Intel unveils technology roadmap through 2024, Raptor Lake is coming later this year

[Jimmy2x]

Forward-looking: Intel CEO Pat Gelsinger gave attendees of Friday's 2022 investor meeting a look at the company's strategy for growth and technical advancement over the next several years. The presentation touched on topics ranging from upcoming datacenter technologies to plans for automotive-focused foundry services. It also provided attendees with new details surrounding Intel's server and client computing roadmaps.

According to the milestones presented in the meeting, the first of the Intel 7 process node-based product line, known as Sapphire Rapids, is poised to hit the market in the first quarter of 2022.

Later products will include Emerald Rapids in 2023 and new P-core and E-core configurations for 2024's Sierra Forest and Granite Rapids product lines.

Intel's Client Computing Group (CCG) provided attendees with even more upcoming technology to look forward to. Raptor Lake, the next generation of Intel's desktop processors built on the Intel 7 process node, was put on display in a live demonstration and is planned to ship sometime in the second half of 2022.

The company claims this new generation of CPUs will provide double-digit performance increases and greater core and thread counts while allowing users to remain on Alder Lake's current LGA 1700 platform.

The presentation identified several more generations of Intel CCG's processors that will be built on the upcoming Intel 4 process node. These include the Meteor Lake product line in 2023 and the follow up Arrow Lake product line in 2024.

Beyond 2024, the company plans to leverage their IDM 2.0 strategy to support Lunar Lake and other future product lines.

Also read: Intel Arc laptop GPUs incoming, desktop GPUs pushed to Q2 2022

Intel also provided information on the highly anticipated Intel Arc graphics lineup. The company expects to ship upwards of 4 million discrete GPUs in 2022 alone. This ambitious number includes Arc Alchemist GPUs in Q1 2022 for laptops and graphics cards for desktops and workstations sometime in the second and third quarter. "Celestial," Arc's ultra-enthusiast tier, is also said to be in development, however no timeline for release was disclosed.

Permalink to story.

https://www.techspot.com/news/93470-intel-unveils-technology-roadmap-through-2024-raptor-lake.html

 

[passwordistaco]

24 cores with 32 threads...

The only configuration I can think of that would match is 8 P cores (16 threads) and 16 E cores (16 threads).

Intel really likes their E cores, I guess.

Edit: I guess for Intel chips with P and E cores, the # of P cores = threads - total cores

 

[nodfor]

24 cores with 32 threads...

The only configuration I can think of that would match is 8 P cores (16 threads) and 16 E cores (16 threads).

Intel really likes their E cores, I guess.

It probably will be something like that.
More P cores won't give a significant boost in gaming perf, or in apps that don't scale well with more cores, so they use e-cores to increase perf in multi threaded apps while keeping power consumption in check

 

[Puiu]

It probably will be something like that.
More P cores won't give a significant boost in gaming perf, or in apps that don't scale well with more cores, so they use e-cores to increase perf in multi threaded apps while keeping power consumption in check

More like they can't put more than 8 big cores in that package size without destroying the perf/W. They are forced to use E cores.

 

[Sausagemeat]

Core count is irrelevant. Intels quad core i3 12100 is about the same in multithreaded performance an 8 core R7 1700 according to the benchmarks. And that’s what we should look at, the actual performance of the chip. Another example is a 6 core Ryzen 5600X is very similar to an 8 core 3700X in terms of multithreaded performance.

Judging CPUs on core count is like working out how fast a car goes based only on how many cylinders the engine has. You can’t.

Besides how do we even know these parts have all these cores? Because windows tells us? Just a few years ago one of these manufacturers were fined under a class action lawsuit for lying about the core count of their chips and these chips reported in windows with the false amount of cores yet they performed similar to chips which had a core count of half of that advertised.

please, let’s stop focussing on the core count numbers we are given and look at the actual performance you would get instead. It makes much more sense.

 

[Hardware Geek]

Core count is irrelevant. Intels quad core i3 12100 is about the same in multithreaded performance an 8 core R7 1700 according to the benchmarks. And that’s what we should look at, the actual performance of the chip. Another example is a 6 core Ryzen 5600X is very similar to an 8 core 3700X in terms of multithreaded performance.

Judging CPUs on core count is like working out how fast a car goes based only on how many cylinders the engine has. You can’t.

Besides how do we even know these parts have all these cores? Because windows tells us? Just a few years ago one of these manufacturers were fined under a class action lawsuit for lying about the core count of their chips and these chips reported in windows with the false amount of cores yet they performed similar to chips which had a core count of half of that advertised.

please, let’s stop focussing on the core count numbers we are given and look at the actual performance you would get instead. It makes much more sense.

I suspect the majority of people reading tech news understands that core counts and clock speeds aren't the best way to assess the performance of a processor. Unfortunately most people don't understand that and will simply see higher core counts and assume that means it's better/faster.

 

[hahahanoobs]

I suspect the majority of people reading tech news understands that core counts and clock speeds aren't the best way to assess the performance of a processor. Unfortunately most people don't understand that and will simply see higher core counts and assume that means it's better/faster.

He thinks every PC is for games.
It's just noise to people that know better.

 

[Skjorn]

More like they can't put more than 8 big cores in that package size without destroying the perf/W. They are forced to use E cores.

That is pretty much exactly what they said.... they need to keep the power in check for desktop users

 

[MasterMace]

I see Intel is going to rename their 10nm node multiple times in their roadmap?

 

[captaincranky]

I see Intel is going to rename their 10nm node multiple times in their roadmap?

I know right? If anybody bought into Intel's "road map" from a decade past, we should be, by of necessity, be measuring process widths in angstroms, and not nanometers by now..

 

[theruck]

I suspect the majority of people reading tech news understands that core counts and clock speeds aren't the best way to assess the performance of a processor. Unfortunately most people don't understand that and will simply see higher core counts and assume that means it's better/faster.

its because its most of the time true https://nanoreview.net/en/cpu-list/cinebench-scores and within a CPU generation it is the one and most simple performance expectation parameter for the consumer. some years ago there were people speaking about how Megahertz do not matter and ppl still counted MHz until core counts replaced them

 

[captaincranky]

please, let’s stop focussing on the core count numbers we are given and look at the actual performance you would get instead. It makes much more sense.

The i3-12100 has a much higher base clock than its supposedly more potent i5 "big brothers". "Assuming", (if can we assume this?), it has the same IPC numbers, it could out perform them on single threaded workloads.

I actually want to do a build with one of these, if and when they become available.

 

[DubiousFishHook]

They tried their hardest as usual.. to make precise, small incremental upgrades so they can milk people. If theres a accidental breakthrough & they research forth a super chip thats 30 trillion times faster than the best super computer it would go in the trash bin.

 

[captaincranky]

They tried their hardest as usual.. to make precise, small incremental upgrades so they can milk people. If theres a accidental breakthrough & they research forth a super chip thats 30 trillion times faster than the best super computer it would go in the trash bin.

But really, isn't that the basic strategy of all businesses? Whether they get you hooked on coffee, (Starbucks), iPhones, (Apple), Video games, (Nvidia), Playing the underdog has worked out very well for AMD and Elon Musk.(who BTW, is almost a trillionaire)

.Believe what you choose. Try this; AMD loves you so much, they lose a dollar on every CPU they sell, and try to make it up with "volume". Do you feel better now?

 

[Gimp65]

They tried their hardest as usual.. to make precise, small incremental upgrades so they can milk people. If theres a accidental breakthrough & they research forth a super chip thats 30 trillion times faster than the best super computer it would go in the trash bin.

totally nonsense, if they had that they would own all markets. Sure they have much crazier stuff in their labs than what we have on the market today, all kinds of company's has that, they even show some of it now and then. One thing is having a thing in your lab, another thing is actually making it profitable and in volume. If you really think intel is holding back, you have no idea.

 

[kiwigraeme]

totally nonsense, if they had that they would own all markets. Sure they have much crazier stuff in their labs than what we have on the market today, all kinds of company's has that, they even show some of it now and then. One thing is having a thing in your lab, another thing is actually making it profitable and in volume. If you really think intel is holding back, you have no idea.

Yeah why hasn't Ford/GM release the engines that run on water - we all know they bought the tech and binned it 80 odd years ago huh? huh? ....../s

 

[Hardware Geek]

Hopefully the graphics cards don't get delayed again and the processor launches on time. I can't wait to see it go head to head with zen 4. Competitive pricing and sufficient supply are a luxury I miss.

 

[Mr Majestyk]

24 cores with 32 threads...

The only configuration I can think of that would match is 8 P cores (16 threads) and 16 E cores (16 threads).

Intel really likes their E cores, I guess.

The core counts for Raptor Lake have already been released officially. It's still only 8P cores for 13900 and 16 E cores. Gracemont cores are only getting minor improvements, Raptor cores are about 10-15% IPC uplift. Zen 4 is the biggest architectural change in the series, IPC uplift alone is 25-30%, 2x L2 cache, possible v-cache, and they are supposedly releasing 4c cores with lower IPC uplifts, less cache but could release 24/32 all 4c core cpu's for those that need huge MT performance. 4c cores will destroy Gracemont and will still be faster than Zen3+ but 10-20% slower than regular Zen 4 cores.

This year's cpu battle will be very interesting. Next year though it's even better, Zen 5 is confirmed as big.little and 4c cores will be the E cores while Meteor Lake soldiers on with Gracemont++.

 

[TheBigFatClown]

They tried their hardest as usual.. to make precise, small incremental upgrades so they can milk people. If theres a accidental breakthrough & they research forth a super chip thats 30 trillion times faster than the best super computer it would go in the trash bin.

No, they would just do something to "cripple" certain parts of it and sell it to different audiences at different price points.

But seriously, pushing the natural laws of the universe past their known boundaries on a never ending basis has got to require some serious brain power (R&D). I don't think they can afford to hide or throw away technological breakthroughs that they might stumble upon. Especially, now that Intel is playing second fiddle to AMD.

On a lighter note, I'm not finding myself becoming all that delighted by any of these fancy new code names for upcoming CPUs. The last one I thought was decent was "Coffee Lake" and that's only because I'm a caffeine addict. I was going to buy one of those for that very reason and along came AMDs Ryzen series. So, this shows myself that code names alone aren't enough for me to blow a wad of hard earned cash on a new CPU but they do get one's attention. Mine anyway.

I think the "Lakes" have been done to death over the past few years. Now, we're moving into the rapids? Hehehe. Dirk the Daring, anyone?

 

[Nintenboy01]

Yeah, if the IPC gain is only 10-15% over Alder Lake then I'm pretty sure Zen 4 will surpass Raptor Lake, or at least trade blows with it.

 

[SixTymes]

The best ceo they have had in years. poor guy has so much to correct and do because of the ineptness just before him.

 

[SixTymes]

No, they would just do something to "cripple" certain parts of it and sell it to different audiences at different price points.

But seriously, pushing the natural laws of the universe past their known boundaries on a never ending basis has got to require some serious brain power (R&D). I don't think they can afford to hide or throw away technological breakthroughs that they might stumble upon. Especially, now that Intel is playing second fiddle to AMD.

On a lighter note, I'm not finding myself becoming all that delighted by any of these fancy new code names for upcoming CPUs. The last one I thought was decent was "Coffee Lake" and that's only because I'm a caffeine addict. I was going to buy one of those for that very reason and along came AMDs Ryzen series. So, this shows myself that code names alone aren't enough for me to blow a wad of hard earned cash on a new CPU but they do get one's attention. Mine anyway.

I think the "Lakes" have been done to death over the past few years. Now, we're moving into the rapids? Hehehe. Dirk the Daring, anyone?

its a word, you are complaining about a word... lmao.

 

[Raptor Lake]

The i3-12100 has a much higher base clock than its supposedly more potent i5 "big brothers". "Assuming", (if can we assume this?), it has the same IPC numbers, it could out perform them on single threaded workloads.

I actually want to do a build with one of these, if and when they become available.

Nooooo! This is a another reason why MHz shouldn't be compared. They cpu won't ever be running at base frequency, it will mostly run at the turbo boost frequency. Check any benchmarks and reviews to see for yourself.

 

[captaincranky]

Nooooo! This is a another reason why MHz shouldn't be compared. They cpu won't ever be running at base frequency, it will mostly run at the turbo boost frequency. Check any benchmarks and reviews to see for yourself.

I beg to differ. It would probably be just tickin' over while I'm starting arguments at Techspot.

And even if it were to be blasting away, I'd have installed a cooler twice the size of what it called for. "Ya say it's 65 watts eh"? "Then this 140 mm Noctua ought to do the trick".

 

[Bamda]

Sorry but CPUs aren't a big deal, there are plenty of excellent CPUs to purchase. Where Intel needs to concentrate on is delivering decent GPUs. The market is starving for GPUs, not CPUs.