• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

All-core (P-Core) frequencies of Alder Lake CPUs?

The 12900K is getting 'only' 4.9ghz on all cores with turbo boost enabled, in this review:
https://www.youtube.com/watch?v=Xt7yWzu9EQY

Really useful slide here too:
https://youtu.be/Xt7yWzu9EQY?t=1302

Manually overclocking to get that extra 100mhz (from 4.9ghz to 5ghz) seems to push the CPU temperature up from ~84 degrees to 100 degrees :rolleyes:

According to this, the locked Alder Lake CPUs are clocked 100mhz lower than the locked versions:
https://www.techspot.com/review/2391-intel-core-i7-12700/

So, the 12700/12700f all core frequency (P-Cores) would be 4.6ghz I think? Can anyone confirm that?
 
Last edited:
So, the 12700/12700f all core frequency (P-Cores) would be 4.6ghz I think? Can anyone confirm that?

The 12700K's all-core is 4.7 Ghz, so I think you're correct, buuuut, PL2 is usually different between K and non-K so the clocks (and relative performance) wouldn't be directly comparable unless the power limits were unlocked on the motherboard.
 
So, is it actually worth getting the unlocked CPU versions, if they can't turbo to 5ghz?

Realistically, I think a 12700K at 5.0ghz is on the edge of acceptable load temperatures (and power requirements). Running it at turbo speeds (4.9ghz) is probably preferable.

Intel really needs something like the 12600 and 12700 (locked versions), that can turbo to 5ghz on all (P) cores. They could leave out the E-Cores.

The turbo speeds on the 12600K are particularly disappointing, at just 4.5ghz.
 
Last edited:
I think the issue I have with current Alder Lake CPUs, is the relatively low base clocks, as detailed here:
https://en.wikipedia.org/wiki/Alder_Lake_(microprocessor)#CPU

The 12600K has the highest base clock in the whole series, at 3.7ghz. This takes up quite a bit of power, at 125w...

It's still an improvement on my 8 core Comet Lake CPU, with clocks between 4.6-4.7ghz, with power usage of between 120-130w.

Unfortunately, I think the 10nm process just hasn't turned out to be as efficient as Intel hoped back in 2016...

Maybe they will be able to increase these base clocks for Raptor Lake?
 
Last edited:
I think the issue I have with current Alder Lake CPUs, is the relatively low base clocks, as detailed here:
https://en.wikipedia.org/wiki/Alder_Lake_(microprocessor)#CPU

The 12600K has the highest base clock in the whole series, at 3.7ghz. This takes up quite a bit of power, at 125w...

It's still an improvement on my 8 core Comet Lake CPU, with clocks between 4.6-4.7ghz, with power usage of between 120-130w.

Unfortunately, I think the 10nm process just hasn't turned out to be as efficient as Intel hoped back in 2016...

Maybe they will be able to increase these base clocks for Raptor Lake?

The clocks aren't entirely comparable to comet lake though, because it's a completely different architecture. Comet lake is like the 'ultimate' revision of Skylake and it's a more efficient architecture in games than the 'coves.

I wouldn't be upgrading from a 10700K though, 8 core Skylake at 4.7 Ghz isn't going to bottleneck a 3070 at 4K. Intel say that they anticipate a CPU upgrade every 5 years, so you have 3 to go! :D

From what I heard, Raptor Lake is not targeting power consumption, or higher clocks, they're going for cache.
 
It's weird though isn't it, Intel has the ability to produce 10nm CPUs, clocked at 5ghz (turbo boost), but they choose not to. This could apply to both 6 and 8 core CPUs, we know the 12400 can handle 5ghz as BLCKing has proved this. Instead, Intel favour unlocked K models, that can only reach 5ghz when running at high temperatures (>80 Celsius on most coolers, >90 Celsius for the 12900K).

So, it would be a complete own goal not to release Raptor Lake CPUs that can turbo to 5ghz on all cores, we know this is what AMD is planning for Zen 4 (they have already said this is possible on the Ryzen 6000 mobile series)

What Raptor Lake needs is a 10% boost in turbo boost clock speeds across the board, as well as a substantial increase in L3 Cache. Then, then maybe it won't be totally outclassed by Zen 4.
 
Last edited:
8 core Skylake at 4.7 Ghz isn't going to bottleneck a 3070 at 4K.

I think I'll leave it a bit. Next upgrade will hopefully be an RTX 3080 FE, if I can ever get one.

There's definitely some games that don't care what resolution you run at, they will fully utilize your CPU regardless and you will lose plenty of frames. This can happen in Total War games, as there isn't really a limit on how much the game has to process, even more so in the Laboratory mode. Some games really like CPU cache too, it's difficult to know when you've hit a bottleneck.
 
It's weird though isn't it, Intel has the ability to produce 10nm CPUs, clocked at 5ghz (turbo boost), but they choose not to. This could apply to both 6 and 8 core CPUs, we know the 12400 can handle 5ghz as BLCKing has proved this. Instead, Intel favour unlocked K models, that can only reach 5ghz when running at high temperatures (>80 Celsius on most coolers, >90 Celsius for the 12900K).

So, it would be a complete own goal not to release Raptor Lake CPUs that can turbo to 5ghz on all cores, we know this is what AMD is planning for Zen 4 (they have already said this is possible on the Ryzen 6000 mobile series)

What Raptor Lake needs is a 10% boost in turbo boost clock speeds across the board, as well as a substantial increase in L3 Cache. Then, then maybe it won't be totally outclassed by Zen 4.

It's pretty typical that they do this, because Intel love segmentation. I don't think they have any intention of raptor lake being an am5 smack-down, they'll do what they usually do, make it just enough to capture a few headlines for the high-end, while offering a reasonable improvement in generational performance across all other lines.

AMD have more incentive to move things forward and strengthen their brand, because they have to work much harder to capture market/mind share from Intel. Intel sell everything they can make even when they're behind, so why bother?

You'll get a decent affordable upgrade eventually, but they'll make you wait for it :D
 
I don't necessarily think it's by design, I think they've just hit a wall with what they can do with their 10nm technology, which took them years to even bring to desktop CPUs, at equivalent clock speeds to 14nm. These constraints led them to come up with a hybrid design on 10nm, so they could at least complete with 16 core Zen 3 CPUs in multithreaded situations. Zen 3 has been excellent for AMD, pretty much better in every way than Intel's 14nm desktop CPUs.

I'd like Intel to announce that they are working on further improvements to the 10nm process, but none of their roadmaps have indicated this, only that they want to move to 7nm EUV as soon as possible. Historically, Intel's CPUs have only seen substantial improvements with process improvements also. A new architecture on an old process just doesn't seem to provide much / any improvement.

The thing that makes me doubt the performance of Raptor Lake CPUs is the lack of information on any official roadmaps, despite the inclusion of Meteor Lake (7nm EUV) CPUs.
 
Last edited:
I don't necessarily think it's by design, I think they've just hit a wall with what they can do with their 10nm technology, which took them years to even bring to desktop CPUs, at equivalent clock speeds to 14nm. These constraints led them to come up with a hybrid design on 10nm, so they could at least complete with 16 core Zen 3 CPUs in multithreaded situations. Zen 3 has been excellent for AMD, pretty much better in every way than Intel's 14nm desktop CPUs.

I'd like Intel to announce that they are working on further improvements to the 10nm process, but none of their roadmaps have indicated this, only that they want to move to 7nm EUV as soon as possible. Historically, Intel's CPUs have only seen substantial improvements with process improvements also. A new architecture on an old process just doesn't seem to provide much / any improvement.

The thing that makes me doubt the performance of Raptor Lake CPUs is the lack of information on any official roadmaps, despite the inclusion of Meteor Lake (7nm EUV) CPUs.

My understanding of it is that Raptor Lake will have much bigger caches and a decent IPC bump, so I don't think significantly higher clocks will be needed, at least if we're talking about games. It is typical that as the process matures, they up the clocks.

We're very early in the cycle of the architecture, so there's plenty of room for optimisations, whereas 14nm largely relied on clock increases for more performance since 4th gen. AMD wasn't competitive with Nehalem and it's successor (Sandy), right through to Zen's release, so there was little need for Intel to invest in improving what they had. If I recall correctly, Coffee Lake onwards was supposed to be on 10nm (and 10th and 11th gen weren't what they planned), but they ended up wedging 2 and then 4 more cores on 14nm to compete with Ryzen.

Rocket Lake did offer some improvement in their main weaknesses and there were some cases that the improved performance from the new architecture did actually show, even though overall it flopped. From my perspective, Alder Lake proves that the architecture is decent and has some legs as e.g. the 12400 is very competitive even at relatively low clocks.

I actually have more doubt that AMD can sustain the level of development Zen has shown.
 
Why? AMD has a new CPU socket (AM5), DDR5 RAM support (likely DDR5 5200mhz like Zen 4 server CPUs) and TSMC's 5nm EUV process technology. 5nm alone will allow AMD to overtake Intel both on transistor density and power efficiency. AMD has also said their prototype can do 5ghz on all cores. The new AM5 socket should allow them to utilize more power if needed, beyond 140 watts - useful for 16 core or more CPUs.
 
Why? AMD has a new CPU socket (AM5), DDR5 RAM support (likely DDR5 5200mhz like Zen 4 server CPUs) and TSMC's 5nm EUV process technology. 5nm alone will allow AMD to overtake Intel both on transistor density and power efficiency. AMD has also said their prototype can do 5ghz on all cores. The new AM5 socket should allow them to utilize more power if needed, beyond 140 watts - useful for 16 core or more CPUs.
Just because the scale of improvement with one architecture seems unrealistic to maintain long-term.
 
My 12600kf runs 5ghz oc on all p cores no sweat. With my h115i it maxes out @70c full load. Before that with the msi mag core liquid 240 it was 80c.

I think the turbo speed depends on the motherboard. My i5 boosted to 4.9ghz before the oc.
 
Last edited:
My 12600kf runs 5ghz oc on all p cores no sweat. With my h115i it maxes out @70c full load. Before that with the msi mag core liquid 240 it was 80c.

I think the turbo speed depends on the motherboard. My i5 boosted to 4.9ghz before the oc.

Yes, not sure where g67575 is getting the temps from.
Maybe if I stress test my 12600k in ycruncher it hits hight temps, but with an old d14 noctua when gaming i am in the 70's at 5GHZ p core 4GHz ecore.
 
@=XDC=FluphyBunny Good to know with your air cooler. Hopefully it should also be fine in summer when it warms up more.
My 420mm AIO is good, as you would expect, with my 2700k at 5.1 all P core and 4Ghz E core. But applications like y-cruncher or, to a lesser extent, the benchmark Cinebench R23 can raise the temps. But that doesn't represent my typical use.
Ok to use y-cruncher etc for stability testing but not to evaluate temps etc in respects of typical use, for me.
 
Yes, not sure where g67575 is getting the temps from.
Maybe if I stress test my 12600k in ycruncher it hits hight temps, but with an old d14 noctua when gaming i am in the 70's at 5GHZ p core 4GHz ecore.

I've seen quite a few posts about alder Lake being power hungry and hot. I think this is to do with vcore being to high. On my Gigabyte z690 the vcore is set to auto. Performance profile is standard and the vcore never goes over the limit. Stress test maxes at 80c and gaming mid 50s. I was told by several people not to even try running it with air cooling. Your having no problems.
 
Auto vcore is awful on my MSI Z690 and 12700k, would have it spiking to nearly 1.4v on some occassions. It will happily work at stock clocks as low as 1.22v without performance dropping.
 
Back
Top Bottom