Xeon Phi

History

Background

The Larrabee microarchitecture (in development since 2006) introduced very wide (512-bit) SIMD units to an x86 architecture based processor design, extended to a cache-coherent multiprocessor system connected via a ring bus to memory; each core was capable of four-way multithreading. Due to the design being intended for GPU as well as general purpose computing, the Larrabee chips also included specialised hardware for texture sampling.{{citation |archive-url= https://web.archive.org/web/20100528102342/http://www.anandtech.com/show/3738/intel-kills-larrabee-gpu-will-not-bring-a-discrete-graphics-product-to-market |url-status= dead |archive-date= 28 May 2010

Another contemporary Intel research project implementing x86 architecture on a many-multicore processor was the 'Single-chip Cloud Computer' (prototype introduced 2009), a design mimicking a cloud computing computer datacentre on a single chip with multiple independent cores: the prototype design included 48 cores per chip with hardware support for selective frequency and voltage control of cores to maximize energy efficiency, and incorporated a mesh network for inter-chip messaging. The design lacked cache-coherent cores and focused on principles that would allow the design to scale to many more cores.

The Teraflops Research Chip (prototype unveiled 2007) is an experimental 80-core chip with two floating-point units per core, implementing a 96-bit VLIW architecture instead of the x86 architecture. The project investigated intercore communication methods, per-chip power management, and achieved 1.01 TFLOPS at 3.16 GHz consuming 62 W of power.

{{anchor|Knights Ferry}}Knights Ferry

Intel's Many Integrated Core (MIC) prototype board, named Knights Ferry, incorporating a processor codenamed Aubrey Isle was announced 31 May 2010. The product was stated to be a derivative of the Larrabee project and other Intel research including the Single-chip Cloud Computer.

The development product was offered as a PCIe card with 32 in-order cores at up to 1.2 GHz with four threads per core, 2 GB GDDR5 memory, and 8 MB coherent L2 cache (256 KB per core with 32 KB L1 cache), and a power requirement of ≈300 W, built at a 45 nm process. In the Aubrey Isle core a 1,024-bit ring bus (512-bit bi-directional) connects processors to main memory. Single-board performance has exceeded 750 GFLOPS. The prototype boards only support single-precision floating-point instructions.

Initial developers included CERN, Korea Institute of Science and Technology Information (KISTI) and Leibniz Supercomputing Centre. Hardware vendors for prototype boards included IBM, SGI, HP, Dell and others.

Knights Corner

The Knights Corner product line is made at a 22 nm process size, using Intel's Tri-gate technology with more than 50 cores per chip, and is Intel's first many-cores commercial product.

In June 2011, SGI announced a partnership with Intel to use the MIC architecture in its high-performance computing products. In September 2011, it was announced that the Texas Advanced Computing Center (TACC) will use Knights Corner cards in their 10-petaFLOPS "Stampede" supercomputer, providing 8 petaFLOPS of compute power. According to "Stampede: A Comprehensive Petascale Computing Environment" the "second-generation Intel (Knights Landing) MICs will be added when they become available, increasing Stampede's aggregate peak performance to at least 15 PetaFLOPS."

On 15 November 2011, Intel showed an early silicon version of a Knights Corner processor.

On 5 June 2012, Intel released open source software and documentation regarding Knights Corner.

On 18 June 2012, Intel announced at the 2012 Hamburg International Supercomputing Conference that Xeon Phi will be the brand name used for all products based on their Many Integrated Core architecture. In June 2012, Cray announced it would be offering 22 nm 'Knight's Corner' chips (branded as 'Xeon Phi') as a co-processor in its 'Cascade' systems.

In June 2012, ScaleMP announced a virtualization update allowing Xeon Phi as a transparent processor extension, allowing legacy MMX/SSE code to run without code changes. An important component of the Intel Xeon Phi coprocessor's core is its vector processing unit (VPU). The VPU features a novel 512-bit SIMD instruction set, officially known as Intel Initial Many Core Instructions (Intel IMCI). Thus, the VPU can execute 16 single-precision (SP) or 8 double-precision (DP) operations per cycle. The VPU also supports Fused Multiply-Add (FMA) instructions and hence can execute 32 SP or 16 DP floating point operations per cycle. It also provides support for integers. The VPU also features an Extended Math Unit (EMU) that can execute operations such as reciprocal, square root, and logarithm, thereby allowing these operations to be executed in a vector fashion with high bandwidth. The EMU operates by calculating polynomial approximations of these functions.

On 12 November 2012, Intel announced two Xeon Phi coprocessor families using the 22 nm process size: the Xeon Phi 3100 and the Xeon Phi 5110P. The Xeon Phi 3100 will be capable of more than 1 teraFLOPS of double-precision floating-point instructions with 240 GB/s memory bandwidth at 300 W. The Xeon Phi 5110P will be capable of 1.01 teraFLOPS of double-precision floating-point instructions with 320 GB/s memory bandwidth at 225 W. The Xeon Phi 7120P will be capable of 1.2 teraFLOPS of double-precision floating-point instructions with 352 GB/s memory bandwidth at 300 W.

On 17 June 2013, the Tianhe-2 supercomputer was announced by TOP500 as the world's fastest. Tianhe-2 used Intel Ivy Bridge Xeon and Xeon Phi processors to achieve 33.86 petaFLOPS. It was the fastest on the list for two and a half years, lastly in November 2015.{{cite web

Design and programming

The cores of Knights Corner are based on a modified version of P54C design, used in the original Pentium. The basis of the Intel MIC architecture is to leverage x86 legacy by creating an x86-compatible multiprocessor architecture that can use existing parallelization software tools. Programming tools include OpenMP, OpenCL, Cilk/Cilk Plus and specialised versions of Intel's Fortran, C++ and math libraries.

Design elements inherited from the Larrabee project include x86 ISA, 4-way SMT per core, 512-bit SIMD units, 32 KB L1 instruction cache, 32 KB L1 data cache, coherent L2 cache (512 KB per core), and ultra-wide ring bus connecting processors and memory.

The Knights Corner 512-bit SIMD instructions share many intrinsic functions with AVX-512 extension . The instruction set documentation is available from Intel under the extension name of KNC.

Knights Landing

Code name for the second-generation MIC architecture product from Intel. Intel officially first revealed details of its second-generation Intel Xeon Phi products on 17 June 2013. Intel said that the next generation of Intel MIC Architecture-based products will be available in two forms, as a coprocessor or a host processor (CPU), and be manufactured using Intel's 14 nm process technology. Knights Landing products will include integrated on-package memory for significantly higher memory bandwidth.

Knights Landing contains up to 72 Airmont (Atom) cores with four threads per core, using LGA 3647 socket supporting up to 384 GB of "far" DDR4 2133 RAM and 8–16 GB of stacked "near" 3D MCDRAM, a version of the Hybrid Memory Cube. Each core has two 512-bit vector units and supports AVX-512 SIMD instructions, specifically the Intel AVX-512 Foundational Instructions (AVX-512F) with Intel AVX-512 Conflict Detection Instructions (AVX-512CD), Intel AVX-512 Exponential and Reciprocal Instructions (AVX-512ER), and Intel AVX-512 Prefetch Instructions (AVX-512PF). Support for IMCI has been removed in favor of AVX-512.

The National Energy Research Scientific Computing Center announced that Phase 2 of its newest supercomputing system "Cori" would use Knights Landing Xeon Phi coprocessors.

On 20 June 2016, Intel launched the Intel Xeon Phi product family x200 based on the Knights Landing architecture, stressing its applicability to not just traditional simulation workloads, but also to machine learning. The model lineup announced at launch included only Xeon Phi of bootable form-factor, but two versions of it: standard processors and processors with integrated Intel Omni-Path architecture fabric. The latter is denoted by the suffix F in the model number. Integrated fabric is expected to provide better latency at a lower cost than discrete high-performance network cards.

On 14 November 2016, the 48th list of TOP500 contained two systems using Knights Landing in the Top 10.

The PCIe based co-processor variant of Knight's Landing was never offered to the general market and was discontinued by August 2017. This included the 7220A, 7240P and 7220P coprocessor cards.

Intel announced they were discontinuing Knights Landing in summer 2018.

Models

All models can boost to their peak speeds, adding 200 MHz to their base frequency when running just one or two cores. When running from three to the maximum number of cores, the chips can only boost 100 MHz above the base frequency. All chips run high-AVX code at a frequency reduced by 200 MHz.

Knights Mill

Knights Mill is Intel's codename for a Xeon Phi product specialized in deep learning, initially released in December 2017. Nearly identical in specifications to Knights Landing, Knights Mill includes optimizations for better utilization of AVX-512 instructions. Single-precision and variable-precision floating-point performance increased, at the expense of double-precision floating-point performance.

Models

Knights Hill

Knights Hill was the codename for the third-generation MIC architecture, for which Intel announced the first details at SC14. It was to be manufactured in a 10 nm process.

Knights Hill was expected to be used in the United States Department of Energy Aurora supercomputer, to be deployed at Argonne National Laboratory. However, Aurora was delayed in favor of using an "advanced architecture" with a focus on machine learning.

In 2017, Intel announced that Knights Hill had been canceled in favor of another architecture built from the ground up to enable Exascale computing in the future. This new architecture was expected for 2020–2021; however, this was also cancelled due to the discontinuation of the Xeon Phi.

Programming

One performance and programmability study reported that achieving high performance with Xeon Phi still needs help from programmers and that merely relying on compilers with traditional programming models is insufficient. Other studies in various domains, such as life sciences and deep learning, have shown that exploiting the thread- and SIMD-parallelism of Xeon Phi achieves significant speed-ups.

Competitors

• Nvidia Tesla, a direct competitor in the HPC market
• AMD Radeon Pro and AMD Radeon Instinct direct competitors in the HPC market

References

References

1. Ian Cutress & Anton Shilov. (7 May 2019). "The Larrabee Chapter Closes: Intel's Final Xeon Phi Processors Now in EOL".
2. (8 Nov 2012). "Intel® Xeon Phi™ Coprocessor System Software Developers Guide". Intel.
3. robert-reed. (4 February 2013). "Best Known Methods for Using OpenMP on Intel Many Integrated Core (Intel MIC) Architecture".
4. (1 March 2013). "Intel Xeon Phi Coprocessor High Performance Programming". [[Morgan Kaufmann]].
5. Hruska, Joel. (8 May 2019). "Intel Quietly Kills Off Xeon Phi".
6. (6 December 2009). "Intel scraps graphics chip based on Larrabee". [[Reuters]].
7. (2016). "Knights Landing: Second-Generation Intel Xeon Phi Product". [[IEEE Micro]].
8. "TOP500 - June 2013". TOP500.
9. "June 2023 {{!}} TOP500 Supercomputer Sites".
10. "Intel Powers the World's Fastest Supercomputer, Reveals New and Future High Performance Computing Technologies".
11. W1zzard. (24 July 2018). "Intel is Giving up on Xeon Phi - Eight More Models Declared End-Of-Life".
12. Marc Sauter. (20 June 2016). "Knights Landing: Intel veröffentlicht Xeon Phi mit bis zu 7 Teraflops - Golem.de".
13. Charlie Demerjian. (3 July 2006). "New from Intel: It's Mini-Cores!". [[The Inquirer]].
14. (August 2008). "Larrabee: A Many-Core x86 Architecture for Visual Computing". [[ACM Transactions on Graphics]].
15. Tony Bradley. (3 December 2009). "Intel 48-Core "Single-Chip Cloud Computer" Improves Power Efficiency". [[PC World (magazine).
16. "Intel Research : Single-Chip Cloud Computer". [[Intel]].
17. Ben Ames. (11 February 2007). "Intel Tests Chip Design With 80-Core Processor". IDG News.
18. "Intel Details 80-Core Teraflops Research Chip - X-bit labs". xbitlabs.com.
19. "Intel's Teraflops Research Chip". Intel.
20. Anton Shilov. (12 February 2007). "Intel Details 80-Core Teraflops Research Chip". Xbit laboratories.
21. Rupert Goodwins. (1 June 2010). "Intel unveils many-core Knights platform for HPC". [[ZDNet]].
22. (31 May 2010). "Intel News Release : Intel Unveils New Product Plans for High-Performance Computing". Intel.
23. Mike Giles. (24 June 2010). "Runners and riders in GPU steeplechase". people.maths.ox.ac.uk.
24. (December 2010). "Intel Many Integrated Core Architecture". Intel.
25. Rick Merritt. (20 June 2011). "OEMs show systems with Intel MIC chips". [[EE Times]].
26. Tom R. Halfhill. (18 July 2011). "Intel Shows MIC Progress". [[The Linley Group]].
27. Gareth Halfacree. (20 June 2011). "Intel pushes for HPC space with Knights Corner". [[Net Communities Limited, UK]].
28. Andrea Petrou. (20 June 2011). "SGI wants Intel for super supercomputer". news.techeye.net.
29. (22 September 2011). ""Stampede's" Comprehensive Capabilities to Bolster U.S. Open Science Computational Resources". [[Texas Advanced Computing Center]].
30. "Stampede: A Comprehensive Petascale Computing Environment". IEEE Cluster 2011 Special Topic.
31. Yam, Marcus. (16 November 2011). "Intel's Knights Corner: 50+ Core 22nm Co-processor". [[Tom's Hardware]].
32. Sylvie Barak. (16 November 2011). "Intel unveils 1 TFLOP/s Knights Corner". [[EE Times]].
33. James Reinders. (5 June 2012). "Knights Corner: Open source software stack". Intel.
34. Radek. (18 June 2012). "Chip Shot: Intel Names the Technology to Revolutionize the Future of HPC - Intel Xeon Phi Product Family". Intel.
35. Prickett Morgan, Timothy. (18 June 2012). "Intel slaps Xeon Phi brand on MIC coprocessors". 222.theregister.co.uk.
36. Intel Corporation. (18 June 2012). "Latest Intel Xeon Processors E5 Product Family Achieves Fastest Adoption of New Technology on Top500 List". marketwatch.com.
37. Raj Hazra. (18 June 2012). "Intel Xeon Phi coprocessors accelerate the pace of discovery and innovation". Intel.
38. Rick Merritt. (18 June 2012). "Cray will use Intel MIC, branded Xeon Phi". EETimes.
39. Terrence O'Brien. (18 June 2012). "Intel christens its 'Many Integrated Core' products Xeon Phi, eyes exascale milestone". Engadget.
40. Jeffrey Burt. (18 June 2012). "Intel Wraps Xeon Phi Branding Around MIC Coprocessors". [[eWeek]].
41. Merritt, Rick. (8 June 2012). "Cray will use Intel MIC, branded Xeon Phi". eetimes.com.
42. Latif, Lawrence. (19 June 2012). "Cray to support Intel's Xeon Phi in Cascade clusters". theinquirer.net.
43. (20 June 2012). "ScaleMP vSMP Foundation to Support Intel Xeon Phi". ScaleMP.
44. George Chrysos. (12 November 2012). "Intel Xeon Phi X100 Family Coprocessor - the Architecture".
45. IntelPR. (12 November 2012). "Intel Delivers New Architecture for Discovery with Intel Xeon Phi Coprocessors". Intel.
46. Agam Shah. (12 November 2012). "Intel ships 60-core Xeon Phi processor".
47. Johan De Gelas. (14 November 2012). "The Xeon Phi at work at TACC". AnandTech.
48. (30 July 2012). "Intel's 50-core champion: In-depth on Xeon Phi". Ziff Davis, Inc..
49. (2013). "Manycore Parallelism through OpenMP". Springer.
50. Rick Merritt. (20 June 2011). "OEMs show systems with Intel MIC chips". [[EE Times]].
51. (23 November 2012). "Efficient Hybrid Execution of C++ Applications using Intel Xeon Phi Coprocessor".
52. (20 June 2011). "News Fact Sheet: Intel Many Integrated Core (Intel MIC) Architecture ISC'11 Demos and Performance Description". Intel.
53. [https://cug.org/proceedings/cug2013_proceedings/includes/files/pap199.pdf Tesla vs. Xeon Phi vs. Radeon. A Compiler Writer's Perspective] // The Portland Group (PGI), CUG 2013 Proceedings
54. "Intel Many Integrated Core Architecture (Intel MIC Architecture) - RESOURCES (including downloads)". Intel.
55. (7 September 2012). "Intel Xeon Phi Coprocessor Instruction Set Architecture Reference Manual". Intel.
56. "Intel Developer Zone: Intel Xeon Phi Coprocessor". Intel.
57. "Intel® Intrinsics Guide".
58. "Intel SE3110X Xeon Phi 3110X Knights Corner 6GB Coprocessor-No Cooling -SabrePC.com -SabrePC.com".
59. "Intel Xeon Phi Coprocessor 3120A (6GB, 1.100 GHz, 57 core) Product Specifications". Intel ARK (Product Specs).
60. "Intel Xeon Phi Coprocessor 3120P (6GB, 1.100 GHz, 57 core) Product Specifications". Intel ARK (Product Specs).
61. "Intel Xeon Phi 31S1P - BC31S1P".
62. "Intel Xeon Phi Coprocessor 5110P (8GB, 1.053 GHz, 60 core) Product Specifications". Intel ARK (Product Specs).
63. "Intel Xeon Phi Coprocessor 5120D (8GB, 1.053 GHz, 60 core) Product Specifications". Intel ARK (Product Specs).
64. "Intel Xeon Phi SE10P".
65. "Intel Xeon Phi SE10X".
66. "Intel SC7110P Xeon Phi 7110P Knights Corner Coprocessor -SabrePC.com -SabrePC.com".
67. "Intel SC7110X Xeon Phi 7110X Knights Corner Coprocessor -SabrePC.com -SabrePC.com".
68. "Intel Xeon Phi Coprocessor 7120A (16GB, 1.238 GHz, 61 core) Product Specifications". Intel ARK (Product Specs).
69. "Intel Xeon Phi Coprocessor 7120D (16GB, 1.238 GHz, 61 core) Product Specifications". Intel ARK (Product Specs).
70. "Intel Xeon Phi Coprocessor 7120P (16GB, 1.238 GHz, 61 core) Product Specifications". Intel ARK (Product Specs).
71. "Intel Xeon Phi Coprocessor 7120X (16GB, 1.238 GHz, 61 core) Product Specifications". Intel ARK (Product Specs).
72. (2013-11-25). "Intel Xeon Phi 'Knights Landing' Features Integrated Memory With 500 GB/s Bandwidth and DDR4 Memory Support - Architecture Detailed". WCCFtech.
73. Sebastian Anthony. (26 November 2013). "Intel unveils 72-core x86 Knights Landing CPU for exascale supercomputing". [[ExtremeTech]].
74. [https://www.tomshardware.com/news/intel-xeon-phi-knights-landing,32121.html Tom's Hardware: Intel Xeon Phi Knights Landing Now Shipping; Omni Path Update, Too]. 20 June 2016
75. James Reinders. (23 July 2013). "AVX-512 Instructions". Intel.
76. "Cori".
77. "2016 ISC High Performance: Intel's Rajeeb Hazra Delivers Keynote Address".
78. Pradeep Dubey. (20 June 2016). "How Intel Xeon Phi Processors Benefit Machine Learning/Deep Learning Apps and Frameworks".
79. "Introducing the Intel Xeon Phi Processor – Your Path to Deeper Insight".
80. "Highlights - November 2016 | TOP500".
81. (24 August 2017). "Intel Quietly Drops Xeon Phi 7200 Coprocessors". Phoronix.
82. "Product Change Notification 116378 - 00".
83. "Intel Xeon Phi processor: Your Path to Deeper Insight".
84. "Intel Xeon Phi 7220A - SC7220A / SC7220AEB".
85. "Intel Xeon Phi 7220P - SC7220P".
86. "Intel Xeon Phi 7240P - SC7240P".
87. "Intel Xeon Phi Processor 7210 (16GB, 1.30 GHz, 64 core) Product Specifications". Intel ARK (Product Specs).
88. "Intel Xeon Phi Processor 7210F (16GB, 1.30 GHz, 64 core) Product Specifications". Intel ARK (Product Specs).
89. "Intel Xeon Phi Processor 7230 (16GB, 1.30 GHz, 64 core) Product Specifications". Intel ARK (Product Specs).
90. "Intel Xeon Phi Processor 7230F (16GB, 1.30 GHz, 64 core) Product Specifications". Intel ARK (Product Specs).
91. "Intel Xeon Phi Processor 7250 (16GB, 1.40 GHz, 68 core) Product Specifications". Intel ARK (Product Specs).
92. "Intel Xeon Phi Processors".
93. "Intel Xeon Phi Processor 7250F (16GB, 1.40 GHz, 68 core) Product Specifications". Intel ARK (Product Specs).
94. "Intel Xeon Phi Processor 7290 (16GB, 1.50 GHz, 72 core) Product Specifications". Intel ARK (Product Specs).
95. "Intel Xeon Phi Processor 7290F (16GB, 1.50 GHz, 72 core) Product Specifications". Intel ARK (Product Specs).
96. (17 August 2016). "Intel Announces Knight's Mill: A Xeon Phi for Deep Learning". Anandtech.
97. (19 December 2017). "Intel Lists Knights Mill Xeon Phi on ARK: Up to 72 cores at 320W with QFMA and VNNI". Anandtech.
98. SC14: Supercomputing '14; International Conference for High Performance Computing, Networking, Storage and Analysis in year 2014
99. Eric Gardner. (25 November 2014). "What public disclosures has Intel made about Knights Landing?". [[Intel Corporation]].
100. ALCF staff. (9 April 2015). "Introducing Aurora".
101. ALCF staff. (9 April 2015). "Aurora".
102. (23 May 2017). "Some Surprises in the 2018 DoE Budget for Supercomputing". Next Platform.
103. (16 June 2017). "Is Aurora Morphing into an Exascale AI Supercomputer?". Inside HPC.
104. (13 November 2017). "Unleashing high performance computing today and tomorrow". Intel IT Peer Network.
105. (13 November 2017). "Intel quietly kills off next-gen Knights Hill Xeon Phi chips". Tech Report.
106. (2014). "Test-Driving Intel Xeon Phi".
107. (29 June 2015). "Accelerating DNA Sequence Analysis using Intel Xeon Phi".
108. (30 June 2015). "The Potential of the Intel Xeon Phi for Supervised Deep Learning".
109. Jon Stokes. (20 June 2011). "Intel takes wraps off 50-core supercomputing processor plans".
110. Johan De Gelas. (11 September 2012). "Intel's Xeon Phi in 10 Petaflops supercomputer". AnandTech.