Recent from talks
Nothing was collected or created yet.
Intel Quark
View on Wikipedia
Intel Quark is a line of 32-bit x86 SoCs and microcontrollers by Intel, designed for small size and low power consumption, and targeted at new markets including wearable devices. The line was introduced at Intel Developer Forum in 2013, and discontinued in January 2019.[1]
Quark processors, while slower than Atom processors, are much smaller and consume less power. They lack support for SIMD instruction sets (such as MMX and SSE)[2] and only support embedded operating systems.
Quark powers the (now discontinued) Intel Galileo developer microcontroller board.[3] In 2016 Arduino released the Arduino 101 board that includes an Intel Quark SoC.[4][5] The CPU instruction set is, for most models, the same as a Pentium (P54C/i586) CPU.[6]
History
[edit]The first product in the Quark line is the single-core 32 nm X1000 SoC with a clock rate of up to 400 MHz. The system includes several interfaces, including PCI Express, serial UART, I²C, Fast Ethernet, USB 2.0, SDIO, power management controller, and GPIO. There are 16 kB of on-chip embedded SRAM and an integrated DDR3 memory controller.[7][8]
A second Intel product that includes Quark core, the Intel Edison microcomputer, was presented in January 2014. It has a form factor close to the size of an SD card, and is capable of wireless networking using Wi-Fi or Bluetooth.[9]
In January 2015, Intel announced the sub-miniature Intel Curie module for wearable applications, based on a Quark SE core with 80 kB SRAM and 384 kB flash.[10] At the size of a button, it also features a 6-axis accelerometer, a DSP sensor hub, a Bluetooth LE unit and a battery charge controller.
Intel announced the end-of-life of its Quark products in January 2019, with orders accepted until July 2019 and final shipments set for July 2022.[1][11]
List of processors
[edit]This section contains instructions or advice. (September 2021) |
"Lakemont" (32 nm)
[edit]The name Lakemont has been used in reference to the processor core in multiple Quark-series processors.[12]: 4 [13]: 42
"Clanton"
[edit]Source:[14]
- All models support i586 instruction set, with x87 FPU and NX bit
- Temperature range: -40 °C to +85 °C for X10x1 models, 0 °C to +70 °C, for X10x0 models
- Secure boot supported on X102x models
- DDR3 controller with one 16-bit channel
- Two PCI-Express 2.0 controllers, with 1 lane each.
- USB Controller with two USB 2.0 Host ports and one USB 2.0 device port
- Two 10/100 MBit Ethernet controllers
- Integrated memory card reader supporting SDIO 3.0, eMMC 4.41 and SD 3.0
- Serial I/O supporting SPI, UART (serial port) and I2C
(The L2 cache column shows the size of the L1 cache.)
| Model | sSpec number |
Cores | Clock rate | GPU frequency |
L2 cache |
I/O bus | Memory | Voltage | TDP | Socket | Release date | Part number(s) |
Release price (USD) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Quark X1000 |
|
1 | 400 MHz | — | 16 KB | PCIe | DDR3-800 | 0.95–1.1 V | 2.2 W
|
|
Q4'13 |
|
$9.63 |
| Quark X1001 |
|
1 | 400 MHz | — | 16 KB | PCIe | DDR3-800 | 0.95–1.1 V | 2.2 W
|
|
Q2'14 |
|
$11.77 |
| Quark X1010 |
|
1 | 400 MHz | — | 16 KB | PCIe | DDR3-800 (ECC) | 0.95–1.1 V | 2.2 W
|
|
Q1'14 |
|
$10.16 |
| Quark X1011 |
|
1 | 400 MHz | — | 16 KB | PCIe | DDR3-800 (ECC) | 0.95–1.1 V | 2.2 W
|
|
Q2'14 |
|
$12.31 |
| Quark X1020 |
|
1 | 400 MHz | — | 16 KB | PCIe | DDR3-800 (ECC) | 0.95–1.1 V | 2.2 W
|
|
Q2'14 |
|
$11.45 |
| Quark X1020D |
|
1 | 400 MHz | — | 16 KB | PCIe | DDR3-800 (ECC) | 0.95–1.1 V | 2.2 W
|
|
Q1'14 |
|
$10.70 |
| Quark X1021 |
|
1 | 400 MHz | — | 16 KB | PCIe | DDR3-800 (ECC) | 0.95–1.1 V | 2.2 W
|
|
Q2'14 |
|
$13.39 |
| Quark X1021D |
|
1 | 400 MHz | — | 16 KB | PCIe | DDR3-800 (ECC) | 0.95–1.1 V | 2.2 W
|
|
Q2'14 |
|
$12.85 |
"Silver Butte"
[edit]- Implements only a limited subset of the 32-bit x86 instruction set (e.g. segmentation, BCD/string instructions, AF/PF flags,
XCHGare not supported)[15]
| Model | sSpec number |
Cores | Clock rate | GPU frequency |
L2 cache |
I/O bus | Memory | Voltage | TDP | Socket | Release date | Part number(s) |
Release price (USD) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Quark D1000 |
|
1 | 32 MHz | — | AHB-Lite, APB[16]: 30 | eSRAM | 1.62–3.63 V |
|
|
Q3'15 | DMNIAD01SLVBT
|
$2.54 |
"Mint Valley"
[edit]- Supports i586 instruction set, without x87.
| Model | sSpec number |
Cores | Clock rate | GPU frequency |
L2 cache |
I/O bus | Memory | Voltage | TDP | Socket | Release date | Part number(s) |
Release price (USD) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Quark D2000 |
|
1 | 32 MHz | — | AHB-Lite,[13]: 72 APB[13]: 96 | eSRAM | 1.62–3.63 V | 0.025 W
|
|
Q3'15 | FND2000
|
$2.54 |
"Atlas Peak"
[edit]- Supports i586 instruction set, without x87.
| Model | sSpec number |
Cores | Clock rate | GPU frequency |
L2 cache |
I/O bus | Memory | Voltage | TDP | Socket | Release date | Part number(s) |
Release price (USD) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Quark SE C1000 |
|
1 | 32 MHz | — | 8 KB | AHB-Lite, APB | eSRAM | 1.8–3.3 V | 0.025 W
|
|
Q4'15 | LMCQ1000
|
$10.32 |
Segfault bug
[edit]Intel Quark SoC X1000 contains a bug (#71538)[17] that "under specific circumstances" results in a type of crash known as a segfault. The workaround implemented by Intel is to omit LOCK prefixes (not required on single-threaded processors) in the compiled code.[18] While source-based embedded systems like those built using the Yocto Project can incorporate this workaround at compile time, general purpose Linux distributions such as Debian are deeply affected by the bug. Such a workaround is not easy to implement in binaries meant to support multithreading too as they require LOCK prefixes to function properly.[19]
See also
[edit]References
[edit]- ^ a b "Product Change Notification 116715-00" (PDF). Intel Quality Document Management System. Intel. 2019-01-18. Archived (PDF) from the original on 2021-04-18. Retrieved 2020-09-29.
- ^ Turley, Jim (October 16, 2013). "Intel Quark Provides Spin, Charm, and Strange New Low-end x86 MCU Line Emerging from the Lab". EEJournal. Archived from the original on January 8, 2014.
- ^ "Intel® Galileo Datasheet". Archived from the original on 2013-10-12. Retrieved 2013-10-07.
- ^ "Arduino 101". Archived from the original on 2020-09-29. Retrieved 2018-03-23.
- ^ JavaFX 9 by Example, Chapter on Arduino
- ^ "Intel Quark SoC X1000 Core - Developer's Manual". Archived from the original on 2014-10-19. Retrieved 2014-10-19.
- ^ Flaherty, Nick (2013-10-07). "Intel Tackles SoC With Quark". EETimes. Archived from the original on 29 February 2016. Retrieved 9 October 2013.
- ^ Intel® Quark SoC X1000 (16K Cache, 400 MHz) Specifications Archived 2014-01-08 at the Wayback Machine, Intel
- ^ Gareth Halfacree (7 January 2014). "Intel unveils Quark-based Edison microcomputer". BitTech. Archived from the original on 2014-01-10. Retrieved 2014-01-07.
- ^ "Intel® Curie Module: Unleashing Wearable Device Innovation". Intel. 2015-01-06. Archived from the original on 2015-09-06. Retrieved 12 April 2015.
- ^ Shilov, Anton (2019-01-22). "Intel Discontinues Quark SoCs and Microcontrollers". AnandTech. Archived from the original on 2020-11-30. Retrieved 2020-09-29.
- ^ Intel Quark SoC X1000 Debug Operations. Intel Corporation. 2014.
- ^ a b c Intel Quark microcontroller D2000. Intel Corporation. 2015.
- ^ Intel, Quark SoC X1000 Datasheet, May 2014. Archived on Mar 16, 2022.
- ^ Intel, Quark Microcontroller D1000 Programmer's Reference Manual, order no. 332913-02, nov 2015, page 14. Archived on Apr 13, 2021.
- ^ Intel Quark Microcontroller D1000 Datasheet. Intel Corporation. 2015. Archived from the original on 2016-02-23. Retrieved 2016-02-15.
- ^ "Intel Quark SoC X1000 Software - Release Notes" (PDF). Revision 002. 22 May 2014. p. 21. Archived (PDF) from the original on 1 August 2020. Retrieved 17 February 2020.
- ^ "debian-glibc: Bug#738575: pthread: segfault in libpthread on Intel Galileo board". Archived from the original on 2021-02-09. Retrieved 2016-11-11.
- ^ "#738575 - pthread: Segfault in libpthread on Intel Galileo board - Debian Bug report logs". Archived from the original on 2019-04-17. Retrieved 2014-12-04.
External links
[edit]- Intel Quark SoC Documents
- Quark family // Intel ARK (Product Specs)
- Intel's 'Quark' lineup targets wearables. Intel is making the tiniest of chips that can power wearable devices. // CNET, September 10, 2013
- Intel Unveils Tiny Quark Chips for Wearable Devices. Processor Is One-Fifth the Size of Its Low-End Atom Chip // The WSJ, September 10, 2013
- Intel introduces Quark, a tiny chip for the internet of things and wearable computing // The Verge, 2013-09-10
- With Quark, Intel blesses the market of chips for wearable devices // Ventureboat, 2013-09-12
Intel Quark
View on GrokipediaOverview
Introduction
Intel Quark is a family of 32-bit x86-compatible system-on-chips (SoCs) and microcontrollers (MCUs) developed by Intel, introduced in 2013 to address the needs of embedded systems with an emphasis on small size, low power consumption (with a thermal design power of up to 2.2 W), and compatibility with the established x86 software ecosystem.[3][2][10] These processors were engineered to enable x86 instruction set execution in resource-constrained environments, distinguishing them from Intel's higher-performance Core and Atom lines by prioritizing efficiency over raw computational speed.[7][4] A key differentiator of the Quark family lies in its simplified architecture, optimized for non-traditional computing scenarios such as wearables, sensors, and Internet of Things (IoT) devices, where it sought to compete against the prevalent ARM-based solutions dominating the low-power embedded space.[3][7] By featuring the new Lakemont core based on the i486 architecture, designed to achieve one-fifth the die size and one-tenth the power draw of an Atom core, Quark aimed to bring familiar x86 tools and binaries to markets traditionally inaccessible to Intel's architecture.[3][4][11] Announced on September 10, 2013, at the Intel Developer Forum, the Quark lineup saw its first products, including development boards like Galileo, launch in the fourth quarter of that year.[4][2][12] The family encompassed SoC variants suited for application processing tasks and MCU variants focused on real-time control functions, broadening Intel's reach into IoT ecosystems before the entire line reached end-of-life in January 2019, with final shipments concluding in July 2022.[1][9][13]Design Goals
Intel Quark was developed with the primary goal of bringing x86 compatibility to ultra-low-power embedded devices, enabling developers to leverage the mature ecosystem of x86 software tools and binaries in markets traditionally dominated by ARM architectures.[14] This approach targeted clock speeds ranging from 32-400 MHz on a 32nm process node, allowing Quark to compete with ARM-based microcontrollers (MCUs) in terms of size and efficiency while supporting real-time operations in resource-constrained environments.[15] By maintaining backward compatibility with the Pentium-class instruction set architecture (ISA), Quark facilitated easier porting of existing codebases, reducing development time for applications in connected devices.[16] To achieve these objectives, Intel made deliberate trade-offs in the design, including a reduced ISA subset without 64-bit support and strictly in-order execution, which minimized die size to approximately one-fifth that of the Atom core and reduced power consumption to about one-tenth.[15] The Lakemont core served as the foundational element, emphasizing simplicity over advanced features like out-of-order execution or floating-point units to prioritize energy efficiency.[16] Additionally, the architecture incorporated modularity through configurable soft intellectual property (IP) blocks, allowing customization for specific IoT use cases such as sensor integration or peripheral scaling, thereby addressing the diverse needs of embedded systems without over-provisioning resources.[17] Quark's market positioning centered on the burgeoning Internet of Things (IoT), wearables, and industrial sensors, as part of Intel's strategy to extend the x86 ecosystem beyond traditional PCs and challenge ARM's entrenched lead in low-power computing.[17] Power targets were set aggressively low, with system-on-chip (SoC) variants aiming for 1-2W thermal design power (TDP) and MCU implementations reaching as little as 25mW active power, enabling battery life spanning months or years in duty-cycled scenarios.[16] Security was a key emphasis, featuring isolated execution environments and integrated protections against tampering, essential for safeguarding connected devices in safety-critical applications.[17]History
Announcement and Development
Intel's Quark processor family emerged from the company's strategic push into low-power embedded computing and the burgeoning Internet of Things (IoT) sector during the early 2010s, building on prior efforts in micro-server and connected device architectures. The project focused on creating a minimalistic x86-compatible core to compete in markets dominated by ARM-based solutions, with internal development emphasizing extreme power efficiency for battery-operated and always-on applications. The Lakemont core, codenamed for the initial designs, represented a stripped-down in-order x86 implementation derived from simplified elements of Intel's Atom lineage, targeted initially at 32 nm fabrication to achieve sub-watt power envelopes suitable for sensors and gateways.[18][16] The family was publicly unveiled on September 10, 2013, at the Intel Developer Forum (IDF) in San Francisco, where CEO Brian Krzanich positioned Quark as "the next big thing after Atom" for enabling pervasive computing in connected devices. Krzanich highlighted the Quark X1000 SoC as Intel's smallest processor to date, measuring one-fifth the die size of contemporary Atom cores while consuming one-tenth the power, with full synthesizability allowing partners to integrate custom IP blocks. The announcement underscored Intel's ambition to extend x86 ecosystem compatibility— including software tools and peripherals— to IoT edge nodes, contrasting with the power-hungry designs of traditional servers. Led by Intel's core architecture team under the broader "new devices" group headed by executive vice president Mike Bell, the initiative aimed to foster developer adoption through open interfaces and reference designs.[3][19][20] Early collaborations were established to accelerate ecosystem integration, notably with Arduino for maker-friendly hardware compatibility. Just weeks after the IDF reveal, on October 3, 2013, at Maker Faire Rome, Krzanich and Arduino co-founder Massimo Banzi announced the partnership, culminating in the Intel Galileo development board—the first Quark-based product—designed to run Arduino software stacks on x86 hardware. This alliance targeted educators and hobbyists, with Intel committing to donate 50,000 Galileo boards to 1,000 universities worldwide over 18 months to promote IoT prototyping.[21] Pre-launch efforts involved rigorous internal testing of power management and integration, with prototypes demonstrating viability in real-world scenarios such as sensor hubs and wireless gateways. At the IDF keynote, Krzanich showcased early reference designs, including a digital bracelet prototype integrating Quark for wearable health monitoring, alongside ingestible medical sensors and identification chips. Sampling of Quark reference boards to developers began in Q4 2013, enabling pre-production validation focused on ultra-low-power operation in disconnected environments.[3][22]Production and Adoption
The first Intel Quark system-on-chip (SoC), codenamed Clanton and featuring the Quark X1000, entered volume production in the fourth quarter of 2013 following its announcement earlier that year. Fabricated on Intel's 32 nm process technology at company-owned facilities, these processors were optimized for low-power embedded applications, with production ramping up to meet demand during the 2014-2015 surge in Internet of Things (IoT) development.[23] [24] Key product launches expanded the Quark lineup for diverse IoT roles. The Quark X1000 debuted in 2013, targeting gateways and connectivity hubs with its integrated Pentium-class core.[3] Microcontroller variants followed in 2015, including the D1000 and D2000, which offered 32 MHz operation and enhanced I/O for ultra-low-power sensor and control tasks.[25] These releases supported broader ecosystem integration, such as in maker development boards like the Galileo (launched October 2013) and Edison (announced January 2014 and available September 2014).[26] [27] Market reception highlighted Quark's role in bridging x86 compatibility with embedded constraints, though it competed against ARM's mature, cost-effective ecosystem. Adoption grew in industrial IoT through partnerships, such as Siemens' integration of the Quark X1000 into its SIMATIC IOT2000 gateways for secure data aggregation in manufacturing environments.[28] Developers praised the x86 instruction set for simplifying debugging and leveraging existing tools, unlike ARM's fragmented software landscape.[5] Quark contributed to Intel's embedded segment expansion, with the IoT Group reporting $2.1 billion in revenue for 2014 (up 19% from 2013) and $2.3 billion in 2015 (up 7% year-over-year), driven by demand for connected devices.[29] Despite this, adoption remained confined to niches like industrial controls and wearables, where Quark's higher manufacturing costs relative to ARM alternatives limited broader penetration.[30]Discontinuation
Intel issued an official end-of-life (EOL) notice for most Intel Quark variants on January 18, 2019, through Product Change Notification (PCN) 116715-00, covering SoCs such as the X1000, X1010, X1020, and X1021 series, as well as microcontrollers like the D1000, D2000, and SE C1000.[8] Final orders for these products were accepted until July 19, 2019, with committed last shipments scheduled for July 17, 2022, allowing for extended inventory clearance.[8] Earlier discontinuations affected development boards like Galileo and Edison in 2017, with their last order date set for September 16, 2017.[31] The phase-out stemmed from a market shift toward ARM and RISC-V architectures, which offered superior cost and power efficiency for low-end IoT devices, overshadowing x86-based solutions like Quark.[9] Despite initial hype around IoT applications, Quark experienced low sales volumes, failing to achieve widespread adoption in embedded and maker ecosystems.[9] Intel cited shifting customer demand to other products as the primary reason, pivoting toward more capable low-power x86 lines such as the Atom E3900 series (Apollo Lake) for edge computing and, later, integrating similar capabilities into Core Ultra processors.[8][32] The discontinuation impacted the maker community significantly, as support for Quark-based boards like Galileo and Edison ended abruptly in 2017, limiting ongoing projects and community-driven development.[33] Software updates and technical support for Quark products ceased following the EOL timeline, prompting users to migrate to alternative platforms.[34] Inventory shipments continued until 2022 to fulfill existing commitments, but production halted entirely post-2019.[8] As successors, Intel transitioned customers to Atom-based embedded solutions, including the E3900 series for IoT and edge applications, which provided enhanced performance while maintaining low power profiles.[32] Broader IoT platforms like Up Squared, powered by Atom E3900 processors, emerged as direct replacements, offering expanded I/O and integration for industrial and maker use cases.[35]Architecture
Lakemont Core
The Lakemont core serves as the foundational CPU microarchitecture for all Intel Quark processors, providing a low-power, 32-bit x86-compatible design optimized for embedded and IoT applications. It employs an in-order, five-stage in-order pipeline based on the Pentium ISA, supporting core instructions from the IA-32 architecture up to Pentium-level compatibility, excluding SIMD extensions such as MMX and SSE. The core lacks hyper-threading and out-of-order execution, focusing instead on deterministic performance for real-time tasks, with clock speeds configurable from 32 MHz in microcontroller variants to up to 400 MHz in SoC implementations.[16][18][11] Lakemont is a simplified 32-bit x86-compatible microarchitecture based on an i486-like in-order design supporting the Pentium ISA, resulting in a simplified structure with a 32-byte instruction queue, four write buffers, and support for real mode, protected mode, virtual 8086 mode, and system management mode. Memory management includes a 4-way set-associative 32-entry TLB and segmented addressing with 32-bit linear addresses, while the floating-point unit handles 16-, 32-, and 64-bit operations with PC-type error reporting. Cache implementations vary by variant: higher-performance models feature a 16 KB unified L1 cache (4-way set-associative, 16-byte lines, MESI protocol, write-back or write-through configurable), whereas ultra-low-power versions omit L1 caching in favor of 8 KB on-die SRAM for zero-wait-state access. Exclusions such as 64-bit integer support in the integer unit, second-level caching, and dynamic bus sizing further reduce complexity and power. JTAG-compliant debug support is integrated, including breakpoint instructions (opcode 0CCH), single-stepping via the TF bit, and up to four hardware breakpoints using debug registers DR0-DR3.[16][18][11] Power efficiency is central to the design, achieved through dynamic voltage scaling (e.g., 1.8 V active to 1.35 V retention) and multiple low-power states, including C0 (active), C1/C2 (clock gating on halt), stop grant (reducing current to 20-50 mA), and deep sleep modes drawing under 10 µW idle or 1.3-3.4 µA. Fabricated on Intel's 32 nm process, the core's compact footprint—approximately 5 mm²—enables overall SoC power consumption below 1 W, with features like cache disabling (via CR0 bits) and instruction-level power gating supporting MCU-like operation in sleep states. These optimizations ensure compatibility with battery-powered devices while maintaining x86 software portability.[16][18][11]SoC and MCU Features
The Intel Quark SoC, exemplified by the X1000 series, integrates a DDR3/DDR3L memory controller supporting up to 2 GB of addressable memory via a 16-bit data bus at speeds of 800 MT/s.[36] It also features USB 2.0 with two host ports (EHCI/OHCI compliant) and one device port, both operating at up to 480 Mbps, alongside an Ethernet MAC supporting dual 10/100 Mbps ports with MII/RMII interfaces and IEEE 802.3x full-duplex capabilities.[36] For expansion, it includes two PCIe 2.0 root ports each with one lane at 2.5 GT/s, enabling connectivity to peripherals like additional storage or network adapters.[36] Additionally, up to 108 configurable GPIO pins provide interfacing for sensors and actuators, with 6 pins remaining active in S3 suspend mode for wake-up functions.[36] In the MCU variants, such as the Quark D1000 and D2000, ultra-low power modes enable efficient operation for embedded tasks, including active modes at approximately 25 mW (D1000 at 32 MHz) and deep sleep states under 5 µA.[37][11] These MCUs incorporate a 12-bit SAR ADC with up to 19 channels and sampling rates of 2.28 MSps (D2000) or 2.4 MSps (D1000), alongside analog comparators but no integrated DAC.[37][11] A real-time clock driven by a 32.768 kHz oscillator supports battery-backed operation with alarm and wake capabilities, allowing timekeeping without an operating system for basic sensor monitoring tasks.[37][11] Security provisions across Quark SoC and MCU lines include a hardware random number generator for cryptographic key generation and secure boot mechanisms in SE variants, which enforce firmware authentication via on-die Boot ROM and OTP memory to prevent unauthorized code execution.[36][38] The SE microcontroller further supports isolated execution environments through flash protection regions and NVM access controls, enabling trusted computing for sensitive IoT applications.[38][39] Connectivity emphasizes modularity for custom embedded boards, with support for external Wi-Fi and Bluetooth modules via standard interfaces.[36] Core peripherals include two UARTs (16550-compliant, up to 2 Mbps with DMA and flow control), two SPI controllers (masters up to 16 MHz, slaves up to 3.2 MHz), and up to two I2C buses (100-400 kbps master/slave modes), facilitating integration with sensors and peripherals in GPIO-limited designs.[37][11][39]Processors
Clanton
The Clanton, codenamed for the Intel Quark SoC X1000 series, served as the flagship variant in the Quark lineup, optimized for application processing tasks in gateways and development boards. It incorporates a single Lakemont x86 core clocked at 400 MHz, 512 KB of embedded ECC-protected SRAM for on-chip memory, a thermal design power of 2 W, and fabrication on a 32 nm process node.[40][41] The base X1000 configuration delivers comprehensive SoC capabilities, including peripherals like USB 2.0 ports, Ethernet MAC, and PCIe interfaces, while the X1010 and X1020 variants incorporate adjustments such as reduced cache sizes or peripheral subsets for cost optimization; the X1010 and X1020 add ECC memory support, with the X1020 also enabling secure boot functionality.[42][2] In practical deployments, the Clanton found primary use in development boards like the Intel Galileo Gen 2, where it enabled embedded software execution with support for Linux-based systems such as the Yocto Project distribution.[43][44] Introduced in 2013 following its announcement at the Intel Developer Forum, the Clanton series became the highest-volume Quark processor due to its versatility in early IoT and maker ecosystems, before reaching end-of-life status in 2019 with final orders accepted that July.[45]Silver Butte
The Intel Quark Silver Butte, also known as the Quark D1000, is a microcontroller unit (MCU) designed for ultra-low-power applications requiring simple control and direct sensor interfacing in space-constrained embedded systems.[46] It features a single-core Lakemont processor operating at a base frequency of 33 MHz, with configurable frequencies down to 1 MHz for power optimization.[37] The device includes 8 KB of zero-wait-state SRAM for data storage, along with 32 KB of code flash and 4 KB of data flash, but lacks a DRAM controller to maintain its minimal footprint and power profile.[37] Fabricated on a 32 nm process node, it achieves a thermal design power (TDP) of 0.025 W in active mode at full frequency, enabling operation from a single 1.6–3.6 V supply suitable for battery-powered scenarios.[2][46] Key peripherals emphasize I/O density for sensor integration, including up to 24 GPIO pins configurable for various functions, a 19-channel 12-bit successive approximation register (SAR) analog-to-digital converter (ADC) sampling at up to 2.4 MSps, and six high-speed comparators for signal conditioning.[37] It supports multiple timers—two general-purpose 32-bit timers, one CPU-integrated 32-bit timer, and a watchdog timer—that can generate pulse-width modulation (PWM) signals for motor control or LED driving, alongside serial interfaces such as two UARTs, SPI (up to 16 Mbps), and I²C (up to 400 kbps).[37] The D1000 is optimized for bare-metal programming to ensure deterministic performance and minimal overhead, though its architecture allows integration with lightweight RTOS like FreeRTOS in resource-constrained environments.[47] Packaged in a compact 6 mm × 6 mm 40-pin QFN, it offers a smaller form factor compared to the Clanton SoC's larger BGA options, making it ideal for devices where size is critical.[37][36] Targeted at low-end embedded applications such as wearable sensors, RFID tags, industrial transducers, and smart metering peripherals, the Silver Butte enables efficient data acquisition and basic processing without the complexity of higher-end SoCs.[48] Its power management features, including halt and standby modes drawing as low as 2 μA, support extended operation in battery-limited setups like remote monitoring nodes.[37] Launched in the third quarter of 2015 with initial shipments in November, the D1000 saw lower production volumes than the more feature-rich Quark SoCs, reflecting its niche focus on ultra-simple control tasks.[46][49] Intel discontinued the D1000 in 2019, with last orders accepted in July of that year and final shipments concluding in July 2022.[50]Mint Valley
The Intel Quark D2000, codenamed Mint Valley, is a 32-bit x86-compatible microcontroller optimized for ultra-low-power applications in battery-operated devices. It features a single-core Lakemont processor running at 32 MHz, based on the Pentium ISA without a floating-point unit, paired with 8 KB of on-chip SRAM and 32 KB of on-die flash memory for instruction storage, along with 8 KB OTP flash and 4 KB OTP data flash. Advanced power management includes multiple low-power states with deep sleep currents below 3.5 µA, enabling average power consumption under 30 mW in active modes and less than 1 mW in sleep, supported by power gating and a hybrid oscillator for efficient clocking.[51][11][52] Key features emphasize sensor integration and peripheral connectivity, including always-on modes for low-power sensor monitoring, one I²C interface (up to 1 Mbps), one SPI master/slave (up to 16 MHz), two UARTs, 25 GPIOs, a 19-channel SAR ADC (up to 2.28 MSPS), and DMA channels for efficient data handling. It supports real-time operating systems like Zephyr RTOS, facilitating development for edge computing tasks without the overhead of full OSes. These capabilities make it suitable as a sensor hub in resource-constrained environments, with brief reference to peripherals like the ADC for analog signal processing.[11][53][54] Targeted at battery-powered IoT nodes, fitness trackers, and remote sensors, the D2000 prioritizes extended battery life in mobile edge scenarios, such as wearable health monitors and environmental sensing devices, distinguishing it from higher-power industrial variants through its focus on sub-milliwatt sleep efficiency. Launched in Q3 2015, it was positioned for the growing mobile IoT market but reached end-of-life status with last orders in July 2019 and final shipments in July 2022.[51][54][8]Atlas Peak
The Intel Quark SE C1000, codenamed Atlas Peak, represents a secure-oriented microcontroller in the Quark family, emphasizing hardware-based protection for edge devices handling sensitive data. Built around the Lakemont x86-compatible core operating at 32 MHz with an 8 KB L1 instruction cache, it includes 80 KB of on-die SRAM and 384 KB of on-die flash memory (split between 192 KB for host and 192 KB for sensor subsystems), alongside support for up to 1 MB of external SRAM. Operating voltages range from 1.62 V to 1.98 V for the core, with flexible I/O up to 3.3 V, the device achieves ultra-low power consumption, including deep sleep modes below 1 μA and active states optimized for always-on sensing.[55][56] Key security capabilities center on an integrated hardware security engine that accelerates cryptographic functions such as AES-128/192/256, SHA-1/SHA-256, RSA-1024/2048, ECC-256/384, and HMAC, complemented by a true random number generator (TRNG) for key generation. Memory isolation is enforced through four isolated memory regions (IMRs) with agent-based access controls, configurable memory protection regions (MPRs), and on-die NVM read/write protections, enabling partitioned execution environments for trusted code. Additional safeguards include secure boot mechanisms, secure firmware updates, JTAG lockout to block unauthorized debugging, and 8 KB of one-time programmable (OTP) memory for immutable storage, all designed to mitigate physical and logical attacks in resource-constrained settings.[39][55] Targeted at niche, security-critical embedded markets, the Quark SE C1000 supports applications like smart tags and readers, medical/biometric controllers, and secure IoT gateways, where it leverages Intel's broader trusted execution framework for protected cryptographic operations and data isolation. Introduced in the fourth quarter of 2015 as part of Intel's push into intelligent edge computing, it filled a specialized role in low-power, x86-based secure processing until its discontinuation, with final orders accepted through July 2019 and shipments ceasing in 2022.[57][39][30]Applications
Development Boards
The Intel Galileo Gen 1, released in 2013, was the first development board based on the Intel Quark SoC, specifically the Clanton X1000 processor running at 400 MHz, and was designed to be hardware and software compatible with the Arduino Uno R3 ecosystem, including support for existing shields.[58][26] It featured 256 MB of DDR3 RAM, an integrated Ethernet port for connectivity, and ran a Linux distribution based on the Yocto Project, enabling more complex computations than traditional AVR-based Arduino boards.[58][43] The Intel Galileo Gen 2, launched in 2014, built on its predecessor with enhancements for expanded prototyping, including a full-size mini-PCIe slot for additional peripherals like Wi-Fi modules and an 8 MB NOR flash for storage, while retaining the 400 MHz Quark X1000 processor, 256 MB DDR3 RAM, and full Arduino compatibility.[59][60] Like the Gen 1, it supported the open-source Arduino IDE for sketch development and the Yocto Linux kernel, allowing hobbyists to prototype projects involving sensor integration and networked applications.[61][43] The Intel Edison, introduced in 2014 as a compact module roughly the size of an SD card (approximately 25 mm x 35.5 mm), integrated a dual-core Intel Atom processor with a Quark microcontroller core for efficient real-time tasks, 1 GB of LPDDR RAM, 4 GB eMMC storage, built-in Wi-Fi and Bluetooth connectivity, and came bundled with a developer kit featuring a breakout board for easier access to GPIO pins and interfaces. The Atom cores operate at 500 MHz and the Quark core at 100 MHz.[62][63] It also supported the Arduino IDE and Yocto Linux, facilitating rapid prototyping of compact IoT devices such as wearable sensors or small robotics.[43] These boards played a pivotal role in bridging x86 architecture with the maker community, introducing Intel's low-power processors to education and hobbyist environments where Arduino compatibility lowered the entry barrier for experimenting with advanced features like Ethernet and wireless modules in projects focused on robotics and environmental monitoring.[64] Tutorials, shield expansions, and community resources remained actively available through platforms like the Intel Makers community until 2019, fostering widespread adoption in prototyping workflows despite the boards' eventual discontinuation in 2017.[64][65]IoT and Embedded Systems
Intel Quark processors have been deployed in industrial IoT gateways, particularly through Siemens' SIMATIC IOT2000 and IOT2040 platforms, which integrate the Intel Quark x1000 and x1020 SoCs to enable secure data aggregation and connectivity in factory automation environments.[66] These gateways support high-level programming via TIA Portal and open-source Linux distributions, facilitating real-time monitoring and control in manufacturing settings where reliability and 24/7 operation are essential.[28] For instance, the IOT2040 model features 1 GB DDR3 RAM, multiple Ethernet ports, and serial interfaces, allowing integration with sensors for predictive maintenance and process optimization in industrial embedded systems.[67] In building automation and smart environments, Quark-based solutions power hubs and sensor networks, as seen in Yanzi Networks' deployments using the Quark SoC X1000 for access points and the Quark SE SoC for motion-sensor boards.[57] Yanzi's IoT platform leverages Quark's low-power x86 architecture to connect wireless sensors for energy management and occupancy detection, with over 20 installations in Sweden by 2015 that extended to cloud-based analytics for commercial properties.[68] This enables scalable embedded systems for office and facility automation, where Quark handles protocol translation and data processing at the edge. For wearables and environmental sensors, the Intel Quark microcontroller D2000 series supports battery-efficient applications such as fitness trackers and monitoring devices, with its 32 MHz i486 core and integrated peripherals minimizing external components for prolonged operation.[14] In smart city initiatives, Quark-powered gateways were used in Dublin's 2014 environmental sensor project to collect data on air quality and traffic, demonstrating deployment in urban IoT nodes for public infrastructure.[69] Similarly, in transportation and industrial sectors, D2000 variants enable condition monitoring in control units, such as vibration sensors on vehicles or machinery, to predict maintenance needs and reduce downtime.[14] By 2017, Quark integrations contributed to broader IoT ecosystems, supporting connected nodes across industrial and consumer applications, emphasizing secure, low-power embedded computing.[70]Issues
Segfault Bug
The segmentation fault bug in Intel Quark-based systems primarily affected multithreaded applications utilizing the pthread library, leading to recurrent crashes on Intel Galileo development boards during 2014 and 2015. This issue arose in early Linux ports, where applications like SSH daemons (sshd) and package managers (e.g., apt-get) would fail during operations involving thread creation or synchronization, often immediately after booting or executing simple commands. For instance, creating a non-root user or installing packages triggered segfaults in libpthread, halting system functionality and complicating development workflows.[71] The root cause stemmed from a silicon errata in the Quark SoC X1000 (Clanton variant), where instructions prefixed with LOCK—such as lock cmpxchg used for atomic operations in pthread implementations—could corrupt CPU state if a page fault occurred during execution. This flaw in the memory management unit (MMU) handling interacted poorly with shared libraries, as dynamic loading and threading primitives frequently invoked these atomic sequences. The Quark's reduced ISA, aligned with 486-level compatibility but lacking certain Pentium-era features, further exacerbated the problem by misinterpreting or mishandling these instructions as faults rather than executing them safely. Intel documented this as Errata 24 in their specification update, noting no hardware fix was available and emphasizing software avoidance of LOCK-prefixed page-fault-prone instructions.[72][71] Affected systems were centered on Clanton-based Intel Galileo Gen 1 and Gen 2 boards, running Yocto Linux with kernel 3.8.7 or Debian Wheezy ports on i386 architecture. The bug impacted distributions relying on dynamic linking, with reproducible failures in basic threading tests like pthread_create calls or multithreaded benchmarks, while single-threaded applications often ran unaffected. Broader ecosystem ports, such as CentOS 5 derivatives, exhibited similar symptoms due to shared glibc dependencies.[71] Resolutions involved software mitigations rather than hardware corrections, with Intel's Yocto releases patching the GNU assembler (gas) to replace LOCK prefixes with NOP instructions in critical code paths, including glibc and libpthread, starting from BSP 0.7.5 updates in 2014. Kernel advancements to version 3.14 and beyond, integrated into later Yocto builds around 2015, incorporated enhanced fault handling and microcode tweaks to reduce exposure, though full compatibility required recompilation flags like -momit-lock-prefix in binutils or -with-cpu=i386 during glibc builds. Workarounds included static linking applications to bypass shared library atomics, or preloading libraries like libx1000 to NOP out problematic instructions at runtime. Intel's errata document #329677 outlines these mitigations, recommending developer awareness for embedded threading code.[73][71][72]Compatibility Limitations
The Intel Quark processors adhere to a Pentium-class instruction set architecture (ISA), supporting only 32-bit x86 operations without extensions such as SSE2 or later SIMD instructions, and lacking 64-bit addressing capabilities.[74][42] This restricted ISA leads to "illegal instruction" errors when attempting to execute modern x86 software compiled for broader feature sets, necessitating recompilation with flags like -m32 and -march=pentium to ensure compatibility.[75] Operating system support for Quark is confined to lightweight embedded environments, including Yocto Project-based Linux distributions and Wind River Linux, which provide tailored kernels and root filesystems for the platform's constraints.[43][76] Real-time operating systems (RTOS) like Zephyr are also compatible, particularly for microcontroller variants such as the D2000 series, enabling low-overhead task management in IoT applications.[77] Full desktop operating systems, including Windows, are unsupported due to the absence of necessary drivers and hardware features, while post-end-of-life (EOL) status since 2019 has resulted in gaps for peripheral drivers, complicating updates for newer hardware integrations.[9] The Quark ecosystem suffers from a smaller developer community compared to dominant ARM-based alternatives, with limited public documentation and reliance on NDAs for detailed specifications, hindering widespread adoption in maker and IoT spaces.[30] Debugging is particularly challenging without the comprehensive tools available for higher-end Intel Atom processors, as the Pentium ISA's simplifications require specialized configurations and slow iteration cycles.[30] In hybrid designs like the Intel Edison module, which pairs a Quark co-processor with an Atom core, power management exhibits quirks such as inefficient idle states and elevated consumption during transitions, exacerbated by the platform's discontinued support and outdated Yocto Linux kernel.[78][79] Hardware limitations further constrain compatibility, with no integrated GPU for acceleration, relying instead on CPU-only processing for graphics or multimedia tasks. PCIe connectivity supports Generation 1 speeds (2.5 GT/s) via two x1 root ports.[80] Integration with third-party wireless chips, such as Wi-Fi or Bluetooth modules, depends on SDIO or USB interfaces but faces challenges from incomplete driver ecosystems and the need for custom firmware adaptations post-EOL.[80][9]References
- https://en.wikichip.org/wiki/intel/quark