Intel Alder Lake Processor Tested, Big Cores Ramp Up to 3 GHz

[AleksandarK]

Intel "Alder Lake" is the first processor generation coming from the company to feature the hybrid big.LITTLE type core arrangement and we are wondering how the configurations look like and just how powerful the next-generation processors are going to be. Today, a Geekbench submission has appeared that gave us a little more information about one out of twelve Alder Lake-S configurations. This time, we are getting an 8-core, 16-threaded design with all big cores and no smaller cores present. Such design with no little cores in place is exclusive to the Alder Lake-S desktop platform, and will not come to the Alder Lake-P processors designed for mobile platforms.

Based on the socket LGA1700, the processor was spotted running all of its eight cores at 2.99 GHz frequency. Please note that this is only an engineering sample and the clock speeds of the final product should be higher. It was paired with the latest DDR5 memory and NVIDIA GeForce RTX 2080 GPU. The OpenCL score this CPU ran has shown that it has provided the GPU with more than enough performance. Typically, the RTX 2080 GPU scores about 106101 points in Geekbench OpenCL tests. Paired with the Alder Lake-S CPU, the GPU has managed to score as much as 108068 points, showing the power of the new generation of cores. While there is still a lot of mystery surrounding the Alder Lake-S series, we have come to know that the big cores used are supposed to be very powerful.


View at TechPowerUp Main Site

 

[kayjay010101]

3 whole GHz? That must be a record... I'm blown away :|

 

[dgianstefani]

3 whole GHz? That must be a record... I'm blown away :|

IQ? It's an engineering sample more than six months before retail release.

3ghz cores competitive with 4 to 5ghz cores in early benches, what do you think that means? Or is math too difficult?

 

[ZoneDymo]

"This time, we are getting an 8-core, 16-threaded design with all big cores and no smaller cores present. Such design with no little cores in place is exclusive to the Alder Lake-S desktop platform, and will not come to the Alder Lake-P processors designed for mobile platforms."

wait what? I thought the whole deal with these was the big-little design....now we are getting just traditional cpu's with not little cores?

and I know we talked about this already but man...what is Intel's plan here regarding Rocket Lake...its nice they are transparant about Alderlake being around the corner but....why even release Rocket Lake at this point?

IQ? It's an engineering sample more than six months before retail release.

3ghz cores competitive with 4 to 5ghz cores in early benches, what do you think that means? Or is math too difficult?

not entirely sure how you are going to apply math to this but sure.

 

[dj-electric]

TPU users ability to extrapolate information is immaculate.

Nobody was told the CPU doesnt have the little gracemont cores, and that geekbench even support seeing such thing at this point.
Nobody was told what the final frequency would be for the product, or how boost behavior is going to act like.

 

[Caring1]

TPU users ability to extrapolate information is immaculate.

Based on the fact a 2080 GPU was used in testing, it's easy to extrapolate that the 3 series gpu's were unavailable.

 

[rainxh11]

so what if this cpu can bump up a GPU OpenCL score by 200 points, what the point?, since it's the gpu doing the work
this early leaks are a waste of time, whoever leaked this, plz do a proper cpu geekbench benchmark next time, so we can see what those ES 3.0 ghz can do compared to what we have now
and why rocket lake even exists at this point if something like this is around the corner rocking more IPC & DDR5, rocket lake is looking like the first Gen Core series that got demolished by Sandy Bridge right after

 

[Turmania]

I actually prefer old style before boost days. Boos term is misleading in many fronts and it is only used for marketing lure for it is high frequency. It is misleading and false information on both all core boost and power draw. When I say this, I point at both companies not just one.

 

[Athlonite]

the GPU has managed to score as much as 108068 points

has a fair bit of catching up to do then as my R7 3700X @ 4.25GHz and RX6800 gets 125858 points

 

[1d10t]

Aren't these kind of leaks just make people even convince to skip Rocket Lake?

 

[ZoneDymo]

the GPU has managed to score as much as 108068 points

has a fair bit of catching up to do then as my R7 3700X @ 4.25GHz and RX6800 gets 125858 points

Well yeah, I would imagine an RX6800 beats an RTX2080

 

[Wirko]

"This time, we are getting an 8-core, 16-threaded design with all big cores and no smaller cores present. Such design with no little cores in place is exclusive to the Alder Lake-S desktop platform, and will not come to the Alder Lake-P processors designed for mobile platforms."

wait what? I thought the whole deal with these was the big-little design....now we are getting just traditional cpu's with not little cores?

It has been speculated for a long time that some models wouldn't have any small cores active. Tom's data indicates that Intel is scraping the bottom of the barrel to get operational chips with any number of big and small cores. Too many big cores bad -> goes to notebooks, too many small cores bad -> goes to desktops.

 

[docnorth]

It has been speculated for a long time that some models wouldn't have any small cores active. Tom's data indicates that Intel is scraping the bottom of the barrel to get operational chips with any number of big and small cores. Too many big cores bad -> goes to notebooks, too many small cores bad -> goes to desktops.

I don't know if this is true, but it makes sense.

 

[InVasMani]

It has been speculated for a long time that some models wouldn't have any small cores active. Tom's data indicates that Intel is scraping the bottom of the barrel to get operational chips with any number of big and small cores. Too many big cores bad -> goes to notebooks, too many small cores bad -> goes to desktops.

I defiantly think Intel could bin chips toward desktop and mobile based on the quality of the bigLITTLE chip dies with more of the lower power LITTLE dies that are perfect quality going to mobile and the opposite for desktop. That would make perfect sense to do so for both markets.

 

[Wirko]

I defiantly think Intel could bin chips toward desktop and mobile based on the quality of the bigLITTLE chip dies with more of the lower power LITTLE dies that are perfect quality going to mobile and the opposite for desktop. That would make perfect sense to do so for both markets.

The process of binning must be quite elaborate for these Alder chips. You've got to solve a set of tens of linear and nonlinear equations (freq vs. W vs. $ vs. big vs. little vs. demand) for each chip, which is a procedure generally known today as "AI".

BTW, is there any info or signs pointing to more than one variant of Alder Lake silicon?

 

[Dave65]

IQ? It's an engineering sample more than six months before retail release.

3ghz cores competitive with 4 to 5ghz cores in early benches, what do you think that means? Or is math too difficult?

OUCH, TRIGGERED!
I highly doubt it will go that much higher than 3 GHZ, MAYBE 4!

 

[Vayra86]

I actually prefer old style before boost days. Boos term is misleading in many fronts and it is only used for marketing lure for it is high frequency. It is misleading and false information on both all core boost and power draw. When I say this, I point at both companies not just one.

Its a bit of a dilemma and it goes back to the way marketing is regulated. It can be misleading... but the facts must check out. So what do we get? The highest number without context.

If the marketing was shifted away from 'moar higher' numbers and instead the focus was on real performance metrics (resulting performance, as in, 'compiles in X seconds' or 'framerate') we would not have this problem and boost would not have to be extended to ridiculous just to meet marketing desires. At the same time, boost technology does need to evolve and find the very limit of what it can do to be a 'proper boost'.

It would also bypass the recent discussions we've had on bursty boosting and some CPUs not even meeting the boost they were rated for (Ryzen) albeit briefly. You could wonder how relevant it really is, because you're not using a CPU to watch it gain a certain frequency, you want it to do work.

Perhaps a good comparison is the WLTP rating for E-vehicles. Its a single metric allowing us to compare range between cars based on the same parameters. If you pull up hard and fast a lot and put your heater on all the time, you won't be hitting it - but that's what CPUs have started doing regardless. Power budget is messed with, we get extended limits, etc. What you would want is for the CPU to be more controllable that way. Give us eco and sport toggles - and review CPUs based on those 'stock settings' offered with the product. That's playing fair. @W1zzard already does something along those lines with the testing on max power and turbo in reviews. More control over boost would be the long term desire - a slider that relates CPU performance to the cooling capacity that's placed on it, perhaps even automated / based on a system test. Right now that is sort-of automated but only based on internal metrics, there is no user control, because then you're essentially overclocking.

Note how Intel already developed a tool for at least measuring the maximum output and Nvidia also has automated its overclock (boost) voltage setup.

 

[bug]

I always said the little cores are a waste on desktop. If there are indeed models that sports big cores only, it would seem I was right after all.

 

[Vayra86]

I always said the little cores are a waste on desktop. If there are indeed models that sports big cores only, it would seem I was right after all.

You're right 90% of the time, so I'll join you in that conclusion

If true though... man, Intel really is frankensteining the hell out of all their designs lately. Every little helps...eh big helps.

 

[bug]

You're right 90% of the time, so I'll join you in that conclusion

If true though... man, Intel really is frankensteining the hell out of all their designs lately. Every little helps...eh big helps.

I'm not right 90% of the time, that was Captain's Tom assessment of himself.
Me, I'm happy if I'm right 3 out of 4 times. Depending on the subject at hand, I can be wrong way more than that.

 

[Caring1]

I wonder how Intel Marketing is going to sell this to us, We removed the little cores from Big/ Little CPUs to make it moar betterer for desktops.
Now why didn't anyone else think of making CPUs with just big cores .......

 

[bug]

I wonder how Intel Marketing is going to sell this to us, We removed the little cores from Big/ Little CPUs to make it moar betterer for desktops.
Now why didn't anyone else think of making CPUs with just big cores .......

big.LITTLE was created for mobile. Parts that need to be frugal with power draw.

Not that marketing has a problem selling any kind of crap, but I think they would have had a harder time explaining why you need the low-power cores on a desktop.
And again, it's not confirmed desktop parts will lack little cores. It's just what we believe based on a report from one benchmark. using an engineering sample.

 

[OneMoar]

imagine using geekbench for anything

 

[r9]

Big and small cores design IMO it would make most sense if the small cores had all the fat stripped all those instructions that are not used in 99% of the tasks just stripped off, have like a RISC approach.
That way for the 99% of the tasks the small cores would be as fast as the large ones at much lower power consumption, potentially allowing for higher clocks.

 

[katzi]

Imagine that, intel GLUING CPUs together...

My my...