Made for "Lower Power" a.k.a BGA trash. Sadly that's where the money is, because so many people simply do not care to buy PC for longevity they all want the castrated performance for the little convenience it has and Apple consumers being even worst accepting Soldered SSDs, RAM and everything irreparable garbage.
These designs make a lot of sense when laptops are such a significant part of the market. So many people are not looking for a high-end gaming machine. Laptops are now so useful given that it's easy to plug in another two displays to practically any laptop (USB alt mode + native HDMI 2).
To be fair, upgrades are on a much slower cycle. For instance, I'm typing away happily on my Stream 11 with a Celeron n3060 and its sufficient for the vast majority of my compute needs including entertainment. No I don't stream the games I play or run anything graphically demanding, but there is no compelling need to replace CPUs, add RAM, etc if someone has realistic expectations about what computing actually requires. Most of what people do in terms of home computing is entertainment driven, wasteful, needlessly energy and cost intensive, and promotes physical inactivity anyhow. BGA is the least of our worries as a civilization and species.
BGA is still an enviromental concern for enabling throw away culture. I recently upgraded an old second gen core i laptop from a pentium to a dual core i7 along with an SSD, and that thing FLEW. If it was BGA that still functional machine would be scrap, as the pentium was far too slow for daily use anymore.
They've already done that type of thing before with Kaby Lake-G, with pretty decent results, but I'd expect them to show HBM stacks in the slides if they are doing this again.
see yt review, "Intel Core i7-8705G Benchmarked, RX Vega M + Kaby Lake-G"
so, yeah, the ucie looks like a good idea, and cxl over ucie would be an interesting feature, but I haven't seen Intel describing any details like that for Meteor Lake.
I wonder if we will have a repeat of the Athlon 64 era. AMD integrated the memory controller first and had a good product for a while until Intel did the same thing but things seemed to go down hill after that. Some of that was definitely bad moves on AMD's part but some of it is just no longer having the advantage of making a change earlier. Hopefully AMD stays more competitive this time.
I think this time the hope is that Intel stays competitive...;) How amusing: "the Intel 4 process, which was formerly known as Intel’s 7nm process"...;)
AMD was castrated then by intel's monopolistic tactics and being forced to use their supply limited and expensive foundry. I do not think Lisa would make the same idiotic mistake.
Intel played a major part, but AMD also made multiple bone headed decisions, such as spending money on ATi or building globalfoundaries while their athlons stagnated, then rushing to release phenom when the core 2 landed and ROFLstomped athlong 64.
One never knows but I doubt it'll happen again that easily. Back then, AMD underestimated Intel, rested on their laurels somewhat, and weren't that aggressive with K10/Phenom. As Papermaster or someone said a while ago, it's going to be tock, tock, tock all the way now.
Which is fine, since desktops are 99% pointless; for everything that isn't 100% CPU a laptop with a monitor works better (whether it's in a laptop-shaped case or a Mac Mini-shaped case), for gaming a console works better, for things that are 100% CPU a server works better.
Astonishingly incorrect. All programs are limited to some degree by Amdahl's Law. Combining the use of bigger performance cores to handle the master threads and of PPA-optimized "small cores" to improve perf/$ metrics is universally useful and a step forward in CPU design.
You may argue that Intel's current implementation isn't very useful for g*mers, but that's because games are slow as hell to adapt to hardware changes - and honestly, no one really cares about what g*mers think?
Hybrid tech adds latency and complexity that is unnecessary for a Desktop PC. Remember these are machines which are hooked directly to the power-socket, and prioritise performance for certain applications. So a "Standard" Setup like 8x Large cores in Intel's monolithic chips, is the best option.
On top of this, servers run differently to your standard Desktop PC and Mobile PC. They prioritise security above all. So a Large Setup like 64x Large Cores in AMDs chiplet design is the best option.
For Mobile uses, there's a decent range. From the 4W Thin Tablet devices, with say 4x small cores. To the 20W Slim Laptops devices, with say 2x Large Cores and 4x small cores. And up to the 40W Heavy Notebooks, with say 4x Large Cores and 4x small cores. In these circumstances, Hybrid Computing as developed/perfected by ARM and Apple is the best option.
you seem so very sure, can't say I share your belief. I have a Ryzen 5800U in an ultra book where it obviously struggles to clock all eight cores beyond 4 GHz at both 15 and 28 Watt TDP presets, but I am doubtful that replacing the last 2 Zen 3 cores with something of an E-core equivalent would yield additional compute from the same power/heat budget except in some purpose built synthetic benchmark. In a 24x7 power optimized edge or home server I can imagine assigning certain tasks to E cores to let P cores slumber. But then I might even want some control over RAM, because that can easily take a bigger bite than core power savings with DIMMs.
Well, what I said is true as long as all factors are controlled. So that means each chipset is using the: same software (eg Windows10 Pro), same lithography (eg 6nm), same architecture (not huge difference like Zen1 vs Zen3+).
A large server that you use as a Cloud, needs to handle lots of queries at the same time. Here security is paramount, followed by multi-threaded performance, then efficiency. So AMD's chiplet design which can hold many cores in one chiplet, and combine many chiplets into a large chipset works great. You can greatly increase multi-threaded performance this way, and it won't decrease the efficiency by much, and will keep things secure. If one user requires more performance (or cache) the system might steal resources from another core, and this will take a hit to latency. But this latency is acceptable since there's a tertiary latency with the external connection between the User and the Server.
On a Desktop PC, something like Intel's Core i9-9900k works well. And that's running on an outdated +14nm node, on an outdated architecture (Skylake), which was able to match a new architecture (Zen2) on a 7nm node. If you were able to refresh that with a newer architecture on a newer node, it will perform better against the chiplet design of AMD. The reason is that on a monolithic design, the whole chipset is stamped out by the same good node and it is a closer integrated system. Since this is a computer for a single user, performance (single-core) is much more useful than outright total/multithreaded performance and the system is connected to a power socket. So some efficiency is traded for more performance (ALL Large Cores).
Now stepping down to Heavy Notebooks, here it is expected to run on battery power. And occasionally be plugged into a power socket to extract more power. For that reason, it is very beneficial to use small cores for Power Efficiency and Balance them out with some Large Cores. Since this is a single user, there's no concern about security between each cores/packages, so having a weird Hybrid Design is not very important.
Moving down to Slim Laptops, here it's expected to be used exclusively on battery power. However, the nature of the form-factor does allow the installation of a decent-sized battery and a pretty potent cooling solution. So there is a pretty even balance between having small cores and Large Cores. Again, inter-core security isn't a problem here, and it does reduce efficiency a bit... but it becomes a benefit by providing ample performance at acceptable power drain.
At the lowest end of the spectrum, at least for x86, are Thin Tablets. These use a passive cooling method, and due to battery/heat concerns don't charge particularly fast. Here Large Cores can come to the detriment of the experience. It might deliver an initial jolt of performance, but it will use up a notable amount of battery power, then heat up the device, which causes it run slow for the remainder of that duration. So the most optimal solution is really to use all small cores solution, for maximum efficiency.
Congrats Kangal on typing a lot of words. They're still not accurate though! Your cursory understanding of chiplet design doesn't make you an expert.
You have decided that heat is completely irrelevant to a desktop or server-class design. This is completely wrong. One of the big challenges when Zen 2 came out was that the chiplets produced hot spots that were extremely difficult to cool because the small dies did not disperse heat across a continuous silicon surface. Adding little cores is a perfect solution because they do not dramatically change the architecture of the tile/chiplet, add silicon to allow better heat dispersion, and also do not dramatically change the power profile of the tile. The net result is that the little cores are basically "free" in a tiled/chiplet architecture.
If it wasn't obvious, this is part of why AMD is also releasing Zen 4c.
Thank You. I am an amateur at best, much like most commenters here such as yourself.
Where did I say heat was irrelevant, because I can't see it up there. Desktop PCs are designed for One User with Maximum Performance in mind. That is why they are connected direct to a power-socket, physically large, and have highest cooling system (per user) compared to the other options. Yes, Desktop PCs cool more than Servers because Servers run parallel many users at the same time, and they are more balanced due to budget constraints (electricity).
Your statement about Zen 4C is completely wrong. There is very low information about those cores, so anything you say is just misinformation. They could in-fact be "small cores" as you infer, usually achieved by taking an older architecture (eg Zen1) to miniaturising and optimising it for efficiency and area. That's what Intel did with their Pentium-Celeron-Atom cores, and later with their Skylake-eCores. Or these might actually be the same Zen4 cores, but they are using better silicon, packaged differently on-die, and are better voltage regulated. They may even have cut-down on the L1/L2 cache slightly, or removed a bit of processing/hardware which is not useful for Cloud Computing. At the end of the day, it is all speculation.
"On a Desktop PC, something like Intel's Core i9-9900k works well. And that's running on an outdated +14nm node, on an outdated architecture (Skylake), which was able to match a new architecture (Zen2) on a 7nm node. If you were able to refresh that with a newer architecture on a newer node, it will perform better against the chiplet design of AMD. The reason is that on a monolithic design, the whole chipset is stamped out by the same good node and it is a closer integrated system. Since this is a computer for a single user, performance (single-core) is much more useful than outright total/multithreaded performance and the system is connected to a power socket. So some efficiency is traded for more performance (ALL Large Cores)."
There's two options for that hypothetical die. The die shrink with no other changes leads to no performance gain. Ballooning the die size back to the same area on a monolithic design just leads to leakage, horrible binning, and resulting heat increase. The same can be said for clockspeed increases. The only valid extrapolation from the paragraph I quoted is that you don't care about efficiency when scaling the design up. Otherwise it would be obvious why the 9900K is a dead end.
These issues are not isolated. They are not yet solved. I am sure they can be conquered but it's nontrivial. Point is -- building dies to better diffuse heat is a must when the "uncore" elements that used to add cooling surface are separated out.
Zen 4c is guaranteed to be smaller than Zen 4. I can guarantee that AMD is not wasting any space on its Genoa platform in either of its high-end configurations using Zen 4 and Zen 4c. The expected difference is cache, but that's still a component of the core.
I'm a bit astonished by the inaccuracies in your Intel history section. There's no such thing as a Skylake eCore. Intel drew upon P3 for early Atom designs but Silvermont and later are dramatically different.
Not true. If you’ve got 30 threads and 26 of them are time or performance insensitive then you can schedule them on the efficiency cores. Doing so gives your performance cores a larger slice of the power budget, which in turn means you performance sensitive tasks can run faster and hotter.
IE a file copy, a network stack, an indexer, and a virus scan can all run on a set of low power cores that only sips 5W, leaving you main app the remaining 145W; the alternative is to split your performance cores in those tasks.
Intel's Alder Lake eight e-cores are reportedly more performant than the four cores in my three year old NUC box... so looks like they have a point to me. The Raptor Lake chips are supposed to be drop-in upgrades, adding 8 more e-cores, so looks like a good upgrade path to me.
I've also read some comments about using process lasso or a windows power profile as a way to effectively go fanless, so I'd be interested in trying to run exclusively on e-cores and silence the fans when I don't want to hear them.
as a reference, wikichip fuse has an article "Intel’s Gracemont Small Core Eclipses Last-Gen Big Core Performance" which presents Intel slides on Gracemont relative performance vs the Skylake cores.
Not at all. If I can fit 4 efficiency cores in the same area as 1 performance core and get greater multi-threaded performance than that performance core then it definity has merit on desktop. At the end of the day we are still die size limited. Getting more cores on that die is certainly useful.
AT has become a real ghost town with the highest activity being some robot "Deals".
Really great writers left without parting words like Johan de Gelas or Andrei Frumusanu, some parted with words, but even the last full timer had very few things to say for some time now.
Do you have a legal obligation to keep the site running or do you actually believe you can get out of this stall?
Closing AT would be terrible, but it doesn't get any better by dragging it out like a 19th century opera death.
You know what's going on. No one pays for online content anymore. They are at the mercy of their advertisers. All those writers were good guys and probably made good money going into industry. It is what it is. No point in harassing the guys working their ass off to try and keep it going.
Agree. No surprise if they moved to Youtube and Patreon to support their writing, leaks and information. Just like how GamerNexus and similars do their services.
Asking what the plan is isn't "harrasment" and that spineless wimpering is what leads to the internet becoming so sanitized that nobody can make a living on it anymore.
You are asking people with expertise and skillset that qualifies them to get paid well, but you expect them to work for free ... because you are entitled to it.
🤷🏻♂️
Why do you think people have moved to TikTok / and Clickbait YTube video in 10-15mins format rather than doing any deep dive or critical analysis of any kind?
Perhaps they should follow the direction of Linus Media Group. Have a bunch of professionals who are great at the difficult technical stuff. Then have a handful of people that have great personalities in front of the camera. And drive the business first by info-tainment videos, and follow up with the professional articles.
Otherwise, when all the great people at AT leave and are replaced by bad actors, the important but/heart has been replaced. You've essentially lost that company to obscurity, ignorance, or advertising for big corporations.
A format pivot like that requires vision, leadership buy-in, and funding. There is also risk that AT will not be able to shoulder aside existing media-centric companies such as LMG that already have built an extensive fan base. Future (or whatever company owns AT now) may not have the interest of financing to support an effort like that and my own opinion is that they waited too long and are simply not in a position to acquire the cash to make a pivot.
It's too late for an Anandtech pivot. The talent drain due to the funding drain wiped out that chance a while ago. Who is left? Ryan, Gavin, Brett, E.F? They don't all live on the same continent so I think a LMG pivot is out. If Ryan is smart, he's already working his next industry job right now. The Purch/Future money is drying up, the budget was slashed a long time ago, the owners don't care, the ads don't pay....
And we're left with sites that are devolving, with things like the post hiding and political nonsense at Ars. That was never a great site but it has drastically declined in terms of the ability to have a meaningful conversation — due to the implementation of the voting game. Extremetech has gay-bashing as part of its main course now. The admin there even speaks directly to the person using a bigoted alias, as if there's nothing strange about it.
Anandtech is one of the few remaining bastions of civilized discourse and it seems that's no longer what people in this world are looking for. They don't want to debate without absurd Orwellian games like post hiding voting systems. Echo chambers are not interesting to me.
Well, typically "gay bashing" usually means something along the lines of " I'm not a huge fan of the alphabet group".
Ars Reddita is far worse, with the hiding of the Dr. CheesePizza and using the same lame voting system that turned reddit into an insane echo chamber of losers.
‘Well, typically "gay bashing" usually means something along the lines of " I'm not a huge fan of the alphabet group".’
Use that absurd formulation to try to justify posting bigoted comments on the basis of race, sex, and religion and see how far it gets you.
People who are unable to appreciate valuable members of society due to their unnecessary false beliefs can choose to keep their opinions to themselves. Not being a ‘fan’ doesn’t entitle them to attack people.
Someone has been posting there since February under the name GodSavesH0m0sNo (and various permutations). Joel Hruska, the site admin, even chatted in the comments to that person amiably, as if it’s totally fine to have a portion of his readership under attack.
I know someone who wrote to the parent company and was completely ignored.
Anandtech cant churn out reasonable content anymore. Their primary base wanted GPU and CPU and Case reviews, something that younger creators are better at addressing. Anandtech's older audience likes the deep dives, but anand has lost much of the manpower that makes it happen.
I don't view the comments on those other sites, and hardly ever visit them anyhow, but think the internet as a whole has degenerated. What is bad and vain in human nature, just springs out amplified. Possibly, the less we read online, the better. I wonder what further steps down this spiral staircase the so-called Metaverse and Web 3.0 will take us.
The Internet has exposed human corruption schemes across the globe. A major psychological response has been for people to use verbal warfare to try to fill the void where their faith in human goodness used to be.
Sites also use design to encourage the warfare, like the voting + comment hiding systems. Those change the discussion process into a game where shunning is the payoff.
I also think it's given people the opportunity to play amateur philosopher, writer, and slactivist. Social media, infinite scroll, dopamine-producing design, etc. surely can't be good for the human brain.
And yes, I agree and have a low opinion of comment-voting systems, which seem to have the effect of suppressing unorthodox, out-of-the-way views.
It’s the implementation that matters. People can be mistreated in a forum that doesn’t involve technology. The implementations on the Internet continue to become more China-like, more gamist (such as enticing people to form cliques to shun others via downvoting) and more fraudulent.
For instance, Disqus posted an announcement years ago that claimed that the best way to improve a forum is to have more and more censorship. The same announcement claimed that that isn’t censorship at all. The company rolled out the feature of people thinking their post is being seen by others but it is only visible to them.
Post hiding is becoming very common and it is always implemented in a deceptive manner. The most basic fact about that is that argumentum ad populum is a fallacy.
People are being trained to think it’s normal for someone’s input to not be obviously blocked or erased (so that the censorship is able to be evaluated) — but instead to always face the possibility of being erased/blocked without anyone knowing. It’s an extremely Orwellian situation.
Social media is what people make it. It’s not good nor bad by default. It also makes a difference — who benefits and who suffers. So, the opinions about what is ‘good’ and what isn’t usually come down, as always, to money talking. Exploitation.
I think the more censorship there is in any field, the more blind people become to it. At length, they believe there isn't any censorship; for the orthodoxy has prevailed in snuffing out individuality. Also, the deceptive design tactics we're seeing across the industry are perhaps a reflection of dishonesty, increasing dishonesty, in the people behind it.
On argumentum ad populum, well, that's just another instance where democracy can play tyrant; the fire of general opinion doesn't look too kindly on lone, unorthodox points of view. On the Brave New World vs. Orwellian question, I'm inclined to think the world has tended more to the former, with some touches of the latter. It's easier to control people with the perfumed path of pleasure than dictatorship.
Argumentum ad populum is worse than democracy in a forum setting because the design and the administration select for the population that votes. People who don’t like the design and/or the behaviour/attitudes of the admins (which often includes favoritism and sock puppetry) will go elsewhere. That leaves an echo chamber.
In regular democracy you get the problems of popular ignorance and groupthink. In a more controlled setting like an online forum, you get an increasingly selected population who often think that narrowmindedness is a good thing. Sometimes it is, such as for specialized professional discourse. Many times, though, it’s a disaster like Ars.
For some things video may be better, but most of the time they just waste my time and patience.
I've seen some videos stretching what would be five lines of text into 15 minutes of relentles powertalk. And even if I grant Linus a certain charm, I can only tolerate him in very small doses and I still read much faster than anyone can speak.
For the info I want, text is so much more efficient, and if it's well done, much better abstracted.
I subscribe to some news sites, but logging in means personal tracking, which I dislike.
And Charlie is just mad in terms of pricing.
Off the rails, too, from time to time: his rant on the 5800X3D shows that even his well formed prejudice doesn't always win over hands-on trying first.
Forget Linus, Gamers Nexus and Hardware Unboxed are where it's at.
As for AnandTech, Tom's Hardware is its "sister site" (same owner). The articles are typically of lower quality, but they cover some interesting stuff not found here and they got a 5800X3D review out on time.
Games Nexus also have the same issue of video-only for the vast majority of their output (occasionally there will be an 'article' that is a video transcript).
There's apparently a MASSIVE gap between Intel demo's (even demo's of an actual device, let alone mere "running in the lab" claims) and shipping in volume.
And note that they aren't telling us the transistor density of these tiles (though it surely won't change much between now and ship)? Why? Probably because the message is supposed to be about how i4 achieves a density of 200MTr/mm^2, but the actual tiles are probably 80..100.
Exactly, I remember Intel touting it's 10nm transistor density was 100MT/mm2, yet after the dud Cannon Lake they stopped announcing both transistor density and transistor count (so it couldn't be calculated), yet the best estimates for Tiger Lake was that they were hitting around 50MT/mm2, less than TSMC's 7nm. Fact was most of the +++'s to 10nm process was about increasing gate pitch (not reducing it) to hit higher frequencies and power limits, at the detriment of transistor density.
If this (https://www.digitimes.com/news/a20220503PD216.html... is true, it doesn't bode well for Intel's confidence in Intel 4. Also reinforces my view that all of Intel's claims are suspect until they actually ship.
I don't understand. The post about the "first disaggregated product" contains a photo of a woman holding a package with a single die. How is that disaggregated?
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
Facebook
Twitter
RSS
72 Comments
Back to Article
ballsystemlord - Friday, April 29, 2022 - link
"Gluing chips together"? Naughty. Naughty. ;)Silver5urfer - Friday, April 29, 2022 - link
Made for "Lower Power" a.k.a BGA trash. Sadly that's where the money is, because so many people simply do not care to buy PC for longevity they all want the castrated performance for the little convenience it has and Apple consumers being even worst accepting Soldered SSDs, RAM and everything irreparable garbage.Silver5urfer - Friday, April 29, 2022 - link
Edit - I didn't mean to reply to you but make my own. Somehow that mistake happened.AusMatt - Friday, April 29, 2022 - link
These designs make a lot of sense when laptops are such a significant part of the market. So many people are not looking for a high-end gaming machine. Laptops are now so useful given that it's easy to plug in another two displays to practically any laptop (USB alt mode + native HDMI 2).TomWomack - Saturday, April 30, 2022 - link
I do not care if I can't repair or replace things that are already big enough and never break.PeachNCream - Sunday, May 1, 2022 - link
To be fair, upgrades are on a much slower cycle. For instance, I'm typing away happily on my Stream 11 with a Celeron n3060 and its sufficient for the vast majority of my compute needs including entertainment. No I don't stream the games I play or run anything graphically demanding, but there is no compelling need to replace CPUs, add RAM, etc if someone has realistic expectations about what computing actually requires. Most of what people do in terms of home computing is entertainment driven, wasteful, needlessly energy and cost intensive, and promotes physical inactivity anyhow. BGA is the least of our worries as a civilization and species.TheinsanegamerN - Monday, May 2, 2022 - link
BGA is still an enviromental concern for enabling throw away culture. I recently upgraded an old second gen core i laptop from a pentium to a dual core i7 along with an SSD, and that thing FLEW. If it was BGA that still functional machine would be scrap, as the pentium was far too slow for daily use anymore.nandnandnand - Saturday, April 30, 2022 - link
It's not an iGPU, it's a tGPU!JayNor - Saturday, April 30, 2022 - link
What's the difference between tGPU and iGPU? Does the tGPU connect to the CPU via a pcie bus and have its own memory controller?michael2k - Saturday, April 30, 2022 - link
Yes, something like that:https://www.anandtech.com/show/17288/universal-chi...
JayNor - Saturday, April 30, 2022 - link
They've already done that type of thing before with Kaby Lake-G, with pretty decent results, but I'd expect them to show HBM stacks in the slides if they are doing this again.see yt review, "Intel Core i7-8705G Benchmarked, RX Vega M + Kaby Lake-G"
so, yeah, the ucie looks like a good idea, and cxl over ucie would be an interesting feature, but I haven't seen Intel describing any details like that for Meteor Lake.
III-V - Wednesday, May 18, 2022 - link
They were called IGPs back with Intel Clarkdale... calling it a "tGPU" is just embracing a nonsense buzzword.kpb321 - Friday, April 29, 2022 - link
I wonder if we will have a repeat of the Athlon 64 era. AMD integrated the memory controller first and had a good product for a while until Intel did the same thing but things seemed to go down hill after that. Some of that was definitely bad moves on AMD's part but some of it is just no longer having the advantage of making a change earlier. Hopefully AMD stays more competitive this time.WaltC - Saturday, April 30, 2022 - link
I think this time the hope is that Intel stays competitive...;) How amusing: "the Intel 4 process, which was formerly known as Intel’s 7nm process"...;)JayNor - Saturday, April 30, 2022 - link
Intel used tsm N5 on Ponte Vecchio compute tiles, tsm N6 on the ARC GPUs and will reportedly use TSM N3 on the Meteor Lake GPU tiles.AMD could have moved to EUV processing on Milan, but chose to stay on DUV N7. I get the impression that AMD wants to stay on proven nodes.
Intel is taking more risks with N3, but perhaps sees this as necessary to be competitive with Apple.
t.s - Tuesday, May 3, 2022 - link
I'd say more to profit oriented (older node, cheaper)tamalero - Saturday, April 30, 2022 - link
AMD was castrated then by intel's monopolistic tactics and being forced to use their supply limited and expensive foundry.I do not think Lisa would make the same idiotic mistake.
TheinsanegamerN - Monday, May 2, 2022 - link
Intel played a major part, but AMD also made multiple bone headed decisions, such as spending money on ATi or building globalfoundaries while their athlons stagnated, then rushing to release phenom when the core 2 landed and ROFLstomped athlong 64.GeoffreyA - Monday, May 2, 2022 - link
One never knows but I doubt it'll happen again that easily. Back then, AMD underestimated Intel, rested on their laurels somewhat, and weren't that aggressive with K10/Phenom. As Papermaster or someone said a while ago, it's going to be tock, tock, tock all the way now.Duwelon - Friday, April 29, 2022 - link
Hybrid cores are 99% pointless for desktops.brucethemoose - Saturday, April 30, 2022 - link
But extremely useful for laptops, and workstations.bji - Saturday, April 30, 2022 - link
I know. Amazingly I keep checking Anandtech for new articles every single day, not quite sure why I even do it anymore. I think it's force of habit.nandnandnand - Saturday, April 30, 2022 - link
Not if it improves multi-core performance per Watt and mm^2. But it will take you another 5 years to get that.TomWomack - Saturday, April 30, 2022 - link
Which is fine, since desktops are 99% pointless; for everything that isn't 100% CPU a laptop with a monitor works better (whether it's in a laptop-shaped case or a Mac Mini-shaped case), for gaming a console works better, for things that are 100% CPU a server works better.Oxford Guy - Sunday, May 1, 2022 - link
'desktops are 99% pointless'No. Desktops offer a form factor that makes it possible to have a large amount of area devoted to cooling.
Laptops are used far too frequently for workloads that demand too much from the cooling, resulting in very irritating noise.
TheinsanegamerN - Monday, May 2, 2022 - link
"for gaming a console works better"MMmm....nothing beats that 22 FPS buttery smooth experience consoles provide. LOL.
Wereweeb - Saturday, April 30, 2022 - link
Astonishingly incorrect. All programs are limited to some degree by Amdahl's Law. Combining the use of bigger performance cores to handle the master threads and of PPA-optimized "small cores" to improve perf/$ metrics is universally useful and a step forward in CPU design.You may argue that Intel's current implementation isn't very useful for g*mers, but that's because games are slow as hell to adapt to hardware changes - and honestly, no one really cares about what g*mers think?
Wereweeb - Saturday, April 30, 2022 - link
"Aktchually, not all programs!" Yeah yeah, shut up, you got the point, if it's running on a CPU it's likely to be bound by a master thread.Kangal - Saturday, April 30, 2022 - link
Hybrid tech adds latency and complexity that is unnecessary for a Desktop PC.Remember these are machines which are hooked directly to the power-socket, and prioritise performance for certain applications. So a "Standard" Setup like 8x Large cores in Intel's monolithic chips, is the best option.
On top of this, servers run differently to your standard Desktop PC and Mobile PC. They prioritise security above all. So a Large Setup like 64x Large Cores in AMDs chiplet design is the best option.
For Mobile uses, there's a decent range. From the 4W Thin Tablet devices, with say 4x small cores. To the 20W Slim Laptops devices, with say 2x Large Cores and 4x small cores. And up to the 40W Heavy Notebooks, with say 4x Large Cores and 4x small cores. In these circumstances, Hybrid Computing as developed/perfected by ARM and Apple is the best option.
abufrejoval - Saturday, April 30, 2022 - link
you seem so very sure, can't say I share your belief. I have a Ryzen 5800U in an ultra book where it obviously struggles to clock all eight cores beyond 4 GHz at both 15 and 28 Watt TDP presets, but I am doubtful that replacing the last 2 Zen 3 cores with something of an E-core equivalent would yield additional compute from the same power/heat budget except in some purpose built synthetic benchmark. In a 24x7 power optimized edge or home server I can imagine assigning certain tasks to E cores to let P cores slumber. But then I might even want some control over RAM, because that can easily take a bigger bite than core power savings with DIMMs.Kangal - Saturday, April 30, 2022 - link
Well, what I said is true as long as all factors are controlled.So that means each chipset is using the: same software (eg Windows10 Pro), same lithography (eg 6nm), same architecture (not huge difference like Zen1 vs Zen3+).
A large server that you use as a Cloud, needs to handle lots of queries at the same time. Here security is paramount, followed by multi-threaded performance, then efficiency. So AMD's chiplet design which can hold many cores in one chiplet, and combine many chiplets into a large chipset works great. You can greatly increase multi-threaded performance this way, and it won't decrease the efficiency by much, and will keep things secure. If one user requires more performance (or cache) the system might steal resources from another core, and this will take a hit to latency. But this latency is acceptable since there's a tertiary latency with the external connection between the User and the Server.
On a Desktop PC, something like Intel's Core i9-9900k works well. And that's running on an outdated +14nm node, on an outdated architecture (Skylake), which was able to match a new architecture (Zen2) on a 7nm node. If you were able to refresh that with a newer architecture on a newer node, it will perform better against the chiplet design of AMD. The reason is that on a monolithic design, the whole chipset is stamped out by the same good node and it is a closer integrated system. Since this is a computer for a single user, performance (single-core) is much more useful than outright total/multithreaded performance and the system is connected to a power socket. So some efficiency is traded for more performance (ALL Large Cores).
Now stepping down to Heavy Notebooks, here it is expected to run on battery power. And occasionally be plugged into a power socket to extract more power. For that reason, it is very beneficial to use small cores for Power Efficiency and Balance them out with some Large Cores. Since this is a single user, there's no concern about security between each cores/packages, so having a weird Hybrid Design is not very important.
Moving down to Slim Laptops, here it's expected to be used exclusively on battery power. However, the nature of the form-factor does allow the installation of a decent-sized battery and a pretty potent cooling solution. So there is a pretty even balance between having small cores and Large Cores. Again, inter-core security isn't a problem here, and it does reduce efficiency a bit... but it becomes a benefit by providing ample performance at acceptable power drain.
At the lowest end of the spectrum, at least for x86, are Thin Tablets. These use a passive cooling method, and due to battery/heat concerns don't charge particularly fast. Here Large Cores can come to the detriment of the experience. It might deliver an initial jolt of performance, but it will use up a notable amount of battery power, then heat up the device, which causes it run slow for the remainder of that duration. So the most optimal solution is really to use all small cores solution, for maximum efficiency.
lmcd - Sunday, May 1, 2022 - link
Congrats Kangal on typing a lot of words. They're still not accurate though! Your cursory understanding of chiplet design doesn't make you an expert.You have decided that heat is completely irrelevant to a desktop or server-class design. This is completely wrong. One of the big challenges when Zen 2 came out was that the chiplets produced hot spots that were extremely difficult to cool because the small dies did not disperse heat across a continuous silicon surface. Adding little cores is a perfect solution because they do not dramatically change the architecture of the tile/chiplet, add silicon to allow better heat dispersion, and also do not dramatically change the power profile of the tile. The net result is that the little cores are basically "free" in a tiled/chiplet architecture.
If it wasn't obvious, this is part of why AMD is also releasing Zen 4c.
Kangal - Sunday, May 1, 2022 - link
Thank You.I am an amateur at best, much like most commenters here such as yourself.
Where did I say heat was irrelevant, because I can't see it up there. Desktop PCs are designed for One User with Maximum Performance in mind. That is why they are connected direct to a power-socket, physically large, and have highest cooling system (per user) compared to the other options. Yes, Desktop PCs cool more than Servers because Servers run parallel many users at the same time, and they are more balanced due to budget constraints (electricity).
Not sure about what Epyc/Zen2 issues you speak about. They may have been isolated or early day issues, but I haven't heard anything about it. You can read the papers released by AMD on the topic of optimising thermals here:
https://www.cs.utah.edu/wondp/eckert.pdf
https://arxiv.org/pdf/2108.00808.pdf
Your statement about Zen 4C is completely wrong. There is very low information about those cores, so anything you say is just misinformation. They could in-fact be "small cores" as you infer, usually achieved by taking an older architecture (eg Zen1) to miniaturising and optimising it for efficiency and area. That's what Intel did with their Pentium-Celeron-Atom cores, and later with their Skylake-eCores. Or these might actually be the same Zen4 cores, but they are using better silicon, packaged differently on-die, and are better voltage regulated. They may even have cut-down on the L1/L2 cache slightly, or removed a bit of processing/hardware which is not useful for Cloud Computing. At the end of the day, it is all speculation.
lmcd - Monday, May 2, 2022 - link
"On a Desktop PC, something like Intel's Core i9-9900k works well. And that's running on an outdated +14nm node, on an outdated architecture (Skylake), which was able to match a new architecture (Zen2) on a 7nm node. If you were able to refresh that with a newer architecture on a newer node, it will perform better against the chiplet design of AMD. The reason is that on a monolithic design, the whole chipset is stamped out by the same good node and it is a closer integrated system. Since this is a computer for a single user, performance (single-core) is much more useful than outright total/multithreaded performance and the system is connected to a power socket. So some efficiency is traded for more performance (ALL Large Cores)."There's two options for that hypothetical die. The die shrink with no other changes leads to no performance gain. Ballooning the die size back to the same area on a monolithic design just leads to leakage, horrible binning, and resulting heat increase. The same can be said for clockspeed increases. The only valid extrapolation from the paragraph I quoted is that you don't care about efficiency when scaling the design up. Otherwise it would be obvious why the 9900K is a dead end.
As for the cooling issues, hello and welcome! https://www.igorslab.de/en/cpu-heatspreader-in-det...
These issues are not isolated. They are not yet solved. I am sure they can be conquered but it's nontrivial. Point is -- building dies to better diffuse heat is a must when the "uncore" elements that used to add cooling surface are separated out.
Zen 4c is guaranteed to be smaller than Zen 4. I can guarantee that AMD is not wasting any space on its Genoa platform in either of its high-end configurations using Zen 4 and Zen 4c. The expected difference is cache, but that's still a component of the core.
I'm a bit astonished by the inaccuracies in your Intel history section. There's no such thing as a Skylake eCore. Intel drew upon P3 for early Atom designs but Silvermont and later are dramatically different.
michael2k - Saturday, April 30, 2022 - link
Not true. If you’ve got 30 threads and 26 of them are time or performance insensitive then you can schedule them on the efficiency cores. Doing so gives your performance cores a larger slice of the power budget, which in turn means you performance sensitive tasks can run faster and hotter.IE a file copy, a network stack, an indexer, and a virus scan can all run on a set of low power cores that only sips 5W, leaving you main app the remaining 145W; the alternative is to split your performance cores in those tasks.
JayNor - Saturday, April 30, 2022 - link
Intel's Alder Lake eight e-cores are reportedly more performant than the four cores in my three year old NUC box... so looks like they have a point to me. The Raptor Lake chips are supposed to be drop-in upgrades, adding 8 more e-cores, so looks like a good upgrade path to me.I've also read some comments about using process lasso or a windows power profile as a way to effectively go fanless, so I'd be interested in trying to run exclusively on e-cores and silence the fans when I don't want to hear them.
JayNor - Saturday, April 30, 2022 - link
as a reference, wikichip fuse has an article"Intel’s Gracemont Small Core Eclipses Last-Gen Big Core Performance"
which presents Intel slides on Gracemont relative performance vs the Skylake cores.
hecksagon - Tuesday, May 3, 2022 - link
Not at all. If I can fit 4 efficiency cores in the same area as 1 performance core and get greater multi-threaded performance than that performance core then it definity has merit on desktop. At the end of the day we are still die size limited. Getting more cores on that die is certainly useful.abufrejoval - Saturday, April 30, 2022 - link
Guys, what's going on?AT has become a real ghost town with the highest activity being some robot "Deals".
Really great writers left without parting words like Johan de Gelas or Andrei Frumusanu, some parted with words, but even the last full timer had very few things to say for some time now.
Do you have a legal obligation to keep the site running or do you actually believe you can get out of this stall?
Closing AT would be terrible, but it doesn't get any better by dragging it out like a 19th century opera death.
flgt - Saturday, April 30, 2022 - link
You know what's going on. No one pays for online content anymore. They are at the mercy of their advertisers. All those writers were good guys and probably made good money going into industry. It is what it is. No point in harassing the guys working their ass off to try and keep it going.tamalero - Sunday, May 1, 2022 - link
Agree. No surprise if they moved to Youtube and Patreon to support their writing, leaks and information. Just like how GamerNexus and similars do their services.TheinsanegamerN - Monday, May 2, 2022 - link
Asking what the plan is isn't "harrasment" and that spineless wimpering is what leads to the internet becoming so sanitized that nobody can make a living on it anymore.web2dot0 - Saturday, April 30, 2022 - link
You are asking people with expertise and skillset that qualifies them to get paid well, but you expect them to work for free ... because you are entitled to it.🤷🏻♂️
Why do you think people have moved to TikTok / and Clickbait YTube video in 10-15mins format rather than doing any deep dive or critical analysis of any kind?
Don't blame the player, blame the game.
Kangal - Saturday, April 30, 2022 - link
Perhaps they should follow the direction of Linus Media Group. Have a bunch of professionals who are great at the difficult technical stuff. Then have a handful of people that have great personalities in front of the camera. And drive the business first by info-tainment videos, and follow up with the professional articles.Otherwise, when all the great people at AT leave and are replaced by bad actors, the important but/heart has been replaced. You've essentially lost that company to obscurity, ignorance, or advertising for big corporations.
PeachNCream - Sunday, May 1, 2022 - link
A format pivot like that requires vision, leadership buy-in, and funding. There is also risk that AT will not be able to shoulder aside existing media-centric companies such as LMG that already have built an extensive fan base. Future (or whatever company owns AT now) may not have the interest of financing to support an effort like that and my own opinion is that they waited too long and are simply not in a position to acquire the cash to make a pivot.erotomania - Sunday, May 1, 2022 - link
It's too late for an Anandtech pivot. The talent drain due to the funding drain wiped out that chance a while ago. Who is left? Ryan, Gavin, Brett, E.F? They don't all live on the same continent so I think a LMG pivot is out. If Ryan is smart, he's already working his next industry job right now. The Purch/Future money is drying up, the budget was slashed a long time ago, the owners don't care, the ads don't pay....This game is up. I hate it too. But it is.
Oxford Guy - Sunday, May 1, 2022 - link
And we're left with sites that are devolving, with things like the post hiding and political nonsense at Ars. That was never a great site but it has drastically declined in terms of the ability to have a meaningful conversation — due to the implementation of the voting game. Extremetech has gay-bashing as part of its main course now. The admin there even speaks directly to the person using a bigoted alias, as if there's nothing strange about it.Anandtech is one of the few remaining bastions of civilized discourse and it seems that's no longer what people in this world are looking for. They don't want to debate without absurd Orwellian games like post hiding voting systems. Echo chambers are not interesting to me.
nandnandnand - Monday, May 2, 2022 - link
Where is the "gay-bashing"? I don't go to Extremetech often enough to have seen it. Sounds better than Arse Technica: Home of the DrPizza.TheinsanegamerN - Monday, May 2, 2022 - link
Well, typically "gay bashing" usually means something along the lines of " I'm not a huge fan of the alphabet group".Ars Reddita is far worse, with the hiding of the Dr. CheesePizza and using the same lame voting system that turned reddit into an insane echo chamber of losers.
Oxford Guy - Monday, May 2, 2022 - link
‘Well, typically "gay bashing" usually means something along the lines of " I'm not a huge fan of the alphabet group".’Use that absurd formulation to try to justify posting bigoted comments on the basis of race, sex, and religion and see how far it gets you.
People who are unable to appreciate valuable members of society due to their unnecessary false beliefs can choose to keep their opinions to themselves. Not being a ‘fan’ doesn’t entitle them to attack people.
Oxford Guy - Monday, May 2, 2022 - link
‘Where is the "gay-bashing"?’Someone has been posting there since February under the name GodSavesH0m0sNo (and various permutations). Joel Hruska, the site admin, even chatted in the comments to that person amiably, as if it’s totally fine to have a portion of his readership under attack.
I know someone who wrote to the parent company and was completely ignored.
TheinsanegamerN - Monday, May 2, 2022 - link
Anandtech cant churn out reasonable content anymore. Their primary base wanted GPU and CPU and Case reviews, something that younger creators are better at addressing. Anandtech's older audience likes the deep dives, but anand has lost much of the manpower that makes it happen.Stop blaming Anandtech's demise on comment wars.
GeoffreyA - Monday, May 2, 2022 - link
"And we're left with sites that are devolving"I don't view the comments on those other sites, and hardly ever visit them anyhow, but think the internet as a whole has degenerated. What is bad and vain in human nature, just springs out amplified. Possibly, the less we read online, the better. I wonder what further steps down this spiral staircase the so-called Metaverse and Web 3.0 will take us.
Oxford Guy - Monday, May 2, 2022 - link
‘the internet as a whole has degenerated.’The Internet has exposed human corruption schemes across the globe. A major psychological response has been for people to use verbal warfare to try to fill the void where their faith in human goodness used to be.
Sites also use design to encourage the warfare, like the voting + comment hiding systems. Those change the discussion process into a game where shunning is the payoff.
GeoffreyA - Tuesday, May 3, 2022 - link
I also think it's given people the opportunity to play amateur philosopher, writer, and slactivist. Social media, infinite scroll, dopamine-producing design, etc. surely can't be good for the human brain.And yes, I agree and have a low opinion of comment-voting systems, which seem to have the effect of suppressing unorthodox, out-of-the-way views.
Oxford Guy - Tuesday, May 3, 2022 - link
It’s the implementation that matters. People can be mistreated in a forum that doesn’t involve technology. The implementations on the Internet continue to become more China-like, more gamist (such as enticing people to form cliques to shun others via downvoting) and more fraudulent.For instance, Disqus posted an announcement years ago that claimed that the best way to improve a forum is to have more and more censorship. The same announcement claimed that that isn’t censorship at all. The company rolled out the feature of people thinking their post is being seen by others but it is only visible to them.
Post hiding is becoming very common and it is always implemented in a deceptive manner. The most basic fact about that is that argumentum ad populum is a fallacy.
People are being trained to think it’s normal for someone’s input to not be obviously blocked or erased (so that the censorship is able to be evaluated) — but instead to always face the possibility of being erased/blocked without anyone knowing. It’s an extremely Orwellian situation.
Social media is what people make it. It’s not good nor bad by default. It also makes a difference — who benefits and who suffers. So, the opinions about what is ‘good’ and what isn’t usually come down, as always, to money talking. Exploitation.
Oxford Guy - Tuesday, May 3, 2022 - link
Your citation of infinite scrolling, though, is spot-on. I won’t use anything that involves that. I don’t enjoy feeling like a lab rat.GeoffreyA - Wednesday, May 4, 2022 - link
I think the more censorship there is in any field, the more blind people become to it. At length, they believe there isn't any censorship; for the orthodoxy has prevailed in snuffing out individuality. Also, the deceptive design tactics we're seeing across the industry are perhaps a reflection of dishonesty, increasing dishonesty, in the people behind it.On argumentum ad populum, well, that's just another instance where democracy can play tyrant; the fire of general opinion doesn't look too kindly on lone, unorthodox points of view. On the Brave New World vs. Orwellian question, I'm inclined to think the world has tended more to the former, with some touches of the latter. It's easier to control people with the perfumed path of pleasure than dictatorship.
GeoffreyA - Wednesday, May 4, 2022 - link
The lab rat is a good analogy for infinite scroll!Oxford Guy - Wednesday, May 4, 2022 - link
Argumentum ad populum is worse than democracy in a forum setting because the design and the administration select for the population that votes. People who don’t like the design and/or the behaviour/attitudes of the admins (which often includes favoritism and sock puppetry) will go elsewhere. That leaves an echo chamber.In regular democracy you get the problems of popular ignorance and groupthink. In a more controlled setting like an online forum, you get an increasingly selected population who often think that narrowmindedness is a good thing. Sometimes it is, such as for specialized professional discourse. Many times, though, it’s a disaster like Ars.
GeoffreyA - Thursday, May 5, 2022 - link
Well said.abufrejoval - Sunday, May 1, 2022 - link
Only one problem: LMG is video.For some things video may be better, but most of the time they just waste my time and patience.
I've seen some videos stretching what would be five lines of text into 15 minutes of relentles powertalk. And even if I grant Linus a certain charm, I can only tolerate him in very small doses and I still read much faster than anyone can speak.
For the info I want, text is so much more efficient, and if it's well done, much better abstracted.
I subscribe to some news sites, but logging in means personal tracking, which I dislike.
And Charlie is just mad in terms of pricing.
Off the rails, too, from time to time: his rant on the 5800X3D shows that even his well formed prejudice doesn't always win over hands-on trying first.
nandnandnand - Monday, May 2, 2022 - link
Forget Linus, Gamers Nexus and Hardware Unboxed are where it's at.As for AnandTech, Tom's Hardware is its "sister site" (same owner). The articles are typically of lower quality, but they cover some interesting stuff not found here and they got a 5800X3D review out on time.
edzieba - Monday, May 2, 2022 - link
Games Nexus also have the same issue of video-only for the vast majority of their output (occasionally there will be an 'article' that is a video transcript).TheinsanegamerN - Monday, May 2, 2022 - link
GN's dropping of their written reviews on their website still sucks.GeoffreyA - Monday, May 2, 2022 - link
Sadly, many prefer to watch a video than read an article.name99 - Saturday, April 30, 2022 - link
Never forget this golden oldie: CES January 2017https://www.itpro.co.uk/laptops/27867/moore-s-law-...
There's apparently a MASSIVE gap between Intel demo's (even demo's of an actual device, let alone mere "running in the lab" claims) and shipping in volume.
And note that they aren't telling us the transistor density of these tiles (though it surely won't change much between now and ship)? Why? Probably because the message is supposed to be about how i4 achieves a density of 200MTr/mm^2, but the actual tiles are probably 80..100.
Tom_Yum - Saturday, April 30, 2022 - link
Exactly, I remember Intel touting it's 10nm transistor density was 100MT/mm2, yet after the dud Cannon Lake they stopped announcing both transistor density and transistor count (so it couldn't be calculated), yet the best estimates for Tiger Lake was that they were hitting around 50MT/mm2, less than TSMC's 7nm. Fact was most of the +++'s to 10nm process was about increasing gate pitch (not reducing it) to hit higher frequencies and power limits, at the detriment of transistor density.Silma - Wednesday, May 4, 2022 - link
So, if all goes according to plan, Intel will produce 7 nm in 2023 when Apple will probably switch to 4 or 3 nm ?Blastdoor - Wednesday, May 4, 2022 - link
If this (https://www.digitimes.com/news/a20220503PD216.html... is true, it doesn't bode well for Intel's confidence in Intel 4. Also reinforces my view that all of Intel's claims are suspect until they actually ship.The Von Matrices - Wednesday, May 4, 2022 - link
I don't understand. The post about the "first disaggregated product" contains a photo of a woman holding a package with a single die. How is that disaggregated?evilpaul666 - Sunday, May 22, 2022 - link
So much glue.