Intel Client & Server CPU Roadmap Updates: Meteor Lake In 2023, 20A & 18A Powered Xeons & Core Chips Beyond 2024
In addition to its graphics update, Intel also announced its new client and server CPU roadmap which encompasses its Core and Xeon chips.
Intel Client 'Core' & Server 'Xeon' CPU Roadmap Unveiled: 5nm Manufacturing In 2H 2022, 3nm In 2H 2023, 20A In 1H 2024 & 18A In 2H 2024
The latest Technology roadmaps and milestones shown by Intel give a rundown of their upcoming products from the Client and Server divisions. Intel states that they are on track to reclaim the perf per watt leadership by 2025 and have shown a range of products that will help them achieve this ambitious goal.
Intel remains on track to reclaim transistor performance per watt leadership by 2025. Intel’s advanced test and packaging technologies give it unmatched industry leadership that benefits its products and foundry customers and will play a critical part in the pursuit of Moore’s Law. Continuous innovation is the cornerstone of Moore’s Law, and innovation is very much alive and well at Intel.
- Process – Intel 7 is in production and shipping in volume with the launch of 12th Gen Intel® Core™ processors and additional products coming in 2022. Intel 4, our implementation of extreme ultraviolet (EUV) lithography, will be manufacturing-ready in the second half of 2022. It delivers an approximate 20% increase in transistor performance per watt. Intel 3, with additional features, delivers a further 18% performance per watt and will be manufacturing-ready in the second half of 2023. Ushering in the Angstrom Era with RibbonFET and PowerVia, Intel 20A will deliver up to a 15% performance per watt improvement and will be manufacturing-ready in the first half of 2024. Intel 18A delivers an additional 10% improvement and will be manufacturing-ready in the second half of 2024.
- Packaging – Our advanced packaging leadership gives designers options across thermals, power, high-speed signaling, and interconnect density, to maximize and co-optimize product performance. In 2022, Intel will ship leadership packaging technologies in Sapphire Rapids and Ponte Vecchio and start risk production on Meteor Lake. Foveros Omni and Foveros Direct, our advanced packaging technologies unveiled at Intel Accelerated in July 2021, will be manufacturing-ready in 2023.
- Innovation – As Intel looks forward to technologies such as High-NA EUV, RibbonFET, PowerVia, and Foveros Omni and Direct, its leaders see no end to innovation and therefore no end to Moore’s Law. Intel remains undeterred in achieving its aspiration of delivering approximately 1 trillion transistors in a single device by the end of the decade.
Intel Process Roadmap
|Process Name||Intel 10nm SuperFin||Intel 7||Intel 4||Intel 3||Intel 20A||Intel 18A|
|Production||In High-Volume (Now)||In Volume (Now)||2H 2022||2H 2023||2H 2024||2H 2025|
|Perf/Watt (over 10nm ESF)||N/A||10-15%||20%||18%||>20%?||TBA|
|Transistor Architecture||FinFET||Optimized FinFET||Optimized FinFET||Optimized FinFET||RibbonFET||Optimized RibbonFET|
|Products||Tiger Lake||Alder Lake|
Xe-HPC / Xe-HP?
Intel Client CPU Roadmap
Starting with the Intel Client CPU roadmap, the company has already launched its Alder Lake CPU lineup, and its follow-up, the 13th Gen Raptor Lake lineup is on track for launch later this year. For Raptor Lake, 24 cores and 32 threads have been confirmed (8+16). The blue team is also promising up to double-digit performance gains, enhanced overclocking support, a new AI M.2 support, and socket compatibility with the Alder Lake platform (LGA 1700/1800).
CCG Roadmap Updates – Building upon its leadership products today, Intel’s roadmap for next-generation client products includes:
- Raptor Lake – Shipping in the second half of 2022, Raptor Lake will deliver up to double-digit performance increases compared with Alder Lake and come with enhanced overclocking features. Raptor Lake features up to 24 cores and 32 threads and is built on the Intel 7 process node with performance hybrid architecture. Raptor Lake will be socket-compatible with Alder Lake systems.
- Meteor Lake and Arrow Lake – Meteor Lake will be built on Intel 4. Arrow Lake will be the first Intel product using Intel 20A tiles, as well as tiles made using an external process. These products will offer a huge step forward on XPU improvement, with integrated AI and tiled GPU architecture that offers discrete graphics-class performance. Meteor Lake will ship in 2023, and Arrow Lake will follow in 2024.
- Lunar Lake and Beyond – Fueled by its IDM 2.0 strategy, Intel will be using both internal and external process nodes to deliver leadership products.
The next update to the Client CPU roadmap will come in the form of Meteor Lake which is scheduled for manufacturing on the 'Intel 4' process node in 2H 2022 and expected for launch in 2023. The 'Intel 4' node will deliver 20% performance per watt improvement and will be the first to use EUV technology, offering a significant increase in density over 'Intel 7' (10nm ESF). Intel will also be launching a follow-up to Meteor Lake in 2024 known as Arrow Lake.
Both Meteor and Arrow Lake CPUs will feature several different core IPs composed of Intel 4 (Meteor Lake), Intel 20A (Arrow Lake), and TSMC's N3 (Arc GPU / SOC) process node. What's particularly interesting is that starting Meteor Lake, the new CPUs will feature a Tiled-GPU design featuring Arc graphics architecture. Raja Koduri states that this 'tGPU' will be unlike anything previously done on client CPUs and will be different than discrete and integrated graphics designs.
Moving beyond 2024, Intel plans to offer a major overhaul to its Client platform with Lunar Lake and Nova Lake chips. These will likely be scheduled for 2025 and 2026 releases, respectively. The next-generation CPUs will come within a tiled-CPU architecture, featuring an MCM CPU, SOC, and GPU IPs integrated on the same package. Both Lunar Lake & Nova Lake chips are expected to utilize the improved 18A node which will offer an additional 10% improvement in performance per watt and add more enhancements to the RibbonFET architecture. The manufacturing is scheduled for 1H 2024 for 20A and 2H 2024 for 18A chips. Other core IPs will be relying on external foundry nodes.
As you can see from the roadmap block diagrams, the next-gen Client platforms will be utilizing a fully-integrated MCM design, making use of several chiplets that will be integrated on the same CPU package. All in all, the following is the node-specific to the chiplet count we can expect on future client chips from team blue:
- 4 - 3 Chiplet SOC
- 20A - 4 Chiplet SOC
- 18A - 5 Chiplet SOC
Intel Mainstream Desktop CPU Generations Comparison:
|Intel CPU Family||Processor Process||Processors Cores/Threads (Max)||TDPs||Platform Chipset||Platform||Memory Support||PCIe Support||Launch|
|Sandy Bridge (2nd Gen)||32nm||4/8||35-95W||6-Series||LGA 1155||DDR3||PCIe Gen 2.0||2011|
|Ivy Bridge (3rd Gen)||22nm||4/8||35-77W||7-Series||LGA 1155||DDR3||PCIe Gen 3.0||2012|
|Haswell (4th Gen)||22nm||4/8||35-84W||8-Series||LGA 1150||DDR3||PCIe Gen 3.0||2013-2014|
|Broadwell (5th Gen)||14nm||4/8||65-65W||9-Series||LGA 1150||DDR3||PCIe Gen 3.0||2015|
|Skylake (6th Gen)||14nm||4/8||35-91W||100-Series||LGA 1151||DDR4||PCIe Gen 3.0||2015|
|Kaby Lake (7th Gen)||14nm||4/8||35-91W||200-Series||LGA 1151||DDR4||PCIe Gen 3.0||2017|
|Coffee Lake (8th Gen)||14nm||6/12||35-95W||300-Series||LGA 1151||DDR4||PCIe Gen 3.0||2017|
|Coffee Lake (9th Gen)||14nm||8/16||35-95W||300-Series||LGA 1151||DDR4||PCIe Gen 3.0||2018|
|Comet Lake (10th Gen)||14nm||10/20||35-125W||400-Series||LGA 1200||DDR4||PCIe Gen 3.0||2020|
|Rocket Lake (11th Gen)||14nm||8/16||35-125W||500-Series||LGA 1200||DDR4||PCIe Gen 4.0||2021|
|Alder Lake (12th Gen)||Intel 7||16/24||35-125W||600 Series||LGA 1700||DDR5 / DDR4||PCIe Gen 5.0||2021|
|Raptor Lake (13th Gen)||Intel 7||24/32||35-125W||700-Series||LGA 1700||DDR5 / DDR4||PCIe Gen 5.0||2022|
|Meteor Lake (14th Gen)||Intel 4||TBA||35-125W||800 Series?||TBA||DDR5||PCIe Gen 5.0?||2023|
|Arrow Lake (15th Gen)||Intel 20A||40/48||TBA||900-Series?||TBA||DDR5||PCIe Gen 5.0?||2024|
|Lunar Lake (16th Gen)||Intel 18A||TBA||TBA||1000-Series?||TBA||DDR5||PCIe Gen 5.0?||2025|
|Nova Lake (17th Gen)||Intel 18A||TBA||TBA||2000-Series?||TBA||DDR5?||PCIe Gen 6.0?||2026|
Intel Server CPU Roadmap
For servers, Intel plans to offer its first major architectural uplift in the form of Diamond Rapids which will be utilizing the 'Intel 3' process node. This node is expected to offer an 18% uplift in performance per watt and will utilize higher-performance libraries, optimized drive current, and metal stack. Intel states that it will have est wafers of their lead product (Xeon) running in their labs by the second half of 2022.
For the first time, Intel has also provided a first-look at its next-gen 18A powered Xeon CPU for the data center segment. This chip is shown with at least 17 tile blocks, 6 of which are CPU blocks.
The current Sapphire Rapids-SP Xeon CPUs feature 15 cores per CPU tile so an 18A chip with a similar core count will offer up to 90 cores though, by 2025, Intel would be offering more cores per CPU tile so we can definitely expect over 100 cores to match AMD's high-core count offerings in the server segment. Diamond Rapids on the '3' node is expected to offer up to 144 cores and 288 threads so 18A would definitely be offering a much higher core count. Intel has also confirmed that its Granite Rapid (P-Core) & Sierra Forest (E-Core) Xeon SKUs will be based on the 'Intel 3' process node.
- Sapphire Rapids: For its P-core lineup on Intel 7, Intel highlighted Sapphire Rapids, its most feature-rich Xeon to date that extends its data center leadership position in several key areas, including AI performance and security features. Next month, Intel will start the initial shipment of Sapphire Rapids for revenue, leading the industry’s transition to next-generation memory and interface standards, including DDR5, PCIe Gen5, and the new Compute Express Link high-speed interconnect.
- Emerald Rapids: In 2023, Intel will deliver Emerald Rapids, the next-generation P-core processor on the Intel 7 process node. The processor will be a socket-compatible refresh to Sapphire Rapids, improving performance and extending the memory and security benefits in the existing platform, providing an easily adaptable and valuable upgrade for Intel’s customers
- Sierra Forest: As part of this new strategy, Intel presented for the first time details of
Sierra Forest,[GA1] its new E-core-based Xeon processor for the data center. It will deliver performance-per-watt efficiency and unmatched density that bolsters the total cost of ownership benefit. This new product, slated for 2024, is designed for the needs of Intel’s hyper-scale customers as a purpose-built solution with a uniquely optimized core for cloud workloads.
- Granite Rapids: Reinforcing its confidence in the health of the Intel 3 process node, Intel announced that it will upgrade Granite Rapids from Intel 4 to the Intel 3 process. This next-generation P-core Xeon product comes in 2024 and will solidify Intel’s leadership in the industry.
Intel Xeon SP Families (Preliminary):
|Family Branding||Skylake-SP||Cascade Lake-SP/AP||Cooper Lake-SP||Ice Lake-SP||Sapphire Rapids||Emerald Rapids||Granite Rapids||Diamond Rapids|
|Process Node||14nm+||14nm++||14nm++||10nm+||Intel 7||Intel 7||Intel 3||Intel 3?|
|Platform Name||Intel Purley||Intel Purley||Intel Cedar Island||Intel Whitley||Intel Eagle Stream||Intel Eagle Stream||Intel Mountain Stream|
Intel Birch Stream
|Intel Mountain Stream
Intel Birch Stream
|Core Architecture||Skylake||Cascade Lake||Cascade Lake||Sunny Cove||Golden Cove||Raptor Cove||Redwood Cove?||Lion Cove?|
|IPC Improvement (Vs Prev Gen)||10%||0%||0%||20%||19%||8%?||35%?||39%?|
|MCP (Multi-Chip Package) SKUs||No||Yes||No||No||Yes||Yes||TBD (Possibly Yes)||TBD (Possibly Yes)|
|Socket||LGA 3647||LGA 3647||LGA 4189||LGA 4189||LGA 4677||LGA 4677||TBD||TBD|
|Max Core Count||Up To 28||Up To 28||Up To 28||Up To 40||Up To 56||Up To 64?||Up To 120?||Up To 144?|
|Max Thread Count||Up To 56||Up To 56||Up To 56||Up To 80||Up To 112||Up To 128?||Up To 240?||Up To 288?|
|Max L3 Cache||38.5 MB L3||38.5 MB L3||38.5 MB L3||60 MB L3||105 MB L3||120 MB L3?||240 MB L3?||288 MB L3?|
|Memory Support||DDR4-2666 6-Channel||DDR4-2933 6-Channel||Up To 6-Channel DDR4-3200||Up To 8-Channel DDR4-3200||Up To 8-Channel DDR5-4800||Up To 8-Channel DDR5-5600?||Up To 12-Channel DDR5-6400?||Up To 12-Channel DDR6-7200?|
|PCIe Gen Support||PCIe 3.0 (48 Lanes)||PCIe 3.0 (48 Lanes)||PCIe 3.0 (48 Lanes)||PCIe 4.0 (64 Lanes)||PCIe 5.0 (80 lanes)||PCIe 5.0 (80 Lanes)||PCIe 6.0 (128 Lanes)?||PCIe 6.0 (128 Lanes)?|
|TDP Range (PL1)||140W-205W||165W-205W||150W-250W||105-270W||Up To 350W||Up To 375W?||Up To 400W?||Up To 425W?|
|3D Xpoint Optane DIMM||N/A||Apache Pass||Barlow Pass||Barlow Pass||Crow Pass||Crow Pass?||Donahue Pass?||Donahue Pass?|
|Competition||AMD EPYC Naples 14nm||AMD EPYC Rome 7nm||AMD EPYC Rome 7nm||AMD EPYC Milan 7nm+||AMD EPYC Genoa ~5nm||AMD Next-Gen EPYC (Post Genoa)||AMD Next-Gen EPYC (Post Genoa)||AMD Next-Gen EPYC (Post Genoa)|