This article was accepted into the corpus but its outbound wikilinks were never NER-processed — typical at the deepest BFS hop or when the run's entity cap was reached. No expansion funnel to show.
| ARM big.LITTLE | |
|---|---|
| Name | ARM big.LITTLE |
| Developer | ARM Holdings |
| Introduced | 2011 |
| Architecture | ARM architecture |
| Application | System on a chip (SoC), mobile device, embedded system |
| Predecessor | ARM Cortex-A9 family |
| Successor | ARM DynamIQ |
ARM big.LITTLE
ARM big.LITTLE is a heterogeneous processor architecture approach developed by ARM Holdings that pairs high-performance cores with energy-efficient cores to balance throughput and power consumption in mobile devices and embedded systems. It targets workloads spanning interactive applications, background services, and multimedia processing used in platforms by companies such as Samsung Electronics, Qualcomm Incorporated, and MediaTek Inc.. The design influenced subsequent heterogeneous initiatives from firms including Intel Corporation, Apple Inc., and NVIDIA Corporation.
big.LITTLE combines "big" high-performance cores like ARM Cortex-A15 and ARM Cortex-A57 with "LITTLE" energy-efficient cores like ARM Cortex-A7 and ARM Cortex-A53 within a single System on a chip produced by vendors such as Samsung Electronics, TSMC, and GlobalFoundries. It addresses power/performance trade-offs relevant to platforms from HTC Corporation and Sony Mobile to Xiaomi Inc. and Google LLC's Pixel devices. The concept aligns with parallel trends exemplified by heterogeneous strategies from Intel Corporation's Atom line and Apple's transition paths in Apple silicon development.
The architecture leverages clusters of homogeneous cores arranged into asymmetric clusters under a single cache coherency domain or across clusters using interconnects like ARM CoreLink or proprietary fabrics by Qualcomm Incorporated and MediaTek Inc.. Modes of operation include cluster migration, core migration, and heterogeneous multi-processing (HMP), each analogous to scheduling paradigms explored by Microsoft in Windows and by Google LLC in Android. Interactions involve components like the L2 cache, L1 cache, and interconnects such as ARM AMBA and vendor IP from Imagination Technologies. big.LITTLE designs exploit power management frameworks used by Texas Instruments and firmware models influenced by UEFI and standards from JEDEC.
Commercial implementations appear across SoCs: Samsung Exynos series (featuring Cortex-A7/Cortex-A15 pairings and later Cortex-A53/Cortex-A57), Qualcomm Snapdragon series (various asymmetric clusters), MediaTek Helio lines, and custom designs by Apple Inc. and Huawei's HiSilicon Kirin series. Foundry partners like TSMC and Samsung Foundry manufactured many big.LITTLE chips for devices from LG Electronics and Motorola Mobility. ODM and OEM integrations extend to Nokia feature sets and consumer electronics from Sony Corporation and Lenovo Group Limited.
big.LITTLE aims to optimize energy per task by dispatching latency-sensitive threads to "big" cores and background tasks to "LITTLE" cores, improving battery life in devices from Apple iPhone competitors and enhancing thermal envelopes expected by laptop vendors like Dell Technologies and HP Inc.. Benchmarks from organizations such as SPEC and AnTuTu demonstrated workload-dependent gains; reviewers from CNET, The Verge, and Ars Technica reported real-world improvements in multitasking and media playback. The approach interacts with power management units (PMUs) and DVFS strategies seen in systems by Intel Corporation and AMD.
Operating system support evolved with patches to the Linux kernel and scheduler work by contributors from Google LLC, Samsung Electronics, and Linaro, enabling HMP and energy-aware scheduling similar to techniques used in Microsoft Windows power plans and adaptations in Android's schedulers. Middleware and runtime layers in projects from Canonical (software company) and Red Hat influence thread placement, while compiler toolchains from GNU Project and LLVM Project facilitate code generation that benefits asymmetric cores. Virtualization frameworks from VMware, Inc. and KVM face challenges coordinating heterogeneous resources.
big.LITTLE was announced by ARM Holdings in 2011 following research trends in heterogeneous computing explored by academia at institutions like MIT, Stanford University, and University of California, Berkeley. Industry demonstrations and partner collaborations involved Samsung Electronics and Texas Instruments; early commercial products arrived in smartphones and tablets from HTC Corporation and LG Electronics. The evolution toward ARM DynamIQ formalized finer-grained heterogeneity, paralleling shifts by Intel Corporation in hybrid core roadmaps and by Apple Inc. in custom big.LITTLE-like arrangements.
Criticisms include increased complexity for SoC design and software, challenges for compiler and OS developers at organizations like Google LLC and Canonical (software company), and limitations in scheduler heuristics highlighted by researchers at Carnegie Mellon University and ETH Zurich. Some reviewers from The Verge and Ars Technica noted inconsistent gains for single-threaded workloads and thermal throttling under sustained load in devices from Samsung Electronics and Qualcomm Incorporated. The approach competes with alternative strategies from Intel Corporation and custom architectures by Apple Inc. and NVIDIA Corporation.
Category:ARM architectures