder lake, and opted for my 3rd 5950x. 16 cores that are all equal = no problems running multiple apps. Maybe in a few years when all the bugs are worked out, we will see. But a 6 or 8p config for gaming (even if you disable the e cores) would be good for a gamer.

 

[TheELF]

And yes, I realize I can get this going correctly with process lasso but I shouldn't have to. If you're simply typing in Word or something there is no reason the bulk of the processing power can't go to the background Handbrake encoding automatically. Maybe if there were 16 E's or something this behavior wouldn't be so bad but right now it's kind of dumb.

You are the end-user, you are the only system administrator of your own system, you are flynn.
Devs will do whatever they think that will benefit the most people.
You can do whatever you think will benefit you the most.
If you want a super easy way to fix that permanently then just add a key to your registry increasing the priority of handbrake.

https://answers.microsoft.com/en-us/windows/forum/windows_10-performance/how-to-permanently-set-priority-processes-using/2f9ec439-5333-4625-9577-69d322cfbc5e

This is why I did not get an alder lake, and opted for my 3rd 5950x. 16 cores that are all equal = no problems running multiple apps.

Lol... there are problems just different ones, handbrake has this issue because it starts with below normal priority, if you run that on your system and run something else compute heavy with a higher priority, then handbrake will slow down to a crawl or even stop encoding all together until the other thing is done.
This might be preferable to you but it's still not a "no problems" and it still has to be "fixed" by changing the priority of either software.


(If your folding runs at normal priority or higher then handbrake will take forever to finish anything.)

 

[Markfw]

You are the end-user, you are the only system administrator of your own system, you are flynn.
Devs will do whatever they think that will benefit the most people.
You can do whatever you think will benefit you the most.
If you want a super easy way to fix that permanently then just add a key to your registry increasing the priority of handbrake.

https://answers.microsoft.com/en-us/windows/forum/windows_10-performance/how-to-permanently-set-priority-processes-using/2f9ec439-5333-4625-9577-69d322cfbc5e

Lol... there are problems just different ones, handbrake has this issue because it starts with below normal priority, if you run that on your system and run something else compute heavy with a higher priority, then handbrake will slow down to a crawl or even stop encoding all together until the other thing is done.
This might be preferable to you but it's still not a "no problems" and it still has to be "fixed" by changing the priority of either software.


(If your folding runs at normal priority or higher then handbrake will take forever to finish anything.)

You are just reinforcing exactly why Hulk is not happy having to mess around, and why I won't be getting one soon. If there was a 16 P core that did not suck power like crazy, I would get it over the 5950x, as I realize they are more powerful. And if Intel did not default them so high in power, to beat the 5950x, it sure would have been better for many, not having to mess with bios to keep them from being a space heater (specifically the 12900k, the 12700k is OK and below)

 

[Hulk]

This is why I did not get an alder lake, and opted for my 3rd 5950x. 16 cores that are all equal = no problems running multiple apps. Maybe in a few years when all the bugs are worked out, we will see. But a 6 or 8p config for gaming (even if you disable the e cores) would be good for a gamer.

Well it's not a deal breaker for me by any means. For $350 the CPU is an absolute beast!
Just being honest.

 

[scannall]

Well it's not a deal breaker for me by any means. For $350 the CPU is an absolute beast!
Just being honest.

Mark's computing needs are far different from most of the rest of us to be fair. Performance per watt is very important, as I remember a post of his saying he was getting close to the maximum amount of electrical service he could get into his home. And for a very good cause as well.

 

[Markfw]

Mark's computing needs are far different from most of the rest of us to be fair. Performance per watt is very important, as I remember a post of his saying he was getting close to the maximum amount of electrical service he could get into his home. And for a very good cause as well.

That is a good price. But the motherboard price, the HSF situation, the P/e core differences just don't make it work for me, or I would think anybody that multitasks a lot.

 

[Hulk]

That is a good price. But the motherboard price, the HSF situation, the P/e core differences just don't make it work for me, or I would think anybody that multitasks a lot.

Good points but just for reference I paid $350 for the 12700K, $190 for the mobo, $140 for 32GB of DDR4 3600 CL16, and $111 for a Noctua chromax.black cooler. Plus I don't game so the iGPU is all I need. That's $791 for cpu, mobo, memory, and cooler, less than the retail price of the 5950X, for which you need a GPU.

I multitask quite a bit but obviously not 24/7 like you. Clearly the 5950X is the best option for your needs.

 

[Markfw]

Good points but just for reference I paid $350 for the 12700K, $190 for the mobo, $140 for 32GB of DDR4 3600 CL16, and $111 for a Noctua chromax.black cooler. Plus I don't game so the iGPU is all I need. That's $791 for cpu, mobo, memory, and cooler, less than the retail price of the 5950X, for which you need a GPU.

I multitask quite a bit but obviously not 24/7 like you. Clearly the 5950X is the best option for your needs.

just FYI... I just got a 5950x (not even here yet) for $709. But I agree, for you, it was a great deal, especially at $359 for the cpu. I actually thought about that chip, but I have no microcenter even close, so $449 made it out of the question. I already had 32 gig of 3200 cl14 (the good stuff) so for me it was a good xmas gift to myself. Oh, and the 3070TI that I got for $860.

 

[Hulk]

just FYI... I just got a 5950x (not even here yet) for $709. But I agree, for you, it was a great deal, especially at $359 for the cpu. I actually thought about that chip, but I have no microcenter even close, so $449 made it out of the question. I already had 32 gig of 3200 cl14 (the good stuff) so for me it was a good xmas gift to myself. Oh, and the 3070TI that I got for $860.

Nice. The 5950X is still a beast, and an energy efficient one at that. Those 16 Zen 3's chew through anything quickly while using little power. There definitely is something nice about having a few handfuls of Big cores

The new rig was an xmas gift to myself as well. Then on impulse I bought one of those Legendary 40mph RC trucks on Amazon yesterday! Another gift for me! I have two daughters 9 and 13 so they aren't interested in building computers and RC toys so I indulge myself from time-to-time. The most "tech" I do with them is helping the older one with her algebra and explaining to the younger why the sky is blue I used to teach high school physics and chemistry... they don't care although my older one is starting to understand it's nice to have someone around to help with her algebra homework. They do everything with mom, as least I'm good for something.

 

[TheELF]

You are just reinforcing exactly why Hulk is not happy having to mess around, and why I won't be getting one soon. If there was a 16 P core that did not suck power like crazy, I would get it over the 5950x, as I realize they are more powerful. And if Intel did not default them so high in power, to beat the 5950x, it sure would have been better for many, not having to mess with bios to keep them from being a space heater (specifically the 12900k, the 12700k is OK and below)

It's not messing around, that is what using your system is.
You are not forced to mess with the bios although that is the simplest and easiest way to do it.
Software like IXTU or even throttlestop can be used to set up profiles for anything you use to use less power or more power depending on your needs.
Also the non-k variant should show up with the rest of the lineup and there might also be a t version for anybody that can't handle using the bios to reduce tdp manually.

 

[TESKATLIPOKA]

You can't make a particulary accurate estimation of the clock speeds as the different cores will have different power to clock scaling.
Although I think the motherboard isn't playing well at 50w. There are going to be low wattage variants of this die (12900T and the H55 cpus) and the performance drop of here looks like an aberration. If this is really how it performs at 50w then I worry for the 6+8 mobile cpus, 8800 in r23 is well below tiger-h and cezanne-h.

Of course, I know the clockspeed is inaccurate. I don't know which type was working at what clockspeed, so I downscaled both cores by the same ratio based on performance difference in CB. Leaks for mobile ADL show much higher performance, not sure how accurate they are.