https://www.tomshardware.com/features/amd-vs-intel-cpus

The newest AMD Ryzen 5000 wave and the newest Intel hyperclocked Rocket Lake responses have hit the benchmarks, and the net picture is very much more of the same old same old stuff.

Each set of stair steps is a little higher than the last set of stair steps with AMD coming out on top and at least one instance of Intel hyper-clocking their latest stuff hard enough to show up on the list to claim to be "competitive" using over twice the watts of power.

Biggest real world difference between AMD and Intel is that AMD is rated using the cooling system that comes in the box with the processor, while in some cases THERE IS NO COOLER SYSTEM PHYSICALLY AVAILABLE to mate up with the exaggerated claims Intel is making for their latest BS marketing Rocket Lake.   With no real high heat capable cooler system available for laptops, you are actually buying off some bogus Intel specialty test house test results that will not be even close to real in the real world product you are actually purchasing.

The following chart is all about GAMING (Intel's best performer segment) --- and the best performance per dollar chipset value out there is the AMD Ryzen 5 5600X PBO ar $300.



Recommendation is to buy AMD if you can find it.  AMD is getting snatched up world-wide at a point of sale marked up premium prices right now compared to Intel.   AMD still sells direct at recommended retail off their own AMD webpages, but that source of list price supply is very spotty on the best value AMD chipsets (they are instantly sold out).

AMD is making and shipping their entire allocation set of different TSMC wafer allocations worth of finished products, but this is still leaving enough unfilled demand on the table for Intel to continue to survive very nicely.

RUMOR TIME

AMD begins taking early 5nm chiplet shipments from TSMC.

Call them sample lots if you wish as they are indeed full lot runs of chiplets but they are not coming in a constant flow supply fashion.   Apple is releasing some of the 5nm flow they had locked up for two years now as Apple rolls on down to 4nm and 3nm.   AMD is getting part of this old Apple 5nm production flow and is in turn shipping it out as test samples to their major partners so they can get a leg up on the next future wave of AMD.

AMD may do a variety of things with these new more efficient, faster chiplets.    We shall see what is built with them and how the ship quantities ramp up in the real world as AMD gets their full flow of 5nm rolling out into the market place.

Remember, AMD is going to continue to use all of their older lithography allocations ongoing  in a full speed ahead fashion, so as to be able to take more contested market share away from Intel.   This will continue until Intel finally manages to put out something that can actually beat the oldest AMD 7nm chiplet based processors .......   and Intel 14nm simply isn't up to that job.


==================================================


Intel is attempting to buy out RISC-5 design house SiFive ---- "Why?"  you ask.    Simple, they want the most powerful of the SiFive chip designs to be their Intel "little" core as they have no appropriate little core of their own.

SiFive had recently approached Intel to get Intel to run their chip designs for them, so Intel immediately turns on them and tries to do a hostile take over .......  
Got any moral there for the rest of you little guys who might have some tech that is worth stealing ?????    Something simple like don't dangle a hot dog in front of Godzilla when he's hungry ??????

SiFive P550 is the equivalent of the ARM A75 core and would make a good little core for Intel.   The bad news for SiFive is they go from being a small industry leader to ...... nothing, as Intel will spread their people out among their design groups and promptly hack up everything that SiFive had innovated into existence right back out of existence.

https://liliputing.com/2021/06/sifive-performance-p550-is-its-most-powerful-r...

SiFive has some production allocations at TSMC 7nm level, and Intel may want that as well.

Intel makes the front page of the nightly business news for losing somewhere between 4% and 14% of their market share.

Next day, Intel abruptly fires their current mainframe head dog and churns their entire internal organization structure yet again for the 4th time since they fired their old errant bean picker in chief.

Intel dudes, based on your Board of Director's statements these periodic purges will continue until you guys start doing something right .......

One questions if this endless game of musical chairs is really helping Intel do any long term good stuff  ??????    At some point in time lining them all up on the edge of the aqueduct  and pushing every 10th one off with a spear butt becomes sorta counter productive.

The guy with the spear butt keeps saying the same thing while sending his solders over the edge down the long drop .......  "Get off 14nm.  Get 7nm working right."


===================================================


As Huawei doubles down on making their own chipset building equipment (Huawei has no choice, Trumph's entity listing says no state of the art ASML scanners nor any other western tech for Chinese Military linked businesses).   At the same time Intel acts to slowly decimate its own business groups just to get them to move off the dime technologically, all the while Huawei is also busy floundering about seeking a future pathway.




So perhaps it is time for us to stir some rumors about the future of electronics past 2mn.

IBM and Samsung have built some normal style 2nm tech samples and some 2nm Gate All Around samples.  Both of these sample runs ran into signal loss through the tiny connecting wires going to both styles of processors, making the 2nm results a lot less reliable than 3nm & 4nm results.
Soon "smaller" will not be better any more.


===================================================


At this same time AMD has announced they will be building distributed AI and distributing upper level graphics into the guts of every chiplet.  

In essence AMD will be distributing their massive GPU and AI and the localized chunks of systems memory at the localized processor chiplet level keeping it actually located inside their CPU structure.   At 5nm and below, the new extra available flat real estate on the socket makes this expansion possible.    Boosting the L-3 and L-4 memory up vertically on top of the top of the die also helps make more room for this functional AMD expansion.   By keeping most of the action close by inside the chiplet processor itself AMD is avoiding this growing issue of large signal loss due to very small fine connecting wires.  

And this will actually work out to be both faster and better for AMD, as these new chipsets do not need to communicate long distances across a motherboard or through complex buss structures or through secondary communications chipset subsystems all of which stuff always act to slow down the overall data flow.   Aside from the exterior inputs and post computational outputs, everything remains inside the CPU itself.

Ampere and others have recognized these same small wire leakage issues in their own testing and they have developed their "full wafer sized chipset systems" to put 100% all of the computer into the resulting huge wafer sized chipset itself.   AMD is doing the basically the same thing, just using their superior chiplet design in aggregate as a distributed full computer system with very close by local memory and GPU and AI distributed clusters.   The AMD system is much better in making less testing scrap, as a defect in the Ampere system can scrap an entire wafer, while several flaws in the  proposed AMD system result in very minimal chiplet scrap.

Base upon this generic issue of 2nm signal loss due to fine wires, AMD clearly sees that they only have 2-3 future lithography shrinks left for them to utilize ....... and they also see Intel in the background kinda sorta matching AMD's current high end performance levels using much much larger and much cruder lithography running at relatively huge power draws, all the while using lots and lots of sneaky cute little AI tricks just to get the job done "fast enough".

AMD can do the same sorts of AI tricks in their much faster processors, but they will approach these tricks more openly if they go there as AMD/Xylinx owns all the AI tech base patents that all these tricks depend upon.

Intel does not own this tech at all, AMD owns it  ---- thus the Intel sneaky "don't ever talk about it" approach Intel is using.

What is also clear is the consumer need for general processors has doubled during the 2 Covid years and it will double again inside of 3 years as we recover from Covid.  

Intel and Global Foundries are still betting their older tech base is still actually good to go for as long as it physically lasts as the marketplace is telling them that OLDER LITHOGRAPHY IS BETTER THAN NO LITHOGRAPHY and that they can still sell all they can make at this point in time.   AMD currently gets to charge a sizable premium for their best in class chipsets, and all the others must discount their prices to move their inferior designs if demand eases up any at all.

Huawei is also taking this side of the bet, as they have their existing old AMD CPU licenses to produce 7nm and larger chipsets, and this is the tech level where Huawei will be building their new Chinese foundries using IBM/GF style old homemade lithography systems.

More Hot Rumors


https://consumer.huawei.com/ae-en/community/details/TSMC-ANNOUNCES-ITS-FIRST-...

TSMC announces roadmap plans for next 2 years

At TSMC’s 26th Technical Symposium, TSMC officially confirmed that its 5nm and 6nm processes are already in mass production. In addition, the company also announced that it will release a higher version of the 5nm process next year. Furthermore, TSMC officially confirmed that more advanced 2nm, 3nm, and 4nm is in development. As for the 4nm process, it is an ultimate improvement of 5nm. However, the 3nm process is the natural successor to the 5nm process.

In terms of technical indicators, the risk production of the 3nm process (N3) will take place next year. The mass production will commence in 2022. Compared with 5nm, 3nm will reduce power consumption by 25-30% and improve performance by 10-15%. The 4nm (N4) risk production is also scheduled to take place next year. Just like the 3nm process, 4nm mass production will start in early 2022. For TSMC N5 customers, the transition to N4 will be very smooth, which means that the tape-out cost will be greatly reduced.



https://www.tomshardware.com/news/apple-intel-set-to-use-tsmc-3nm-node-for-20...

Confirmation of 3nm rumors

Apple and Intel will be the first to adopt Taiwan Semiconductor Manufacturing Co.'s (TSMC) N3 (3nm) fabrication process when the contract maker of chips deploys in late 2022, Nikkei Asia news agency reported on Friday. Intel is projected to use the technology to make CPUs for PCs and servers, whereas Apple is expected to use the node for its system-on-chips aimed at client devices.  

Apple and Intel are currently 'testing their chip designs' produced using TSMC's N3 process, according to a Nikkei report that cites several people with knowledge of the matter. Given that N3 is about to 'officially' enter risk production mode in the coming weeks and assuming that the information from the news agency is correct, we can speculate that Apple and Intel have already finalized their N3 CPUs and SoCs. Though it is unclear whether they have functional silicon. Commercial production of these chips is set to start in the second half of 2022.

Intel is said to be preparing at least two products made using TSMC's N3 node: one aimed at notebooks, another for servers. There are no details about these processors at present, but Intel has already accidentally confirmed TSMC-made Xeon SoCs aimed at various niche markets. Intel confirmed months ago that it is working with TSMC on its 2023 products, but refrained from revealing any details.  


TSMC 6nm, 5nm and 4nm (Mediatek) are shipping volume production runs right now.   3nm starts at Apple in limited production starting in just a few months.   AMD picks up all the 5nm production that Apple lays down and will start picking up most of the 4nm and 3nm as they become available in volume also as Apple moves on down to a smaller lithography.

Intel has supposedly contracted a few simpler processors (some small and some large) to be run at 3nm to get their foot in the water.   Intel needs some 3nm bragging rights very very badly right now but jumping way ahead to try to run neck and neck with Apple on 3nm processes that were specifically developed and are being specifically tuned for Apple sounds like a fairly high risk endeavor for Intel.   Sounds shaky, in other words.  

Looks like Intel's Board has put a pistol to the heads of existing Intel management team and has very seriously told them that the Board will pull the trigger as many times as it takes to get Intel completely off 14nm ASAP.    One schedule slip or market reversal = one trigger pull.

Intel is also trying at the same time to say their in house 7nm will be almost as good as TSMC's 3nm,  something that sounds a lot more like some more of the typical Intel BS marketing misinformation.    This was BS back when Intel first said it about 5nm, and it smells worse when said about 3nm.  Remember, TSMC will cut Intel off if they keep on acting like a "spoiler" or "completely bogus acting" customer, so Intel should be somewhat mindful of their behavior.  

Remember, Intel is now a direct foundry competitor for TSMC, not just a simple customer.    As such, Intel is not altogether welcome at TSMC right now, having failed to utilize some contracted allocations when they became available in the very recent past.

Next, AMD has already contracted some sizable allocations for 3nm TSMC already, but AMD will use up these allocations up and then try to phase into any additional 3nm chiplet allocations they can get their hands on while keeping up all the applicable earlier lithography nodes of chiplets still flowing at full stream (as AMD is still very very  hungry for more wafers of chiplets).

Remember, AMD is pulling market share away from Intel just as fast as they can get the flow of appropriate chiplets to do it with.   AMD is a signed up money providing ground floor player in all of the new USA and new EU TSMC plants that are under construction right now.  
Intel is not a ground floor "paying player", not yet anyway.   Intel is acting like they are "entitled" right now due to Biden using them as front men for his new tech initiative, but Intel should also note that Biden has not given them their megabucks as he cannot legally do that apart from Congress passing appropriations and getting them approved by both houses.  

TSMC's recent actions to resolve the current automotive supply issues renders Biden's grandiose give away plans somewhat moot ......

SUMMARY

AMD 7nm beats Intel 14nm and most early Intel 10nm
AMD 6nm beats all Intel 10nm
AMD 5nm and 4nm will likely beat early Intel 7nm (whenever, if ever it arrives)
AMD 3nm and 2nm will likely be equivalent to Intel 3nm and 2nm, with the winner having the best integrated thermal design benefits.

Will AMD's new designs with distributed local graphics and distributed local memory and distributed local AI carry the day for AMD?  

We shall see .......



===================================================



https://liliputing.com/2021/07/qualcomm-plans-to-launch-a-laptop-processor-th...

Qualcomm intends to make a M-2 competitor chipset using their NUVIA tech which will take some additional share away from Intel as Intel owns that general market segment right now but is rapidly losing it to the Apple M-1 and M-2 chipsets at this time.



In an interview with Reuters, Qualcomm CEO Cristiano Amon said that he thinks his company will deliver a processor that’s competitive with Apple’s by mid-2022 when the NUVIA purchase finalizes.   This entails additional market share losses to Intel as well as to Apple proper.

There aren’t a lot of details explaining what to expect from that chip in terms of features or performance. But if you’re wondering how Qualcomm plans to catch up to Apple, it’s with the help of some former Apple chip designers that Qualcomm has hired.

Earlier this year Qualcomm acquired a startup called NUVIA for $1.4 billion. That company, which was formed by a team of chip designers with experience working at Google, ARM, Broadcom, AMD, and Apple.  NUVIA brings new, state of the art processor and AI tech to Qualcomm.


CLARIFICATION

Qualcomm is going to be using their new NUVIA designs to do this, as existing ARM designs are simply not as good as the new NUVIA processor designs promise to be.   ARM is expecting to counter against this NUVIA Initiative with some brand new ARM processor designs and and ARM will work very hard to beat NUVIA to market, thus preempting the Qualcomm initiative.

NVIDIA taking over ARM has hit up against the Chinese regulators now, and the Chinese are simply not going to respond to it at all, thus killing the take over when the final due date comes and goes.

https://liliputing.com/2021/07/intels-unpopular-lakefield-chips-reach-end-of-...

Intel dumps embarrassing "very first" Big - Little chipsets out of the lineup.

The Intel Core i3-L13G4 and Core i5-L16G7 processors are five-core chips that combine four low-power, energy-efficient CPU cores based on Intel Tremont architecture (which is basically a version of Intel’s Atom technology) with a single Sunny Cove CPU core (using the same architecture as Intel’s 10th-gen Ice Lake chips).

Theoretically this was meant to allow PC makers to build thin and light computers that offered long battery life while delivering performance on par with what you’d expect from an entry-level machine with an Intel Core processor.

In practice, only a few devices ever shipped with Lakefield processors, including the Samsung Galaxy Book S and Lenovo ThinkPad X1 Fold. The general consensus from reviewers was that you could get better performance from a device with an older Intel Core m3-8100Y processor (like a $630 Microsoft Surface Go 2) than you’d get from these devices, which sold for $1000+ and $2500+ at launch, respectively.

So in that sense, it’s not particularly surprising to see that Microsoft has issued a Product Change Notification indicating that it’s discontinuing its only Lakefield processors. What is a little surprising is just how quickly the company is doing that.

At the time Lakefield came out and got tested no one could see any advantage to Intel's Big Little Lakefield family.   It all felt like an ARM copying "me too" exercise that was actually carried out in a very poor, nonsensical fashion.

Intel has now excised the old team leader for Lakefield, with inference that he did something in error, but instead he did exactly what his bosses at that time told him to do.

Intel is also coming across as sort of a nasty place to be working, right now .........  on top of lying repeatedly to the public about their processors performance.



===================================================



Some new information comes to help explain the abrupt Intel Lakefield dump off and its replacement by Intel Alderlake processors.

https://liliputing.com/2021/07/intel-alder-lake-mobile-chip-lineup-leaked-up-...

Alder Lake processors will use Intel’s “Foveros” 3D stacking technology to package high-performance CPU cores based on “Golden Cove” architecture on the same chip as energy-efficient, Intel Atom-based “Gracemont” CPU cores.

Similar to ARM’s big.LITTLE technology, this allows chips to leverage the right CPU cores for the task at hand. Need more horsepower to complete a job? The big cores kick in and deliver a burst of performance. Running less resource-intensive jobs on your computer? The lower-power Atom cores can probably handle things, extending your laptop or tablet’s battery life.

Alder Lake isn’t Intel’s first crack at this sort of heterogenous computing architecture. The company launched an Intel Lakefield processor with one high-performance “Sunny Cove” core and four energy-efficient “Tremont” cores last year. But while it proved the 3D stacking technology worked, the chip delivered underwhelming performance and Intel has discontinued it just a year after launch.

If the leaked product slide is accurate, it looks like Intel is being much more ambitious this year, with a set of chips arranged across six different product segments, half of which are entirely new (marked by an asterisk below):

*Intel M5 – 5W-7W chips for tablets with Lakefield-like 1 big + 4 small core designs and 48 or 64 GPU execution units

Intel U9 – 9W – 15W chips for ultra-thin laptops including 2 big + 4 or 8 small cores and 80eu or 96 eu graphics (there may also be a 1 big + 4 small + 48eu chip in this range)

Intel U15 – 12W/15W/20W mainstream laptop chips with similar properties to the U9, but higher power consumption (and presumably higher clock speeds)

*Intel U28 – 20W – 28W “performance” chips with either 4 big + 8 small or 6 big + 8 small cores and 96eu graphics

Intel H45 – 35W – 45W “thin enthusiast” chips for gaming laptops and workstations with 4 big + 8 small or 6 big + 8 small CPU cores + 96eu graphics

*Intel H55 – 45-55 watt processors with 8 big and 8 small cores + 32eu graphics, likely because these chips are designed for “muscle laptops” where they’ll be paired with discrete graphics (there may also be a 4 big + 8 small core version)

Intel groups the M5 and U9 chips under the Alder Lake-M family, while the U15, U28, and H45 are considered Alder Lake-P chips. The H55 chips are are part of the Alder Lake-S family. While Intel usually restricts “S” series chips to desktops, it seems like the H55 processors are part of the company’s mobile lineup, despite being BGA socketed chips.


So far these chipsets look like a continuing execution of a family of FAIRLY EXPENSIVE Intel FOVOROS chipsets, same idea that got last year's first generation of the same abruptly dumped off last week.

Good luck determining how powerful the new Intel 14nm processors really are or how many watts of power they actually draw while doing whatever it is that they are doing ....... it sort of depends on whatever it you are using your machine for, in other words.  

Dynamic switching and dynamic throttling on the fly say that Intel processors are going to be a constant soup of "change up" depending on exactly how the laptop maker configures the machine this week and exactly how you are using it right now.

And this configuration and set up can be changed by broad brush strokes by a nightly over the internet update from Intel's servers !!!!?????

Holy built-in cover-up, Batman !!!!

https://www.sammobile.com/news/volkswagen-cars-using-samsung-exynos-auto-chips/

Volkswagen’s new cars are using Samsung’s 4mn Exynos Auto chips

All modern cars use semiconductor chips for various functions, including vehicle controls, internet connectivity, in-vehicle entertainment, navigation, and safety. A few years ago, Samsung expanded into the growing automobile industry with its Exynos Auto chipset, and the company has been supplying chips to Audi since then. It is now being reported that even Volkswagen, the world’s biggest auto brand, has started using Samsung’s chips.

According to a new report from South Korea, Samsung started supplying Exynos Auto chips to Volkswagen in early 2021. The chip is reportedly used for multimedia playback, navigation, and vehicle status control. The Exynos Auto V9 chipset is made using Samsung Foundry’s 8nm FinFET fabrication process. It features eight ARM Cortex-A76 CPU cores, an ARM Mali-G76 MP18 GPU, and an integrated NPU (Neural Processing Unit). Its DSP supports up to six displays, while its ISP supports up to 12 on-vehicle cameras and 4K 120fps videos.

Samsung is planning to expand its Exynos Auto chipset lineup in the future, given that more and more cars will depend on chips for data processing, image and video processing, internet connectivity, voice recognition, and other tasks. Samsung recently unveiled its first ISOCELL image sensor that is specifically designed for automobiles. The company also makes OLED screens, LEDs, and batteries for electric cars, and they are used by BMW and a lot of other auto brands.

Since more cars are bound to switch to full-fledged operating systems such as Android Automotive, QNX, and Apple’s OS (for its rumored smart car), sales of semiconductor chipsets are bound to rise in the near future, and Samsung wants to be fully prepared to cash in. The company is already working with Hyundai, Genesis, and Tesla on various projects.

Intel has promised a 2 year plan to meet automotive needs by building new Intel facilities local in Germany and in Japan.

Intel's grandiose plans are $20 Billion dollars short and two years way way too late.

Samsung, predicted to take the #2 worldwide leadership position in building chipsets away from Intel inside this year, has rung in with Volkswagen already as their main chipset supplier ---- deal is done, finito, completed right now.    TSMC is already the world's leading foundry, so this does not change either with Samsung booting Intel away from #2 position.   Question is, can Intel  hold on to #3 or is it fated to lose that as well?

Done deal, past tense, water over the dam --- Exynos is coming inside Volkswagens starting RIGHT NOW.    Volkswagen is the world's largest car maker, so Samsung now has the Automotive Leadership in chip making all sewed up already.

Samsung and TSMC have been tasked by world governments to end the Automotive Chipset Shortage ASAP and Samsung and TSMC has a scheduling plan to do that within the next quarter that does not involve building new buildings nor costing several 20 billion dollar programs like Intel wishes the EU and Japan to do.

Samsung and TSMC will simply short cell phone production while supplying Automotive Chipsets as a government mandated priority.   This will be all be done inside 4 months.

Samsung swinging over to full on automotive opens up new business potentials for Mediatek and Xiaomi, and Samsung is actually helping Mediatek to move into the larger position in cell phones to stave off Xiaomi's meteoric rise.   Samsung/Mediatek/Xiaomi is a powerhouse trio that is pushing Qualcomm aside quite quickly in cell phones.

You will start to see more ARM computer chipsets and ARM automotive chipsets starting RIGHT NOW.  

Samsung's consortium is moving in on Intel, taking the new markets away from both Intel and Qualcomm.



Intel is a joke and is being treated as such by the EU and Japan.    

Biden's putting Intel in charge of the USA's computer recovery plan is a grave mistake as all Intel is doing now is the same old same old Intel BS self-delusions that are now being pushed over on others as Intel grabs for all the free cash money they can get.

https://seekingalpha.com/news/3715822-intel-in-talks-to-buy-globalfoundries-f...

https://www.wsj.com/articles/intel-is-in-talks-to-buy-globalfoundries-for-abo...


This is a hoot .......

Biden has put Intel in charge of renovating American chip building.   Biden is a weak bulb mentally, and that was all he could think to do, go give snow shovels worth of cash to Intel BS people who will waste it just like they have wasted their own money over the last 10 years.

Intel is not even in the ASML 2 year waiting line to buy any new state of the art lithography equipment.
That ship sailed over a year ago while Bob Swan was still in charge of Intel.

Intel has now burned all its bridges now trying to "consort" with TSMC and Samsung, both of the more modern larger chipmakers are now treating Intel's feeble attempts to get back in to the foundry business like they should treat any brand new weak kneed budding competitor.  

TSMC will now treat Intel like a pay up front high risk customer, which is what Intel is in reality.    Samsung really isn't working with Intel at all at the moment, their relationship was strained to breaking from past "projects" that went south due to inaction on Intel's part.

Intel has some current deadlines from the Biden Administration that they are failing to even make a dent into.    No progress, no more Biden money, in other words.

So, Intel is now trying to secretly buy the corpse of Global Foundries from the Arab conglomerate that currently owns it.

Why?   To meet Biden's deadlines, Intel will GO ABSOLUTELY WASTE ~30 billion~ American Tax Dollars on something that is just about totally worthless for any future oriented uses.

But it is American soil based "production" which is what Intel was charged by Biden to go increase.

This is a bullshite trick by Intel, as the capacity was there last year under Global Foundries and it does not constitute a singe wafer of new capacity that wasn't there a year ago.  

Saying Global was Arab owned and  was "non-American production" is really really really BS stretching things a whole whole lot ..........  
Intel's specialty -- brown vaporous BS

Saying "it is good for automotive uses" is pretty much wrong headed too, as Volkswagen for example doesn't want to use Global at all, they want to use brand new 4nm tech direct from Samsung.    Not any 15 year old moldy 20nm tech from Global Foundry.    Remember, VW is the world's largest car maker now a days ......