Linux 5.16's New Cluster Scheduling Is Causing Regression, Further Hurting Alder Lake

[middy]

Pretty much, yeah. There isn't really anything fundamentally "wrong" with the concept, but it's far from unreasonable to say that you have no interest at all in running gimped cores to save a miniscule amount of power, when all that silicon could have gone to larger caches etc instead.

Even the now-ancient SpeedStep/EIST/etc already provides analogous behavior: idle cores are cut down to half speed or less, with a corresponding reduction in power draw, and completely idle cores can even power-gate - but they're ALL still *capable* of actually performing well, unlike the E-cores. On a *desktop* CPU I'd much rather see improvements in the gating etc than have half the cost of my CPU going to pay for garbage cores that are literally useless for half the things I want a PC to do.

from seeing all the benchmarks for alder lake, i concluded intel is only pushing the e-core thing because they can't scale up their full size cores to high core counts with sane power consumption. e-cores allows them to do that. i really don't think they had power consumption with mind as their focus for it. i really think it was just score scaling as the driving factor. e-cores is allowing them to slap bigger numbers on the box for core counts.

i just hope this doesn't make intel feel content with not working hard to make their normal cores power efficient and use "e-cores" as an excuse to slack off.

[avem]

Looks like one big cluster f*ck.

Which can be trivially fixed as long as you are not an imbecile.

[avem]

from seeing all the benchmarks for alder lake, i concluded intel is only pushing the e-core thing because they can't scale up their full size cores to high core counts with sane power consumption. e-cores allows them to do that. i really don't think they had power consumption with mind as their focus for it. i really think it was just score scaling as the driving factor. e-cores is allowing them to slap bigger numbers on the box for core counts.

i just hope this doesn't make intel feel content with not working hard to make their normal cores power efficient and use "e-cores" as an excuse to slack off.

No, they have e-cores because they offer more MT performance for less power. Intel could have had 10 P cores instead of 8p/8e and MT performance would be worse while power consumption would be a lot higher.

There's no need to actually have more than 6 p-cores because there aren't that many tasks which a ST performance bound.

[F.Ultra]

It's exactly the same. ARM is even more complicated because they often use a big.medium.little approach, while Intel only has big and medium cores.

big.LITTLE can run in an either/or configuration as well, that is how they where initially run before there where schedulers that could use the cores at the same time.

[mSparks]

There's no need to actually have more than 6 p-cores because there aren't that many tasks which a ST performance bound.

Thats what Intel had been telling everyone for the decade before AMD came along, served those of us who have a high need to more CPU compute and took the entire market for servers and developers - the only people still buying x86_64 in any quantity.

If you dont have a need for more full cores, just spend the $200 adding a raspberry pie cluster, it will give you better performance and consume less power than spending $200 on 6 gimpy x86 cores.

The problem with e cores is they dont serve either market. I also highly doubt anyone but a few edge enthusiasts will fall for the BS of calling these 12 or 16 core cpus.

I'd love to see benchmarks of these computing PI.

[geearf]

If you run software that can use many cores, and the smaller cores offer more multi-threaded performance per unit of die area, then the answer to increasing is to keep on adding small cores. The big cores are there to handle tasks that can't be parallelized. 6-8 big cores is still plenty for most users.

Do the smaller cores always offer more MT performance?

I'll admit, I've been wanting for a long time to be able to add a ThrearRipper to my Ryzen system for that exact purpose, it sounds nice in theory, but I don't know if it is best in practice.

Thank you!

[Alexmitter]

Afaik, big.LITTLE is an either/or affair (i.e. you can use the big part or the LITTLE part of a core, not both at the same time). Alder Lake is different, the E cores run alongside P cores. Scheduling logic will be quite different.

That was a thing for the very first generation of big.LITTLE. But not the case for every generation since.

All clusters run at the same time.

[Alexmitter]

This is a conspiracy, i say, to make Intel look bad on Linux!. If people can claim that MS is purposely messing with Ryzen, such as was claimed when AMD introduced the 3990X and now with Win 11, then I can claim that this is done on purpose to sabotage Alder Lake on Linux.

The changes that introduced this regression were done by two intel engineers.

[jaxa]

Do the smaller cores always offer more MT performance?

I'll admit, I've been wanting for a long time to be able to add a ThrearRipper to my Ryzen system for that exact purpose, it sounds nice in theory, but I don't know if it is best in practice.

Thank you!

Not necessarily. For example, you can find cases where a 64-core Threadripper 3990X offers no benefit over a 32-core Threadripper 3970X, because software isn't using all of the cores. On the other hand, most benchmarking deals with one application running at a time, and you can expect applications to support more cores in the future.

P.S. You technically can't add a Threadripper to a Ryzen system because they use a different socket and motherboards.

Do note that the die space of 1 Colden Cove is estimated to be about equivalent to 2 Milan(full Zen3) cores.
And that the Renoir core (cache reduced zen2) is about half the size.

Meaning 1 E core is about the size of a Renoir core, which makes it not that small.

That's why AMD should have a pretty good time implementing their own version of big.LITTLE. You are right to bring up Renoir, because it had a small cache compared to Cezanne and desktop chips, but it peformed reasonably well. AMD's newly announced Zen 4c "cloud-optimized" cores are rumored to be included alongside Zen 5 in Granite Ridge desktop CPUs a couple of years from now. 8 Zen 5 + 16 Zen 4c.

Thats what Intel had been telling everyone for the decade before AMD came along, served those of us who have a high need to more CPU compute and took the entire market for servers and developers - the only people still buying x86_64 in any quantity.

If you dont have a need for more full cores, just spend the $200 adding a raspberry pie cluster, it will give you better performance and consume less power than spending $200 on 6 gimpy x86 cores.

The problem with e cores is they dont serve either market. I also highly doubt anyone but a few edge enthusiasts will fall for the BS of calling these 12 or 16 core cpus.

I'd love to see benchmarks of these computing PI.

Alder Lake is still another beta test of the concept for early adopters, unfortunately. It only manages to beat the 5950X in some scenarios because of improved IPC, DDR5, and higher power draw. Things will become more clear as Intel refuses to increase the big core count in upcoming generations, putting out CPUs with 8 big cores and 16-32 small cores.

Meanwhile on the AMD side, AMD can start using more energy as the AM5 socket is designed to support a 170W TDP instead of 105W (peak power draw is higher). That should help them hit higher clock speeds and support more cores at the top. AMD is also rumored to be doing their own take on big.LITTLE with the newly announced, denser Zen 4c cores. If it serves no market, then why would AMD also be doing it? Maybe because it's actually a good idea, once all the scheduling/regression issues are ironed out. AMD should benefit from being a couple years late to the party.

[uxmkt]

Multiple reviews show you're completely wrong but nowadays it's just fine to say whatever BS you want as long as it's called an opinion.

That's what the downvote button is for, to spot, tag and filter misleading opinions. We do have a downvote button, don't we? We don't? Michael? Michael! Damn, youtube is trying to get rid of the downvotes too!