Recent from talks
Nothing was collected or created yet.
Platform Controller Hub
View on Wikipedia
The Platform Controller Hub (PCH) is a family of Intel's single-chip chipsets, first introduced in 2009. It is the successor to the Intel Hub Architecture, which used two chips–a northbridge and southbridge, and first appeared in the Intel 5 Series.
The PCH controls certain data paths and support functions used in conjunction with Intel CPUs. These include clocking (the system clock), Flexible Display Interface (FDI) and Direct Media Interface (DMI), although FDI is used only when the chipset is required to support a processor with integrated graphics. As such, I/O functions are reassigned between this new central hub and the CPU compared to the previous architecture: some northbridge functions, the memory controller and PCIe lanes, were integrated into the CPU while the PCH took over the remaining functions in addition to the traditional roles of the southbridge. AMD has its equivalent for the PCH, known simply as a chipset since the release of the Zen architecture in 2017.[1] AMD no longer uses its equivalent for the PCH, the Fusion controller hub (FCH).
Overview
[edit]
The PCH architecture supersedes Intel's previous Hub Architecture, with its design addressing the eventual problematic performance bottleneck between the processor and the motherboard. Under the Hub Architecture, a motherboard would have a two-piece chipset consisting of a northbridge chip and a southbridge chip. Over time, the speed of CPUs kept increasing but the bandwidth of the front-side bus (FSB) (connection between the CPU and the motherboard) did not, resulting in a performance bottleneck.[2]
As a solution to the bottleneck, several functions belonging to the traditional northbridge and southbridge chipsets were rearranged. The northbridge and its functions are now eliminated completely: The memory controller, PCI Express lanes for expansion cards and other northbridge functions are now incorporated into the CPU die as a system agent (Intel) or packaged in the processor on an I/O die (AMD Zen 2).
The PCH then incorporates a few of the remaining northbridge functions (e.g. clocking) in addition to all of the southbridge's functions, replacing it. The system clock was previously a connection to a dedicated chip but is now incorporated into the PCH. Two different connections exist between the PCH and the CPU: Flexible Display Interface (FDI) and Direct Media Interface (DMI). The FDI is used only when the chipset requires supporting a processor with integrated graphics. The Intel Management Engine was also moved to the PCH starting with the Nehalem processors and 5-Series chipsets. AMD's chipsets instead use several PCIe lanes to connect with the CPU while also providing their own PCIe lanes, which are also provided by the processor itself.[3][4] The chipset also contains the Nonvolatile BIOS memory.
With the northbridge functions integrated to the CPU, much of the bandwidth needed for chipsets is now relieved.
This style began in Nehalem and will remain for the foreseeable future, through Cannon Lake.
Phase-out
[edit]Beginning with ultra-low-power Haswells and continuing with mobile Skylake processors, Intel incorporated the southbridge IO controllers into the CPU package, eliminating the PCH for a system in package (SOP) design with two dies; the larger die being the CPU die, the smaller die being the PCH die.[5] Rather than DMI, these SOPs directly expose PCIe lanes, as well as SATA, USB, and HDA lines from integrated controllers, and SPI/I²C/UART/GPIO lines for sensors. Like PCH-compatible CPUs, they continue to expose DisplayPort, RAM, and SMBus lines. However, a fully integrated voltage regulator will be absent until Cannon Lake.[needs update]
AMD's FCH has been discontinued since the release of the Carrizo series of CPUs as it has been integrated into the same die as the rest of the CPU.[6] However, since the release of the Zen architecture, there's still a component called a chipset which only handles relatively low speed I/O such as USB and SATA ports and connects to the CPU with a PCIe connection. In these systems all PCIe connections are routed directly to the CPU.[7] The UMI interface previously used by AMD for communicating with the FCH is replaced with a PCIe connection. Technically the processor can operate without a chipset; it only continues to be present for interfacing with low speed I/O.[8] AMD server and laptop CPUs adopt a self contained system on chip (SoC) design instead which doesn't require a chipset.[9][10][11]
Ibex Peak
[edit]The Intel 5 Series chipsets were the first to introduce a PCH. This first PCH is codenamed Ibex Peak.
This has the following variations:
- BD3400 (PCH 3400) Server
- BD3420 (PCH 3420) Server
- BD3450 (PCH 3450) Server
- BD82P55 (PCH P55) Desktop Base
- BD82H55 (PCH H55) Desktop Home
- BD82H57 (PCH H57) Desktop Home
- BD82Q57 (PCH Q57) Desktop Office
- BD82PM55 (PCH PM55) Mobile Base
- BD82HM55 (PCH HM55) Mobile Home
- BD82HM57 (PCH HM57) Mobile Home
- BD82QM57 (PCH QM57) Mobile Office
- BD82QS57 (PCH QS57) Mobile SFF
Issues
[edit]- Bogus USB ports will be detected by desktop PCHs equipped with 6 USB ports (3420, H55) on the first EHCI controller. This can happen when AC power is removed after entering ACPI S4. Applying AC power back and resuming from S4 may result in non detected or even non functioning USB device (erratum 12)
- Bogus USB ports will be detected by mobile PCH equipped with 6 USB ports (HM55) on the first EHCI controller. This can happen when AC power and battery are removed after entering ACPI S4. Applying AC power or battery back and resuming from S4 may result in non detected or even non functioning USB device (erratum 13)
- Reading the HPET comparator timer immediately after a write returns the old value (erratum 14)
- SATA 6 Gbit/s devices may not be detected at cold boot or after ACPI S3, S4 resume (erratum 21)
Langwell
[edit]Langwell is the codename of a PCH in the Moorestown MID/smartphone platform.[12][13] for Atom Lincroft microprocessors.
This has the following variations:
- AF82MP20 (PCH MP20)
- AF82MP30 (PCH MP30)
Tiger Point
[edit]
Tiger Point is the codename of a PCH in the Pine Trail netbook platform chipset for Atom Pineview microprocessors.
This has the following variations:
- CG82NM10 (PCH NM10)
Topcliff
[edit]Topcliff is the codename of a PCH in the Queens Bay embedded platform chipset for Atom Tunnel Creek microprocessors.
It connects to the processor via PCIe (vs. DMI as other PCHs do).
This has the following variations:
- CS82TPCF (PCH EG20T)
Cougar Point
[edit]Cougar Point is the codename of a PCH in Intel 6 Series chipsets for mobile, desktop, workstation, and server platforms. It is most closely associated with Sandy Bridge processors.
This has the following variations:
- BD82C202 (PCH C202) Server
- BD82C204 (PCH C204) Server
- BD82C206 (PCH C206) Workstation and Server
- BD82P67 (PCH P67) Desktop Base
- BD82H67 (PCH H67) Desktop Home
- BD82H61 (PCH H61) Desktop Home
- BD82Z68 (PCH Z68) Combined desktop base and home
- BD82B65 (PCH B65) Desktop Office
- BD82Q67 (PCH Q67) Desktop Office
- BD82Q65 (PCH Q65) Desktop Office
- BD82HM65 (PCH HM65) Mobile Home
- BD82HM67 (PCH HM67) Mobile Home
- BD82QM67 (PCH QM67) Mobile Office
- BD82QS67 (PCH QS67) Mobile SFF
- BD82UM67 (PCH UM67) Ultra Mobile
Issues
[edit]This section's factual accuracy may be compromised due to out-of-date information. (December 2012) |
In the first month after Cougar Point's release, January 2011, Intel posted a press release stating a design error had been discovered. Specifically, a transistor in the 3 Gbit/s PLL clocking tree was receiving too high voltage. The projected result was a 5–15% failure rate within three years of 3 Gbit/s SATA ports, commonly used for storage devices such as hard drives and optical drives. The bug was present in revision B2 of the chipsets, and was fixed with B3. Z68 did not have this bug, since the B2 revision for it was never released. 6 Gbit/s ports were not affected. This bug was especially a problem with the H61 chipset, which only had 3 Gbit/s SATA ports. Through OEMs, Intel plans to repair or replace all affected products at a cost of $700 million.[14][15]
Nearly all produced motherboards using Cougar Point chipsets were designed to handle Sandy Bridge, and later Ivy Bridge, processors. ASRock produced one motherboard for LGA 1156 processors, based on P67 chipset, the P67 Transformer. It exclusively supports Lynnfield Core i5/i7 and Xeon processors, using LGA 1156 socket. After revision B2 of Cougar Point chipsets was recalled, ASRock decided not to update the P67 Transformer motherboard, and was discontinued. Some small Chinese manufacturers are producing LGA 1156 motherboards with H61 chipset.
Whitney Point
[edit]Whitney Point is the codename of a PCH in the Oak Trail tablet platform for Atom Lincroft microprocessors.
This has the following variations:
- 82SM35 (PCH SM35)
Panther Point
[edit]
Panther Point is the codename of a PCH in Intel 7 Series chipsets for mobile and desktop. It is most closely associated with Ivy Bridge processors. These chipsets (except PCH HM75) have integrated USB 3.0.[16]
This has the following variations:
- BD82C216 (PCH C216) Workstation/Server
- BD82H77 (PCH H77) Desktop Home
- BD82Z77 (PCH Z77) Combined desktop base and home
- BD82Z75 (PCH Z75) Combined desktop base and home
- BD82B75 (PCH B75) Desktop Office
- BD82Q77 (PCH Q77) Desktop Office
- BD82Q75 (PCH Q75) Desktop Office
- BD82HM77 (PCH HM77) Mobile Home
- BD82HM76 (PCH HM76) Mobile Home
- BD82HM75 (PCH HM75) Mobile Home
- BD82HM70 (PCH HM70) Mobile Home
- BD82QM77 (PCH QM77) Mobile Office
- BD82QS77 (PCH QS77) Mobile Office
- BD82UM77 (PCH UM77) Ultra Mobile
Cave Creek
[edit]Cave Creek is the codename of the PCH most closely associated with Crystal Forest platforms and Gladden[17] or Sandy Bridge-EP/EN[18] processors.
Patsburg
[edit]Patsburg is the codename of a PCH in Intel 7 Series chipsets for server and workstation using the LGA 2011 socket. It was initially launched in 2011 as part of Intel X79 for the desktop enthusiast Sandy Bridge-E processors in Waimea Bay platforms.[19] Patsburg was then used for the Sandy Bridge-EP server platform (the platform was codenamed Romley and the CPUs codenamed Jaketown, and finally branded as Xeon E5-2600 series) launched in early 2012.[20]
Launched in the fall of 2013, the Ivy Bridge-E/EP processors (the latter branded as Xeon E5-2600 v2 series) also work with Patsburg, typically with a BIOS update.[21][22]
Patsburg has the following variations:
Coleto Creek
[edit]Coleto Creek is the codename of the PCH most closely associated with Highland Forest platforms and Ivy Bridge-EP[23] processors.
Lynx Point
[edit]Lynx Point is the codename of a PCH in Intel 8 Series chipsets, most closely associated with Haswell processors with LGA 1150 socket.[24] The Lynx Point chipset connects to the processor primarily over the Direct Media Interface (DMI) interface.[25]
The following variants are available:[26]
- DH82C222 (PCH C222) Workstation/Server
- DH82C224 (PCH C224) Workstation/Server
- DH82C226 (PCH C226) Workstation/Server
- DH82H81 (PCH H81) Desktop Home
- DH82H87 (PCH H87) Desktop Home
- DH82Z87 (PCH Z87) Combined desktop base and home
- DH82B85 (PCH B85) Desktop Office
- DH82Q87 (PCH Q87) Desktop Office
- DH82Q85 (PCH Q85) Desktop Office
- DH82HM87 (PCH HM87) Mobile Home
- DH82HM86 (PCH HM86) Mobile Home
- DH82QM87 (PCH QM87) Mobile Office
In addition the following newer variants are available, additionally known as Wildcat Point, which also support Haswell Refresh processors:[27]
Issues
[edit]A design flaw causes devices connected to the Lynx Point's integrated USB 3.0 controller to be disconnected when the system wakes up from the S3 state (Suspend to RAM), forcing the USB devices to be reconnected although no data is lost.[28][29] This issue is corrected in C2 stepping level of the Lynx Point chipset.[30]
Wellsburg
[edit]Wellsburg is the codename for the C610-series PCH, supporting the Haswell-E (Core i7 Extreme), Haswell-EP (Xeon E5-16xx v3 and Xeon E5-26xx v3), and Broadwell-EP (Xeon E5-26xx v4) processors. Generally similar to Patsburg, Wellsburg consumes only up to 7 W when fully loaded.[31]
Wellsburg has the following variations:
- DH82029 (PCH C612), intended for servers and workstations
- DHX99 (PCH X99), intended for enthusiasts making use of Intel Core i7 59/69XX processors but it is compatible with LGA 2011-3 Xeons.
Sunrise Point
[edit]Sunrise Point is the codename of a PCH in Intel 100 Series chipsets, most closely associated with Skylake processors with LGA 1151 socket.
The following variants are available:[32]
- GL82C236 (PCH C236) Workstation/Server
- GL82H110 (PCH H110) Desktop Home
- GL82H170 (PCH H170) Desktop Home (Note the datasheet linked one that page is incorrect, see via PCH HM170 below)
- GL82Z170 (PCH Z170) Combined desktop base and home
- GL82B150 (PCH B150) Desktop Office
- GL82Q150 (PCH Q150) Desktop Office
- GL82Q170 (PCH Q170) Desktop Office
- GL82HM170 (PCH HM170) Mobile Home
- GL82CM236 (PCH CM236) Mobile Workstation
- GL82QM170 (PCH QM170) Mobile Office
Union Point
[edit]Union Point is the codename of a PCH in Intel 200 Series chipsets, most closely associated with Kaby Lake processors with LGA 1151 socket.
The following variants are available:[33]
Lewisburg
[edit]Lewisburg is the codename for the C620-series PCH, supporting LGA 2066 socketed Skylake-X/Kaby Lake-X processors ("Skylake-W" Xeon).
Lewisburg has the following variations:
- EY82C621 (PCH C621), intended for servers and workstations
- EY82C622 (PCH C622), intended for servers and workstations
- EY82C624 (PCH C624), intended for servers and workstations
- EY82C625 (PCH C625), intended for servers and workstations
- EY82C626 (PCH C626), intended for servers and workstations
- EY82C627 (PCH C627), intended for servers and workstations
- EY82C628 (PCH C628), intended for servers and workstations
Basin Falls
[edit]Basin Falls is the codename for the C400-series PCH, supporting Skylake-X/Kaby Lake-X processors (branded Core i9 Extreme and "Skylake-W" Xeon). Generally similar to Wellsburg, Basin Falls consumes only up to 6 W when fully loaded.
Basin Falls has the following variations:
Cannon Point
[edit]Cannon Point is the codename of a PCH in Intel 300 Series chipsets, most closely associated with Coffee Lake processors with LGA 1151 socket.[34]
The following variants are available:[35]
Comet Lake PCH
[edit]This section is empty. You can help by adding to it. (August 2021) |
Rocket Lake PCH
[edit]This section is empty. You can help by adding to it. (August 2021) |
Alder Lake PCH
[edit]Raptor Lake PCH
[edit]See also
[edit]- List of Intel chipsets
- Intel Management Engine (ME)
- I/O Controller Hub (ICH)
- PCI IDE ISA Xcelerator (PIIX)
- System Controller Hub (SCH)
- Embedded controller (EC)
References
[edit]- ^ Burke, Steve (2018-06-01). "What is a Chipset? AMD vs. Intel (Z390 vs. Z490, etc.)". GamersNexus.net. Archived from the original on August 29, 2019. Retrieved 2020-08-09.
- ^ Hook, Brian (17 September 2003). "Breaking the Speed Barrier: The Frontside Bus Bottleneck". Technewsworld. Retrieved 1 February 2016.
- ^ Bonshor, Gavin. "The AMD TRX40 Motherboard Overview: 12 New Motherboards Analyzed". www.anandtech.com. Archived from the original on November 29, 2019.
- ^ Burke, Steve (15 June 2019). "AMD X570 vs. X470, X370 Chipset Comparison, Lanes, Specs, & Differences". www.gamersnexus.net.
- ^ Cutress, Ian. "Intel Releases Broadwell-U: New SKUs, up to 48 EUs and Iris 6100". www.anandtech.com. Archived from the original on January 7, 2015.
- ^ "AMD at ISSCC 2015: Carrizo and Excavator Details". Archived from the original on February 24, 2015.
- ^ "AMD Zen 4 Ryzen 9 7950X and Ryzen 5 7600X Review: Retaking the High-End". Archived from the original on September 27, 2022.
- ^ "The AMD Zen and Ryzen 7 Review: A Deep Dive on 1800X, 1700X and 1700". Archived from the original on March 7, 2017.
- ^ "4th Gen AMD EPYC Processor Architecture" (PDF). AMD. Retrieved 2024-11-03.
- ^ Kennedy, Patrick (2019-04-08). "Supermicro M11SDV-4C-LN4F Review mITX AMD EPYC 3151 Platform". ServeTheHome. Retrieved 2024-08-18.
- ^ Cutress, Andrei Frumusanu, Dr Ian. "AMD 3rd Gen EPYC Milan Review: A Peak vs Per Core Performance Balance". www.anandtech.com. Archived from the original on March 21, 2021. Retrieved 2024-08-18.
{{cite web}}: CS1 maint: multiple names: authors list (link) - ^ Langwell Background Information, Intel, archived from the original on 2012-07-09, retrieved 2010-08-03
- ^ New Intel Atom Processor-Based Platform Using Significantly Lower Power Readies Intel for Smartphone, Tablet Push, Intel, May 4, 2010, retrieved 2010-07-27
- ^ Intel Identifies Chipset Design Error, Implementing Solution, Intel, 31 January 2011, archived from the original on 13 May 2012, retrieved 4 February 2011
- ^ The Source of Intel's Cougar Point SATA Bug, AnandTech, 31 January 2011, archived from the original on February 2, 2011
- ^ "Correction: Ivy Bridge and Thunderbolt - Featured, not Integrated". AnandTech. Archived from the original on June 3, 2011. Retrieved 2014-01-21.
- ^ "Intel Xeon/Core/Pentium/Celeron, Communications Chipset 89xx". Intel. 2014-01-13. Retrieved 2014-01-21.
- ^ "Intel Xeon Processor E5-2600/E5-2400/Communications Chipset 89xx". Ssl.intel.com. Retrieved 2014-01-21.
- ^ "A Look Into Intel's Next Gen Enthusiast Platform : Sandy Bridge E & Waimea Bay", vr-zone.com, April 15, 2010, archived from the original on April 23, 2010, retrieved 2010-07-27
- ^ "Intel plugs both your sockets with 'Jaketown' Xeon E5-2600s • The Channel". Channelregister.co.uk. 2012-03-06. Retrieved 2014-01-21.
- ^ "Ivy Bridge-E Not a Cut-down 8-core, 20 MB LLC Die". techPowerUp.com. 2013-08-12. Retrieved 2014-01-21.
- ^ "Intel carves up Xeon E5-2600 v2 chips for two-socket boxes". www.theregister.com.
- ^ "Intel Announces Highland Forest, a New Platform that will Accelerate Network Transformation". Intel. 2013-12-04. Archived from the original on 2016-01-24. Retrieved 2014-08-09.
- ^ Shawn Knight (2012-11-13). "Leaked Intel slides detail Haswell's Lynx Point chipset". techspot.com. Retrieved 2013-10-30.
- ^ "Intel "Lynx Point" 8-series Chipset Detailed, Completely SATA 6 Gbit/s". techpowerup.com. 2012-02-17. Retrieved 2013-10-30.
- ^ "Products (Formerly Lynx Point)". Intel. Retrieved 2013-10-30.
- ^ Andrew Cunningham (2014-05-11). "New Intel chipsets speed up your storage, but they're missing new CPUs". arstechnica.com. Retrieved 2014-05-13.
- ^ "Lynx Point USB 3.0 Controller Issue Correction Needs New Hardware". techpowerup.com. 2013-03-11. Retrieved 2013-10-30.
- ^ Frank Everaardt (2013-03-01). "USB 3.0 problems for Intel's Haswell". hardware.info. Archived from the original on 2013-10-31. Retrieved 2013-10-30.
- ^ "Haswell C2 stepping availability". arstechnica.com. 2013. Retrieved 2013-10-30.
- ^ "Intel Readies 18-Core Xeon "Broadwell-EP" Microprocessors for Launch in 2015 – Report". xbitlabs.com. Archived from the original on 2013-12-24. Retrieved 2014-01-21.
- ^ "Products (Formerly Skylake)". Intel® ARK (Product Specs). Intel. Retrieved 2015-10-24.
- ^ "Products (Formerly Kaby Lake)". Intel. Retrieved 2017-08-18.
- ^ Shilov, Anton (April 26, 2018). "Intel Outs Z390 & X399 PCHs for Cannon Lake & Coffee Lake CPUs". AnandTech. Archived from the original on April 27, 2018. Retrieved 2018-06-29.
- ^ "Products (Formerly Coffee Lake)". Intel® ARK (Product Specs). Intel. Retrieved 2018-06-29.
Platform Controller Hub
View on GrokipediaOverview
Definition and Role
The Platform Controller Hub (PCH) is a centralized chipset component developed by Intel that integrates multiple input/output (I/O), connectivity, and power management functions into a single chip, serving as the primary I/O controller for Intel-based computing platforms.[1] It acts as the successor to the earlier I/O Controller Hub (ICH), which functioned as a southbridge in traditional motherboard architectures, by consolidating these responsibilities to enhance efficiency and reduce hardware complexity.[5] The PCH's core roles include managing a variety of peripheral interfaces and system resources, such as USB ports for device connectivity, SATA for storage devices, PCIe for expansion cards and high-speed links, LPC for legacy peripherals, high-definition audio for multimedia output, Gigabit Ethernet for networking, and comprehensive power management features to optimize energy use across system states.[1] These functions enable seamless communication between the central processing unit (CPU) and external devices, ensuring reliable data transfer and support for modern platform requirements without relying on additional discrete chips.[5] The PCH integrates with the CPU through the Direct Media Interface (DMI), a high-speed serial link that provides dedicated bandwidth for I/O traffic, distinct from the CPU's direct handling of memory control and graphics, which were previously managed by a separate northbridge component.[1] Introduced in 2009 with the 5 Series chipsets (codenamed Ibex Peak), the PCH was designed to simplify motherboard design by minimizing the overall chip count and interconnects, thereby improving scalability, reducing costs, and facilitating more compact system layouts in desktops, laptops, and servers.[1]Historical Context and Introduction
The I/O Controller Hub (ICH), introduced by Intel in the late 1990s as part of the company's Hub Architecture, served as the southbridge component in traditional motherboard designs, managing input/output operations such as USB, SATA, and PCI connections while interfacing with the northbridge via a dedicated hub link.[6] This architecture, which began evolving in the 1990s and persisted through the 2000s, addressed the growing complexity of PC peripherals but faced limitations as processor performance advanced, requiring higher inter-chip bandwidth to handle escalating data demands from integrated graphics, storage, and networking.[7] In 2008, Intel transitioned to the Platform Controller Hub (PCH) as an evolution of the Hub Architecture, debuting it alongside the Nehalem-based Core i7 processors on November 17, with the high-end X58 chipset (codenamed Tylersburg). The X58 platform paired the Nehalem CPU—featuring an integrated memory controller—with an I/O Hub (IOH) that embodied the initial PCH concept, connected via the new Direct Media Interface (DMI) at up to 10 Gb/s to support the processor's QuickPath Interconnect (QPI) and offload peripheral management from the CPU die.[7] This shift was driven by the need for greater bandwidth efficiency in multi-core systems, allowing the CPU to prioritize computational performance over I/O tasks.[7] The full realization of the PCH as a single-chip solution arrived in 2009 with the rollout of the 5 Series chipsets (codenamed Ibex Peak), starting in September, which eliminated the separate northbridge entirely for mainstream platforms and centralized I/O control in the PCH connected directly to the CPU via DMI.[1] By consolidating functions like PCIe lanes, SATA ports, and USB controllers into one hub, the PCH enabled scalable platform designs that have underpinned Intel's x86 architectures ever since.[7]General Architecture
The Platform Controller Hub (PCH) serves as a centralized I/O controller in Intel platforms, integrating multiple peripheral interfaces and system management functions into a single chip to offload tasks from the CPU. At its core, the PCH architecture features a block diagram centered around flexible high-speed I/O (HSIO) lanes that can be configured for various protocols, including multiple SATA ports (up to 8 in recent variants) supporting 6 Gb/s speeds with AHCI mode, up to 14 USB 2.0 ports with additional high-speed USB 3.x/3.2 support via xHCI, numerous PCIe lanes (up to 28 Gen 3/4 in modern designs) distributed across root ports, and an integrated Gigabit Ethernet LAN controller compliant with IEEE 802.3 for 10/100/1000 Mbit/s connectivity.[8][5] These elements connect peripherals to the system while minimizing external components, enabling efficient data routing and resource sharing across the platform. The PCH also integrates the Intel Management Engine (ME) for firmware-based security, remote provisioning, and features like Trusted Execution Technology (TXT) and Virtualization Technology for Directed I/O (VT-d).[8] Power management in the PCH is handled through support for Advanced Configuration and Power Interface (ACPI) states, including S0 (full on), S3 (suspend-to-RAM), S4 (hibernate), and S5 (soft off), along with low-power modes like S0ix for rapid resume and Deep Sx for extended idle efficiency. Thermal monitoring is facilitated by the Platform Environment Control Interface (PECI), a single-wire serial bus that allows real-time temperature reporting from the CPU and other components to the PCH, enabling proactive throttling and fan control to maintain system stability.[8][9] These features ensure compliance with power standards while optimizing energy use in varying workloads.[8] The PCH communicates with the CPU via the Direct Media Interface (DMI) protocol, specifically versions 2.0 and 3.0, which provide a point-to-point serial link with up to 8 GT/s bandwidth across x4 or x8 lanes, supporting bidirectional data transfer at rates up to approximately 8 GB/s in aggregate (3.94 GB/s per direction for DMI 3.0 x8). Additionally, the architecture incorporates integrated General Purpose Input/Output (GPIO) pins for flexible system signaling and wake events, a System Management Bus (SMBus) version 2.0 operating at 100 kHz for low-speed device communication compatible with I²C, and Serial Peripheral Interface (SPI) flash interfaces for secure firmware storage and BIOS execution, typically supporting up to 33 MHz clock speeds and multiple flash devices. The PCH evolved from earlier I/O Controller Hub (ICH) designs by consolidating more functions into a root complex-like structure.[8][10][5]Phase-out and Future Trends
The shift toward on-package and system-on-chip (SoC) integration of I/O functions began with Intel's Meteor Lake processors in 2023, particularly in mobile platforms, where the traditional discrete Platform Controller Hub (PCH) was replaced by an integrated I/O tile within the multi-chip module package. This I/O tile serves as an extension of the SoC tile, handling physical layer interfaces such as USB, SATA, and PCIe endpoints without requiring a separate motherboard-mounted PCH, thereby reducing latency, power consumption, and board space in thin-and-light laptops.[11] In contrast, discrete PCH designs remain essential for desktop and server platforms to support expansive I/O configurations and upgradability. For instance, Intel's Arrow Lake processors, launched in 2024, pair with the 800 Series chipset family, which functions as a discrete PCH embedded on the motherboard to manage connectivity for peripherals, storage, and networking across consumer, workstation, and enterprise segments. This setup provides up to 34 PCIe lanes and enhanced USB support, ensuring robust performance in high-end desktop environments.[12] As of November 2025, hybrid models continue to dominate Intel's portfolio, with integrated I/O tiles in mobile SoCs such as Lunar Lake (launched 2024) and the forthcoming Panther Lake (expected late 2025/early 2026), while discrete PCHs persist in desktop variants such as Arrow Lake (2024) and its anticipated refresh (expected 2026).[13] However, Intel's broader roadmap prioritizes chiplet-based architectures that minimize reliance on separate I/O hubs by embedding more functions directly into CPU packages, aiming for greater efficiency and scalability in AI-driven computing.[13] Looking ahead, Nova Lake platforms expected in 2026 represent a potential milestone in this evolution, with advanced chiplet designs likely absorbing traditional PCH functions into dedicated CPU tiles to streamline integration and reduce external dependencies in both desktop and mobile systems, according to Intel's roadmap as of October 2025. This approach aligns with Intel's emphasis on disaggregated computing for future AI PCs and data center applications.[13]Desktop PCH Variants
Ibex Peak
Ibex Peak represents the first implementation of Intel's Platform Controller Hub (PCH) architecture for desktop systems, introduced as part of the 5 Series chipset family. Launched in September 2009 with the P55 variant, followed by H55 and Q57 models in January 2010, it marked a shift from previous northbridge-southbridge designs by consolidating I/O functions into a single PCH chip connected to the CPU via Direct Media Interface (DMI) 2.0.[1][14] Key features of the Ibex Peak PCH include support for six SATA ports operating at 3 Gb/s speeds, enabling reliable storage connectivity for consumer and business applications. It also provides up to 12 USB 2.0 ports for peripheral expansion, integrated Gigabit Ethernet capabilities through an onboard MAC (typically paired with an external PHY like the Intel 82577), and DMI 2.0 interface running at 2.5 GT/s for high-bandwidth communication between the PCH and processor. These elements allowed for a more streamlined motherboard layout compared to prior generations, reducing component count while maintaining essential I/O performance.[1][14][15] The chipset was designed to support first-generation Intel Core processors based on the Nehalem and Westmere microarchitectures, including the Core i7, i5, and i3 series on the LGA 1156 socket. Specific variants like the P55 targeted enthusiast builds with discrete graphics, while H55 suited systems with integrated graphics, and Q57 focused on business stability with features like vPro support. This compatibility enabled the transition to the Lynnfield (Nehalem) and Clarkdale (Westmere) CPUs, powering early 32nm desktop platforms.[1][16] Early deployments of Ibex Peak encountered USB 2.0 controller instability, where ports could stall under bulk or control traffic loads, leading to system freezes or unresponsive devices. This issue stemmed from errata in the chipset's USB implementation and affected various 5 Series configurations. Intel addressed it through BIOS updates released in 2010, which incorporated microcode fixes to restore stability without requiring hardware replacement.[17][1]Cougar Point
The Cougar Point, codenamed for Intel's 6 Series chipsets, represents the second-generation desktop Platform Controller Hub (PCH), released in January 2011 alongside the Sandy Bridge family of processors.[18] It was designed to interface with 2nd-generation Intel Core i3, i5, and i7 processors via the LGA 1155 socket and a Direct Media Interface (DMI) link operating at up to 20 Gb/s full duplex.[18] Key desktop variants include the H67 (emphasizing integrated graphics without overclocking) and Z68 (adding overclocking support, RAID via Intel Rapid Storage Technology, and flexible graphics switching between integrated and discrete GPUs).[18] These chipsets marked a step forward in I/O capabilities, building on the prior Ibex Peak generation by enhancing storage and expansion options while maintaining compatibility with the evolving desktop ecosystem. A hallmark upgrade in Cougar Point was its storage configuration, featuring six SATA ports: two at 6 Gb/s (SATA 600) for high-speed SSDs and four at 3 Gb/s (SATA 300), with support for AHCI mode, hot-plug, and optional RAID on Z68 variants.[18] For connectivity, it provided up to 14 USB 2.0 ports (configurable to 10-14 depending on the SKU) across two or four EHCI controllers, alongside the platform's first widespread adoption of USB 3.0, typically implemented via two ports from companion controllers like ASMedia ASM1042 on motherboards.[18] Expansion included eight PCIe 2.0 lanes (5 GT/s) from the PCH for peripherals, while the platform supported PCIe 2.0 x16 lanes directly from the Sandy Bridge CPU for discrete graphics cards, enabling high-bandwidth GPU performance without PCH bottlenecks.[18] Additional features encompassed integrated Gigabit Ethernet (on select variants), six-channel HD Audio, and power management compliant with ACPI 4.0, prioritizing efficiency in desktop builds. However, Cougar Point faced a significant manufacturing defect shortly after launch, identified on January 31, 2011, affecting the four 3 Gb/s SATA ports (ports 2-5) in early B2-stepping units shipped from January 9 to mid-February 2011.[19] The issue stemmed from a design flaw in the SATA controller's internal connections, leading to gradual signal degradation over time, potentially reducing performance or disabling ports and risking data integrity in affected systems from 2011-2012.[19] Intel resolved it in the B3 stepping by revising the trace layout, resuming shipments from February 14, 2011, and offering free RMAs or replacements for impacted motherboards, with no estimated impact on overall sales beyond a $300 million charge.[19] This incident highlighted early challenges in high-volume chipset production but did not affect the 6 Gb/s ports (0 and 1) or USB/PCIe functionality.[20]Panther Point
Panther Point is the codename for the Platform Controller Hub (PCH) used in Intel's 7 Series chipsets, released in April 2012 to accompany the Ivy Bridge family of 3rd-generation Core processors.[21] These chipsets, including desktop variants such as Z77, H77, B75, Q75, and Q77, connect the CPU to peripherals via the Direct Media Interface (DMI) 2.0, supporting up to 2 GB/s bidirectional bandwidth.[22] While compatible with 2nd-generation Sandy Bridge processors on LGA 1155 sockets, Panther Point was primarily designed for Ivy Bridge, enabling features like PCI Express 3.0 on the CPU side.[22] A major advancement in Panther Point was the integration of native USB 3.0 support across all variants, featuring an xHCI controller that provides up to 4 USB 3.0 ports at 5 Gb/s, alongside 10 USB 2.0 ports for a total of 14 ports.[22] This built on the Cougar Point's SATA capabilities by expanding storage options to 6 ports, with 4 operating at 6 Gb/s and 2 at 3 Gb/s, enhancing data transfer rates for high-performance drives.[22] Additionally, it includes 8 lanes of PCIe 2.0, configurable for graphics or storage expansion, and supports RAID configurations for improved data redundancy.[22] Fabricated on a 65 nm process, Panther Point maintained a low thermal design power (TDP) of approximately 6 W for desktop implementations, contributing to efficient system operation without significant increases in power consumption compared to prior generations.[23] This design addressed earlier limitations in USB implementation by embedding the controller directly, eliminating the need for third-party add-ons and improving reliability for high-speed peripherals.[24]Lynx Point
Lynx Point represents the fourth-generation Platform Controller Hub (PCH) for desktop platforms, serving as the core I/O component in Intel's 8 Series and 9 Series chipsets. Introduced in May 2013 alongside the 8 Series chipsets such as Z87 and H87, it was designed to pair with 4th-generation Intel Core processors (Haswell) using the LGA 1150 socket. An updated variant powered the 9 Series chipsets, including Z97 and H97, launched in May 2014 to accommodate Haswell Refresh and 5th-generation Core processors (Broadwell), extending compatibility while maintaining backward support for Haswell. This evolution addressed the increasing bandwidth demands of Haswell architectures by enhancing I/O flexibility without altering the fundamental PCH topology.[25][26] Key features of Lynx Point emphasize expanded connectivity, including up to 8 PCIe 2.0 lanes configurable in various widths (such as x4, two x2, or four x1 per port group) for peripherals and storage expansion, operating at 5.0 GT/s. It provides 6 SATA 6 Gb/s ports supporting AHCI and RAID modes with hot-plug capabilities, enabling robust storage configurations. The 9 Series iteration introduced native M.2 support via two PCIe 2.0 lanes or a SATA interface, facilitating early adoption of compact SSDs and marking a shift toward modular storage standards. Building on prior generations like Panther Point, Lynx Point retained integrated USB 3.0 controllers (up to 6 ports) while prioritizing PCIe and SATA enhancements to match Haswell's performance profile. These capabilities were delivered through a 22 nm process, balancing power efficiency with expanded I/O for mainstream and enthusiast desktops.[25][26][27] Lynx Point chipsets, particularly Z87 and Z97, enabled overclocking on unlocked "K" series processors, but early implementations exhibited instability during high-frequency operation due to voltage regulation challenges in the power delivery subsystem. Unsupported overclocking modes could lead to system crashes or thermal throttling, often linked to adaptive voltage offsets and load-line calibration inaccuracies. These issues were mitigated through BIOS updates released throughout 2014 by motherboard vendors, incorporating Intel microcode revisions for improved voltage stability and overclocking reliability. Such patches ensured better compatibility with extreme configurations, though overclocking remained chipset-variant specific—enabled on Z87/Z97 but disabled on lower-tier models like H87.[26][28]Sunrise Point
Sunrise Point represents the fifth-generation desktop Platform Controller Hub (PCH) developed by Intel, serving as the I/O companion to the Skylake and Kaby Lake processor architectures. Released in August 2015 alongside the 100 Series chipsets, such as Z170 and H170, it marked a significant evolution in desktop platform connectivity by integrating enhanced storage and peripheral support tailored for mainstream consumer and enthusiast systems.[29][30] Key features of Sunrise Point include native support for USB 3.1, with configurations allowing up to two USB 3.1 Gen 2 ports (10 Gb/s) on higher-end variants like Z170 and H170, alongside up to 10 USB 3.0 (5 Gb/s) ports for broader peripheral expansion. It also introduced Intel Rapid Storage Technology (RST) enhancements enabling NVMe RAID configurations, such as RAID 0, 1, 5, and 10, for PCIe-based SSDs, improving storage performance and redundancy without relying solely on CPU lanes. Building on the PCIe 2.0 limitations of the prior Lynx Point PCH, Sunrise Point added PCIe 3.0 lanes directly to the chipset, facilitating faster direct-attached storage and expansion options at up to 8 GT/s per lane.[29][31][32] Sunrise Point supports sixth-generation (Skylake) and seventh-generation (Kaby Lake) Intel Core i processors via the LGA 1151 socket, enabling DDR4 memory and integrated graphics capabilities while handling platform power management and I/O routing. Notably, it was the first PCH to provide integrated support for Intel Optane memory acceleration through Intel RST, allowing hybrid storage setups that pair Optane modules with HDDs or SSDs for improved system responsiveness, particularly in the 200 Series chipset variants compatible with Kaby Lake. This feature debuted with Optane's consumer launch in 2017, leveraging the PCH's storage controller for seamless caching without dedicated hardware RAID modules.[33][34]Comet Lake PCH
The Comet Lake Platform Controller Hub (PCH), part of Intel's 400 Series chipsets, serves as the refreshed desktop I/O controller for the 10th-generation Core i processors codenamed Comet Lake-S. Released in April 2020 alongside the Comet Lake-S CPUs, it introduces the LGA 1200 socket and supports high-core-count configurations up to 10 cores, enabling enhanced multitasking and productivity workloads on desktop systems.[35] The 400 Series includes variants such as Z490 for enthusiast overclocking, B460 for mainstream builds, H470 for business applications, and H410 for entry-level setups, all built on a mature architecture that extends compatibility with prior-generation features while optimizing for the new CPU family.[36] Key enhancements in the Comet Lake PCH focus on expanded connectivity without a full architectural overhaul from the preceding 300 Series. It provides up to 24 PCIe 3.0 lanes from the PCH, allowing flexible allocation for storage, networking, and expansion cards, in addition to the 16 PCIe 3.0 lanes from the CPU for graphics and high-bandwidth peripherals.[36] USB support includes up to six USB 3.2 Gen 2x1 ports operating at 10 Gb/s for faster data transfers, alongside up to 10 USB 3.2 Gen 1x1 ports at 5 Gb/s and 14 USB 2.0 ports, catering to diverse peripheral needs in desktop environments.[36] Storage options feature up to eight SATA 6.0 Gb/s ports with RAID 0/1/5/10 configurations via Intel Rapid Storage Technology, maintaining backward compatibility with NVMe SSDs as established in earlier PCH designs.[36][37] A notable advancement is the enhanced integration of Wi-Fi 6 (802.11ax) through the Intel CNVi interface, which offloads radio processing to the PCH for improved efficiency and lower power consumption compared to discrete wireless solutions. This native support enables dual-band operation at up to 2.4 Gb/s on 5 GHz bands when paired with compatible CNVi modules like the Intel Wi-Fi 6 AX201, facilitating seamless high-speed wireless connectivity in modern desktops.[37] Overall, the Comet Lake PCH prioritizes incremental expansions in lane count and I/O versatility to support the increased demands of 10th-generation processors, bridging to future transitions like PCIe 4.0 in subsequent series.[38]Rocket Lake PCH
The Rocket Lake PCH, integrated into Intel's 500 Series Chipset Family, serves as the platform controller hub for 11th-generation Core processors, codenamed Rocket Lake-S, and was released in March 2021. This chipset lineup includes variants such as Z590, H570, B560, and H510, designed primarily for desktop systems to enhance connectivity and I/O capabilities while maintaining compatibility with existing LGA 1200 sockets. Backward compatibility with 400 Series motherboards is achieved through BIOS updates, allowing users to upgrade to Rocket Lake CPUs without replacing the motherboard.[39][8] A key advancement in the Rocket Lake PCH is its introduction of PCIe 4.0 compatibility, specifically providing up to x4 lanes for high-speed storage devices like NVMe SSDs, enabling doubled bandwidth over PCIe 3.0 for faster data transfers in demanding applications. The PCH itself delivers up to 24 PCIe 3.0 lanes for general expansion, with PCIe 4.0 support limited to select ports (such as 0, 3, 4, 7, and 9) when paired with 11th-generation processors, while the DMI 3.0 interface connects to the CPU at x8 width for improved inter-component communication. USB connectivity expands to support up to 12 USB 3.2 ports, including Gen 2x2 options at 20 Gbps, building on the USB expansions seen in the prior Comet Lake platform for better peripheral integration.[8][40][8] Fabricated on a 14 nm process node, the Rocket Lake PCH represents Intel's final iteration of this technology for desktop platforms before transitioning to 7 nm in subsequent generations. The platform's integrated memory controller (IMC) improvements enable native support for DDR4-3200 memory speeds, enhancing overall system performance for Rocket Lake CPUs without requiring overclocking. These features collectively position the 500 Series as an evolutionary step, focusing on storage acceleration and I/O efficiency for gaming and productivity workloads.[8]Alder Lake PCH
The Alder Lake Platform Controller Hub (PCH), part of Intel's 600 Series chipset family, was released in November 2021 alongside the 12th Generation Intel Core processors.[41] These chipsets, including variants such as Z690, H670, B660, and H610, are designed for desktop platforms and support the hybrid architecture of Alder Lake CPUs, which integrate Performance-cores (P-cores) for high-performance tasks and Efficient-cores (E-cores) for efficiency.[42] The PCH serves as the central I/O hub, managing connectivity while the CPU handles primary processing and memory controller functions. A key advancement in the Alder Lake PCH is its enhanced Direct Media Interface (DMI) 4.0 x8 link operating at 16 GT/s, providing up to approximately 15.75 GB/s of bidirectional throughput between the CPU and PCH.[43] This upgrade supports higher data transfer rates for peripherals. The PCH offers up to 28 total PCIe lanes, including up to 12 PCIe 4.0 lanes and up to 16 additional PCIe 3.0 lanes, enabling robust expansion for storage, networking, and other devices.[44] While the PCH itself does not directly support PCIe 5.0, the overall platform integrates CPU-provided PCIe 5.0 x16 lanes dedicated to graphics cards, marking a significant step in I/O scalability.[45] Memory support in the Alder Lake platform is facilitated through the CPU's integrated memory controller, which handles DDR5 (up to 4800 MT/s) or DDR4 (up to 3200 MT/s) configurations, with the PCH overseeing related I/O operations such as USB and SATA interfaces.[42] This dual-memory compatibility allows flexibility for users transitioning to next-generation standards, though DDR5 adoption emphasizes higher bandwidth potential for Alder Lake's hybrid core design. The 600 Series PCH also includes up to 8 SATA 6 Gb/s ports and extensive USB support (up to 14 USB 2.0, 10 USB 3.2 Gen 1x1, and 10 USB 3.2 Gen 2x1 ports across SKUs), ensuring comprehensive peripheral connectivity.[44]Raptor Lake PCH
The Raptor Lake Platform Controller Hub (PCH) serves as the refreshed iteration of Intel's 700 Series chipsets, designed specifically for the 13th-generation Core i processors in desktop systems. Released in October 2022, these chipsets include variants such as Z790 for enthusiast overclocking capabilities and B760 for mainstream builds, enabling enhanced connectivity and performance optimizations over prior generations. These chipsets primarily pair with Raptor Lake and subsequent compatible Intel Core processors, with support for 12th and 14th Gen via BIOS updates.[46][47] Key enhancements in the Raptor Lake PCH focus on improved I/O integration, supporting up to 20 Gbps USB 3.2 Gen 2x2 ports alongside expanded USB4 and Thunderbolt 4 interfaces capable of 40 Gb/s bidirectional data transfer for high-speed peripherals. It supports up to PCIe 4.0 x4 lanes from the PCH for NVMe storage, with PCIe 5.0 x4 available from the CPU, facilitating faster SSD performance, while maintaining compatibility with DDR5 memory up to 192 GB total capacity across four DIMM slots after BIOS updates.[48][49][50] These chipsets build on the hybrid core architecture introduced in Alder Lake by increasing performance core counts for better multitasking efficiency.[51]Arrow Lake PCH
The Arrow Lake Platform Controller Hub (PCH) serves as the I/O controller for Intel's 15th-generation Core Ultra Series 2 desktop processors, released on October 24, 2024, as part of the 800 series chipsets.[52] These chipsets, including the enthusiast-oriented Z890, mainstream B860, and entry-level H810 variants, enable enhanced connectivity and efficiency for desktop platforms focused on AI workloads and high-performance computing.[53] The PCH integrates with the Arrow Lake CPUs to manage peripherals, storage, and networking, building on prior generations like Raptor Lake by continuing support for DDR5 memory up to 6400 MT/s.[54] A key advancement in the Arrow Lake PCH is its expanded PCIe 5.0 support, providing up to x4 PCIe 5.0 lanes from the chipset on the Z890 model, alongside up to 24 PCIe 4.0 lanes for flexible configurations such as graphics cards, NVMe SSDs, and add-in cards.[55] This configuration allows for doubled bandwidth compared to PCIe 4.0 equivalents, facilitating faster data transfer rates critical for AI training and content creation tasks. The B860 offers PCIe 5.0 x4 but with fewer total lanes, while the H810 limits to PCIe 4.0 without PCIe 5.0 from the PCH.[56] The 800 series PCH also introduces standardized USB4 Gen 3x2 support at 40 Gb/s speeds across variants, enabling Thunderbolt-compatible connectivity for external displays, storage, and peripherals with reduced latency.[54] Up to 10 USB 3.2 Gen 2x2 ports (20 Gb/s) and 14 USB 2.0 ports are available on the Z890, with scaled-down options on lower-tier chipsets, enhancing platform efficiency for multi-device ecosystems. For AI-specific enhancements, the PCH facilitates integration with the CPU's built-in Neural Processing Unit (NPU), delivering up to 48 TOPS for machine learning inference while maintaining low overall system power draw.[12]| Chipset Variant | PCIe 5.0 Lanes (Chipset) | USB 3.2 Ports (Max) | USB4 Ports | Key Use Case |
|---|---|---|---|---|
| Z890 | x4 | 10 (Gen 2x2) | 2 | Enthusiast overclocking and AI PCs |
| B860 | x4 | 6 (Gen 2x2) | 1 | Mainstream productivity |
| H810 | None | 4 (Gen 2x1) | 0 | Entry-level systems |
Mobile PCH Variants
Langwell
Langwell is the codename for Intel's first Platform Controller Hub specifically tailored for ultra-low-power mobile platforms, introduced as part of the Moorestown architecture in 2010 to target mobile internet devices (MIDs) and smartphones.[57] Announced in 2009 alongside the Lincroft system-on-chip, it represented Intel's push into handheld computing with enhanced power efficiency over previous Atom designs like Pine Trail. The PCH interfaced with a dedicated power management IC codenamed Briertown to optimize energy delivery, enabling seamless integration of PC-like I/O with mobile peripherals.[57] Fabricated on a 65 nm process, Langwell was engineered for netbook and MID form factors with a platform TDP around 2 W, prioritizing idle power reduction through distributed power gating and advanced sleep states.[57] It supported the Atom Z600 series processors embedded in the 45 nm Lincroft SoC, delivering core computing alongside integrated graphics from the PowerVR SGX 535 GPU.[57] This combination allowed for battery life estimates of up to 5 hours of Wi-Fi web browsing or 4-5 hours of 3G talk time in reference designs. Key features focused on mobile-centric I/O, including a NAND flash controller for eMMC storage, an SDIO 2.0 controller for wireless connectivity, a low-power audio codec with multi-stream support, and a MIPI-CSI interface for high-resolution cameras up to 720p video.[58] The USB 2.0 host controller enabled connectivity for peripherals, while CE-ATA support provided embedded storage options at up to 1.5 Gb/s transfer rates. Integrated graphics were handled via HDMI output for 1080p video, with provisions for low-voltage differential signaling (LVDS) in display configurations suitable for compact screens. Langwell paralleled the desktop Ibex Peak PCH in using the Direct Media Interface (DMI) for CPU connectivity but omitted full PCIe lanes in favor of power savings.Tiger Point
The Intel Tiger Point, codenamed for the NM10 Express Chipset, served as a key platform controller hub in Intel's low-power mobile platforms, particularly the Pine Trail architecture introduced in late 2009 and extended to the Cedar Trail platform in 2011.[59][60] This chipset facilitated the transition to more integrated, power-efficient designs for netbooks and entry-level desktops, pairing with Atom processors to enable compact systems with reduced bill of materials and improved battery life. Manufactured on a 45 nm process, the NM10's compact 17 mm × 17 mm package represented an 85% reduction in size compared to the prior Intel 945GC chipset, contributing to thinner profiles suitable for portable devices like netbooks and early tablets.[59] Key I/O capabilities of Tiger Point emphasized efficiency and connectivity for mobile use, including up to two SATA 3 Gb/s ports for storage access, supporting low-power modes inherited from earlier designs like the Langwell PCH.[59] It provided up to eight USB 2.0 (Hi-Speed) ports for peripheral expansion, an integrated 10/100 Ethernet MAC for networking, and PCI Express lanes configurable for graphics or add-in cards, alongside LPC interface for legacy device support.[59] The chipset also integrated an SMBus controller for system management and power delivery features aligned with ACPI standards, operating at a typical power draw of around 2.1 W to extend battery runtime in ultraportable systems.[61] While not natively supporting SD card readers, its PCI Express and general-purpose I/O enabled integration of such features via companion controllers on motherboards.[59] Tiger Point primarily supported Intel Atom processors from the N and D series, including single- and dual-core models like the Atom N450 (1.66 GHz) and D510 (1.66 GHz) in Pine Trail, as well as later Cedar Trail variants such as the dual-core Atom D2700 (2.0 GHz, 10 W TDP) for enhanced media playback and graphics via integrated PowerVR capabilities.[62] This compatibility allowed for systems with up to 4 GB DDR3 memory and onboard graphics processing, targeting sub-$300 netbooks with full HD video decode support. By optimizing for low thermal design power and form factor, Tiger Point helped sustain the netbook market through 2012, bridging to more advanced mobile platforms before being superseded by higher-performance chipsets.[62][60]Whitney Point
The Whitney Point is the codename for the Intel 82SM35 Platform Controller Hub (PCH), integrated into the SM35 Express Chipset and designed specifically for the Oak Trail platform in low-power mobile devices such as tablets. Introduced in early 2011 as part of Intel's Atom-based ecosystem, it pairs with the Lincroft SoC (featuring Atom Z600 series processors) to enable support for full-featured operating systems like Windows 7 on battery-constrained hardware. This PCH represents an early evolution in Intel's mobile chipset strategy, shifting from system-on-package designs to a discrete PCH for enhanced I/O flexibility while prioritizing power efficiency.[63][64] A key differentiator of Whitney Point is its ultra-low thermal design power (TDP) of 0.75 W, achieved through advanced power management techniques including clock gating, dynamic voltage scaling, and USB selective suspend modes. It connects to the Lincroft SoC via a CMOS Direct Media Interface (cDMI) with dual uni-directional lanes at 400 MT/s, alongside a CMOS Digital Video Out (cDVO) interface at 800 MT/s for display handling. The architecture employs AMBA/OCP-based internal buses to consolidate system control logic, minimizing external components and overall power draw in mobile form factors. Additionally, it incorporates a hardware cryptographic engine for secure key management and content protection, along with 256 KB of on-chip SRAM dedicated to boot code execution and low-power standby states. Power delivery is managed via an integrated SPI interface to an external power management IC (PMIC), enabling fine-grained control over voltage rails and sleep transitions critical for extending battery life in tablets.[63] Whitney Point's I/O subsystem is tailored for compact, connectivity-rich devices, supporting essential peripherals without excessive pin count. It provides robust storage and expansion options optimized for the era's tablet needs, including serial and parallel interfaces for sensors, storage, and multimedia. The following table summarizes its primary connectivity features:| Interface Type | Details |
|---|---|
| USB | 4 x USB 2.0 high-speed ports (480 Mbps), with power-saving modes |
| SATA | 1 x SATA 2.6 port (3.0 Gbps) for storage devices |
| SD/SDIO/MMC | 2 x SD/SDIO/MMC controller ports; 1 x dedicated SDIO port |
| Display | HDMI 1.3a via cDVO (up to 1080p@30 Hz); supports DVI |
| Audio | Intel High Definition Audio with multi-streaming capabilities |
| Other Serial | 3 x I²C buses; 2 x SPI interfaces; LPC bus for legacy peripherals |
Lynx Point-M
Lynx Point-M is the codename for Intel's mobile Platform Controller Hub (PCH) in the 8 Series family, released in 2013 to support the Haswell platform in high-performance laptops and mobile workstations.[67] It exemplifies variants like the HM86 chipset, designed for premium ultrabooks and notebooks.[67] This PCH pairs with mobile Haswell processors from Intel's 4th-generation Core i series, enabling efficient power management and integrated graphics acceleration.[68] Key enhancements focus on storage and wireless capabilities, including support for M.2 slots for solid-state drives (SSDs) to enable faster boot times and application loading.[69] Wireless Display (WiDi) integration allows seamless streaming of video content to compatible screens without cables, leveraging Haswell's improved media processing. Connectivity includes up to 14 USB ports (4 USB 3.0 at 5 Gbps + 10 USB 2.0).[67] A standout advancement in Lynx Point-M is native support for SATA RAID levels 0, 1, and 5 through Intel Rapid Storage Technology, allowing configuration of striped arrays for performance or mirrored/parity setups for data protection in BIOS. This extends RAID capabilities to mobile systems. Lynx Point-M also supports Thunderbolt 2.0 (up to 20 Gbps) for high-bandwidth external peripherals via a single cable.Sunrise Point-M
The Sunrise Point-M platform controller hub (PCH) was released in 2015 as part of Intel's 100 Series mobile chipsets, exemplified by the CM236 variant designed for high-end mobile workstations and performance laptops. It serves as the I/O companion to 6th-generation Intel Core i processors (codenamed Skylake), providing essential connectivity and power management for thin-and-light form factors. Although primarily optimized for Skylake, it maintains compatibility with select 5th-generation Core i configurations in hybrid designs, enabling smooth transitions in mobile platforms.[70][71] Key features of Sunrise Point-M include support for USB 3.0 (up to 5 Gbps) with up to 14 ports total, including Type-C interfaces for reversible connectivity. The PCH also integrates Ethernet support up to 1 Gb/s via multiple MAC controllers, facilitating reliable wired networking in enterprise laptops. Memory handling is enhanced through PCH I/O pathways that accommodate DDR4 and LPDDR4 configurations, with up to 2 DIMMs per channel and ECC support in workstation variants like CM236, allowing for scalable performance in memory-intensive applications. Additionally, it offers DMI 3.0 (8 GT/s), up to 12 PCIe 2.0 lanes, 8 SATA 6 Gb/s ports, and 3 display outputs for multi-monitor setups.[72][70][73] Sunrise Point-M builds on prior mobile PCH designs by incorporating advanced GPIO capabilities, which enable integration of device sensors for versatile form factors such as 2-in-1 convertibles, supporting features like orientation detection through customizable pin configurations. This optimization enhances user experience in hybrid devices by allowing seamless mode switching between laptop and tablet orientations. The chipset's low TDP of 3.67 W contributes to extended battery life, making it ideal for mobile productivity. It also retains backward compatibility with M.2 storage slots introduced in earlier generations for SSD expansion.[70][73]Cannon Point
Cannon Point served as the transitional Platform Controller Hub (PCH) for Intel's Cannon Lake processors, debuting in limited mobile configurations within the 300 Series chipsets in 2018. This PCH was designed specifically for ultra-low-power 8th-generation Core processors, such as the Core i3-8121U, a 10 nm dual-core 15 W part aimed at thin-and-light laptops. As Intel's first foray into 10 nm fabrication for both CPU and PCH components, Cannon Point represented an early attempt to shrink the platform for better power efficiency in mobile devices, though its scope was confined to low-volume production due to manufacturing challenges.[74][75] The Cannon Point PCH featured PCIe 3.0 support with 8 lanes for peripheral connectivity, USB 3.1 Gen 2 ports inheriting capabilities from prior designs like Union Point, and the introduction of the integrated CNVi interface to streamline Wi-Fi integration via compatible M.2 modules. These elements prioritized compact I/O for battery-constrained systems, enabling features like faster data transfer and simplified wireless hardware without discrete controllers. However, the 10 nm process debut encountered yield issues, resulting in throttled clock speeds and suboptimal performance compared to established 14 nm alternatives, which limited Cannon Point's market penetration to a handful of devices before Intel shifted focus to more reliable architectures.[76][77][75] This brief lifecycle underscored Cannon Point's role as a proof-of-concept for on-chip integration trends, paving the way for subsequent PCH evolutions while highlighting the risks of aggressive process transitions in mobile platforms. Adoption remained low, with the platform phased out by early 2020 in favor of refined 14 nm mobile solutions.[78]Union Point
Union Point is the codename for Intel's mobile Platform Controller Hub (PCH) in the 200 Series family, released in 2016 to support the Kaby Lake platform. It exemplifies variants like the HM175 chipset for high-end mobile workstations and performance laptops. It serves as the I/O companion to 7th-generation Intel Core i processors (codenamed Kaby Lake).[79] Key features include continued USB 3.0 support with improved Type-C integration, Ethernet up to 1 Gb/s, DDR4/LPDDR3 memory with ECC in select variants, DMI 3.0, up to 12 PCIe 2.0 lanes, 8 SATA 6 Gb/s ports, and multiple display outputs. It introduces better power efficiency for convertible devices with enhanced GPIO for sensors. TDP around 4 W for mobile variants.[80]Later Mobile PCH Variants
Following Cannon Point, Intel's mobile PCH evolved with the 400 Series (Comet Lake, 2020), using a 14 nm refresh of Cannon Point, supporting 10th-generation Comet Lake-H processors with PCIe 3.0, USB 3.2 Gen 1, and improved Thunderbolt 3 integration for gaming laptops.[81] The 500 Series (Tiger Lake, 2020) shifted toward integration, with low-power U-series featuring on-die I/O, while H-series used discrete PCH with PCIe 4.0 support from CPU, USB4/Thunderbolt 4 (40 Gbps), and CNVi 2.0 for Wi-Fi 6.[82] The 600 Series (Alder Lake mobile, 2021) introduced hybrid architecture support, PCIe 5.0 (from CPU), up to 28 PCIe 4.0 lanes total, DDR5/LPDDR5, and Wi-Fi 6E via CNVi.[83] The 700 Series (Raptor Lake mobile, 2022) enhanced with more PCIe 5.0 lanes, Thunderbolt 4, and better power management for 13th-gen processors.[84] By 2023-2025, Core Ultra series (Meteor Lake, Arrow Lake mobile equivalents) integrated PCH on-package in the compute tile for ultra-low power, supporting PCIe 5.0, USB4 v2 (80 Gbps), and AI accelerators, reducing discrete components in thin laptops. As of November 2025, discrete PCH persists in high-end H-series for expandability.[85]Server and Workstation PCH Variants
Topcliff
The Topcliff Platform Controller Hub (PCH), codenamed EG20T, represents an early implementation of Intel's PCH architecture tailored for embedded and low-end server systems. Released in 2009 as part of the Queens Bay platform, it serves as the I/O controller for low-power, reliable computing environments such as industrial automation and compact servers.[86][87] Designed for integration with Intel's Atom E6xx series processors (Tunnel Creek), including models like the E620T and E640T, Topcliff supports Atom-based server configurations equivalent to the N450 series but optimized for embedded reliability and extended temperature ranges. These processors operate at clock speeds up to 1.6 GHz with low thermal design power (TDP) of 2-3 W, enabling fanless operation in space-constrained server setups. The PCH connects to the CPU via a PCI Express interface, facilitating efficient data transfer for server workloads focused on storage and networking rather than high-performance computing.[88][89][90] Key features of Topcliff emphasize robust I/O for server connectivity, including two SATA ports supporting speeds up to 3 Gb/s for reliable data storage in embedded RAID or direct-attached configurations. It integrates a single [Gigabit Ethernet](/page/Gigabit Ethernet) controller (10/100/1000 Mbps) to enable networked server operations, with support for additional Ethernet via PCIe expansion for dual-port setups in multi-interface environments. Overall, Topcliff prioritizes system stability and low power over raw speed, making it suitable for 24/7 embedded server applications on a 90 nm process node.[91][92][90][93]Cave Creek
The Intel Cave Creek platform controller hub refers to the server-oriented variants of the C600 series chipset, launched in the first quarter of 2012 as part of Intel's Crystal Forest platform. This chipset family was developed to deliver robust I/O connectivity for enterprise workloads, emphasizing reliability and scalability in two-socket server environments. It interfaces with the CPU via a Direct Media Interface (DMI) at 5 GT/s, handling peripheral functions to offload the processor.[94][95] Cave Creek supports Intel Xeon processors from the Sandy Bridge-EP family, specifically the E5-2600 series, which utilize the LGA 2011 socket. These systems leverage Intel QuickPath Interconnect (QPI) links operating at up to 8 GT/s to enable dual-socket configurations, facilitating cache-coherent communication between processors for demanding computational tasks. The chipset is fabricated using a 65 nm process technology, balancing power efficiency with the integration of multiple controllers.[96][94][95] A hallmark of Cave Creek is its expanded storage subsystem, optimized for data-intensive applications. Server variants such as the C606 provide up to eight ports configurable as SATA or SAS through dual integrated Storage Controller Units (SCUs), with two ports supporting SATA at 6 Gb/s and the remainder at 3 Gb/s for SAS or SATA operation. This setup enables high-availability storage arrays, including RAID levels 0, 1, 5, and 10 via Intel Rapid Storage Technology Enterprise (RSTe), enhancing data redundancy and performance without dedicated add-in cards. Additionally, the chipset delivers eight lanes of PCI Express 2.0 connectivity, configurable in various widths (x1, x2, x4, or x8) for peripherals and networking, complementing the CPU's native PCIe 3.0 support.[94][97]Patsburg
The Patsburg chipset, part of Intel's C610 series, was released between 2012 and 2013 as a platform controller hub (PCH) tailored for server and workstation environments supporting multi-socket configurations.[94] It introduced advancements in scalability for enterprise systems, particularly enabling dual PCH setups to manage I/O demands in up to four-socket platforms through features like PM_SYNC2 synchronization.[94] This design evolved from earlier server PCHs, such as Cave Creek, by incorporating enhanced RAID support for storage reliability in larger deployments.[94] Key features of Patsburg include support for 10GbE networking via PCIe interfaces, facilitating high-bandwidth connectivity in data center applications.[94] It provides up to 14 SATA/SAS ports, combining eight 3 Gb/s SAS ports, two 6 Gb/s SATA ports, and four 3 Gb/s SATA ports, allowing flexible storage configurations for RAID arrays and direct-attached storage.[94] The chipset connects to the CPU via a DMI 2.0 interface at 20 Gb/s, ensuring efficient data transfer in multi-processor setups.[94] Patsburg supports Intel Xeon processors based on the Ivy Bridge-EP and Haswell-EP architectures, specifically the E5-2600 v2 and v3 families, enabling robust performance for compute-intensive workloads in two- to four-socket servers.[94] A notable innovation is the transition from the Low Pin Count (LPC) interface to Serial Peripheral Interface (SPI) for firmware management, which supports up to two 16 MB SPI flash devices and allows for larger BIOS sizes to accommodate complex server boot processes.[94] This shift improved boot efficiency and scalability for enterprise firmware updates.[94]Wellsburg
The Wellsburg chipset, marketed as part of Intel's C610 series and specifically the C612 variant, was launched in September 2014 to serve server and workstation platforms requiring robust I/O capabilities.[98] It succeeded the Patsburg chipset by integrating enhanced connectivity in a single-die design optimized for enterprise scalability.[99] Designed for Intel Xeon processors based on the Haswell-EP and Broadwell-EP microarchitectures, Wellsburg primarily supports the E5-2600 v3 and v4 families, enabling configurations with up to 18 cores per socket and DDR4 memory at speeds up to 2400 MT/s.[100] The chipset connects to the CPU via a Direct Media Interface 2.0 (DMI 2.0) at 5 GT/s, facilitating efficient data transfer between the processor and peripherals.[98] A core feature is its integration with the QuickPath Interconnect (QPI), operating at up to 9.6 GT/s for low-latency, coherent communication in multi-socket systems.[100] This allows for high-bandwidth scaling, with each CPU providing 40 lanes of PCIe 3.0 directly, while the PCH contributes 8 additional PCIe 2.0 lanes configurable as x1, x2, x4, or x8.[99][101] Wellsburg emphasizes advanced Reliability, Availability, and Serviceability (RAS) through features like Intel Virtualization Technology for Directed I/O (VT-d) for secure I/O virtualization and Intel Node Manager for power and thermal monitoring.[98] These capabilities support mission-critical operations by enabling error detection, recovery, and remote management. It supports up to 2-socket configurations for E5-2600 v3 and v4 processors.[102]Lewisburg
The Lewisburg chipset, part of Intel's C620 series, was released in July 2017 as the platform controller hub for server and workstation systems based on the first-generation Xeon Scalable processors (Skylake-SP). Manufactured on a 14 nm process node, it emphasizes data center efficiency through features like advanced power management states (including D0/D3 and Deep Sx), clock gating, and Energy Efficient Ethernet (EEE) to optimize performance per watt in multi-socket environments.[9] A key adaptation in Lewisburg is its support for the mesh architecture and Intel Ultra Path Interconnect (UPI) in Skylake-SP platforms, providing up to three UPI links per CPU for scalable multi-socket connectivity while maintaining backward compatibility with prior QPI-based designs like Wellsburg.[9] The platform delivers up to 48 PCIe 3.0 lanes from the CPU; the PCH provides up to 20 PCIe 3.0 lanes, configurable as up to 20 root ports, enabling high-bandwidth expansion for storage, networking, and accelerators without introducing PCIe 4.0.[9][103] Integrated Ethernet options include 10 GbE (up to four ports via Intel X722) and 25 GbE support with features like RDMA, Data Center Bridging (DCB), and virtualization offloads for cloud and virtualization workloads.[9] Lewisburg pairs with Skylake-SP processors via a DMI 3.0 x4 interface and supports six-channel DDR4 memory up to 2666 MT/s, facilitating up to 1.5 TB per socket with Registered DIMMs (RDIMMs) or Load-Reduced DIMMs (LRDIMMs) for memory-intensive data center applications.[9] Additional efficiency measures include Latency Tolerance Reporting (LTR), Dynamic Link Throttling, and Single Root I/O Virtualization (SR-IOV) with up to 128 virtual functions, enhancing resource utilization in virtualized environments.[9] The chipset also incorporates the Intel Management Engine for out-of-band management and security features like Trusted Platform Module (TPM) support to bolster data center reliability.[9] Subsequent generations, such as C621/C622 for Ice Lake-SP and C741 for Sapphire Rapids (as of 2023), continue to evolve the server PCH architecture with enhanced PCIe 4.0/5.0 support and improved security.Basin Falls
The C422 chipset is a platform controller hub designed for workstation systems supporting the second-generation Xeon Scalable processors based on the Cascade Lake microarchitecture. Released in 2019 as part of the Cascade Lake-W platform, it builds on prior workstation PCH designs by incorporating enhanced security features tailored for professional workloads.[104][105] A key advancement in the C422 is the inclusion of hardware-based mitigations for vulnerabilities such as Spectre and Meltdown, which were first addressed in production silicon with Cascade Lake processors to improve system resilience without relying solely on software patches. This PCH supports up to 64 PCIe 3.0 lanes directly from the CPU in single-socket configurations, enabling high-bandwidth connectivity for GPUs, storage, and networking in demanding workstation environments. Additionally, it provides native support for Intel Optane persistent memory, allowing systems to combine DRAM with non-volatile memory for larger effective memory capacities while maintaining data persistence across power cycles. The C422 specifically enables Software Guard Extensions (SGX) enclaves, an Intel technology that creates hardware-isolated regions of memory for running sensitive code and data in trusted execution environments, enhancing security for applications like secure data analytics and confidential computing in workstation deployments. These features position the C422 as an incremental evolution from server-oriented PCHs like Lewisburg, retaining the DMI 3.0 interface while adding Cascade Lake-specific security hardware without major changes to the interconnect architecture.References
- https://en.wikichip.org/wiki/intel/chipsets/langwell
- https://en.wikichip.org/wiki/intel/platforms/oak_trail
- https://en.wikichip.org/wiki/intel/cores/broadwell_ex
