AES instruction set
An Advanced Encryption Standard instruction set is now integrated into many processors. The purpose of the instruction set is to improve the speed and security of applications performing encryption and decryption using Advanced Encryption Standard (AES).
They are often implemented as instructions implementing a single round of AES along with a special version for the last round which has a slightly different method.
x86 architecture processors
AES-NI (or the Intel Advanced Encryption Standard New Instructions; AES-NI) was the first major implementation. AES-NI is an extension to the x86 instruction set architecture for microprocessors from Intel and AMD proposed by Intel in March 2008.
||Perform one round of an AES encryption flow|
||Perform the last round of an AES encryption flow|
||Perform one round of an AES decryption flow|
||Perform the last round of an AES decryption flow|
||Assist in AES round key generation[note 1]|
||Assist in AES Inverse Mix Columns|
- Westmere based processors, specifically:
- Sandy Bridge processors:
- Ivy Bridge processors
- All i5, i7, Xeon and i3-2115C only
- Haswell processors (all except i3-4000m, Pentium and Celeron)
- Broadwell processors (all except Pentium and Celeron)
- Silvermont/Airmont processors (all except Bay Trail-D and Bay Trail-M)
- Goldmont (and later) processors
- Skylake (and later) processors
Several AMD processors support AES instructions:
- Jaguar processors and newer
- Puma processors and newer
- "Heavy Equipment" processors
- Zen (and later) based processors
Hardware acceleration in other architectures
AES support with unprivileged processor instructions is also available in the latest SPARC processors (T3, T4, T5, M5, and forward) and in latest ARM processors. The SPARC T4 processor, introduced in 2011, has user-level instructions implementing AES rounds. These instructions are in addition to higher level encryption commands. The ARMv8-A processor architecture, announced in 2011, including the ARM Cortex-A53 and A57 (but not previous v7 processors like the Cortex A5, 7, 8, 9, 11, 15) also have user-level instructions which implement AES rounds. In August 2012, IBM announced that the then-forthcoming Power7+ architecture would have AES support. The commands in these architectures are not directly equivalent to the AES-NI commands, but implement similar functionality.
IBM z9 or later mainframe processors support AES as single-opcode (KM, KMC) AES ECB/CBC instructions via IBM's CryptoExpress hardware. These single-instruction AES versions are therefore easier to use than Intel NI ones, but may not be extended to implement other algorithms based on AES round functions (such as the Whirlpool and Grøstl hash functions).
Supporting x86 CPUs
The following chips, while supporting AES hardware acceleration, do not support AES-NI:
- AMD Geode LX processors
- VIA, using VIA PadLock
Programming information is available in ARM Architecture Reference Manual ARMv8, for ARMv8-A architecture profile (Section A2.3 "The Armv8 Cryptographic Extension").
- ARMv8-A architecture
- ARM cryptographic extensions optionally supported on ARM Cortex-A30/50/70 cores
- Cryptographic hardware accelerators/engines
Whilst the RISC-V architecture doesn't include AES-specific instructions, a number of RISC-V chips include integrated AES co-processors. Examples include:
- Dual-core RISC-V 64 bits Sipeed-M1 support AES and SHA256.
- RISC-V architecture based ESP32-C (as well as Xtensa-based ESP32), support AES, SHA, RSA, RNG, HMAC, digital signature and XTS 128 for flash.
- Bouffalo Labs BL602/604 32-bit RISC-V supports various AES and SHA variants.
- Atmel XMEGA (on-chip accelerator with parallel execution, not an instruction)
- SPARC T3 and later processors have hardware support for several cryptographic algorithms, including AES.
- Cavium Octeon MIPS All Cavium Octeon MIPS-based processors have hardware support for several cryptographic algorithms, including AES using special coprocessor 3 instructions.
In AES-NI Performance Analyzed, Patrick Schmid and Achim Roos found "impressive results from a handful of applications already optimized to take advantage of Intel's AES-NI capability". A performance analysis using the Crypto++ security library showed an increase in throughput from approximately 28.0 cycles per byte to 3.5 cycles per byte with AES/GCM versus a Pentium 4 with no acceleration.[failed verification][better source needed]
Most modern compilers can emit AES instructions.
Much security and cryptography software supports the AES instruction set, including the following notable core infrastructure:
- Apple's FileVault 2 full-disk encryption in macOS 10.10+
- NonStop SSH2, NonStop cF SSL Library and BackBox VTC Software in HPE Tandem NonStop OS L-series
- Cryptography API: Next Generation (CNG) (requires Windows 7)
- Linux's Crypto API
- Java 7 HotSpot
- Network Security Services (NSS) version 3.13 and above (used by Firefox and Google Chrome)
- Solaris Cryptographic Framework on Solaris 10 onwards
- FreeBSD's OpenCrypto API (aesni(4) driver)
- OpenSSL 1.0.1 and above
- Go programming language
A fringe use of the AES instruction set involves using it on block ciphers with a similarly-structured S-box, using affine isomorphism to convert between the two. SM4 and Camellia have been accelerated using AES-NI.
- "Intel Software Network". Intel. Archived from the original on 7 April 2008. Retrieved 2008-04-05.
- Shay Gueron (2010). "Intel Advanced Encryption Standard (AES) Instruction Set White Paper" (PDF). Intel. Retrieved 2012-09-20.
- "Intel Product Specification Advanced Search". Intel ARK.
- Shimpi, Anand Lal. "The Sandy Bridge Review: Intel Core i7-2600K, i5-2500K and Core i3-2100 Tested".
- "Intel Product Specification Comparison".
- "AES-NI support in TrueCrypt (Sandy Bridge problem)".
- "Some products can support AES New Instructions with a Processor Configuration update, in particular, i7-2630QM/i7-2635QM, i7-2670QM/i7-2675QM, i5-2430M/i5-2435M, i5-2410M/i5-2415M. Please contact OEM for the BIOS that includes the latest Processor configuration update".
- "Intel Core i3-2115C Processor (3M Cache, 2.00 GHz) Product Specifications".
- "Intel Core i3-4000M Processor (3M Cache, 2.40 GHz) Product Specifications".
- "Following Instructions". AMD. November 22, 2010. Archived from the original on November 26, 2010. Retrieved 2011-01-04.
- Dan Anderson (2011). "SPARC T4 OpenSSL Engine". Oracle. Retrieved 2012-09-20.
- Richard Grisenthwaite (2011). "ARMv8-A Technology Preview" (PDF). ARM. Archived from the original (PDF) on 2018-06-10. Retrieved 2012-09-20.
- Timothy Prickett Morgan (2012). "All the sauce on Big Blue's hot chip: More on Power7+". The Register. Retrieved 2012-09-20.
- "IBM System z10 cryptography". IBM. Retrieved 2014-01-27.
- "AMD Geode LX Processor Family Technical Specifications". AMD.
- "VIA Padlock Security Engine". VIA. Retrieved 2011-11-14.
- Cryptographic Hardware Accelerators on OpenWRT.org
- "VIA Eden-N Processors". VIA. Archived from the original on 2011-11-11. Retrieved 2011-11-14.
- "VIA C7 Processors". VIA. Retrieved 2011-11-14.
- "Arm Architecture Reference Manual Armv8, for Armv8-A architecture profile". ARM. 22 January 2021.
- "Security System/Crypto Engine driver status". sunxi.montjoie.ovh.
- "Linux Cryptographic Acceleration on an i.MX6" (PDF). Linux Foundation. February 2017. Archived from the original (PDF) on 2019-08-26. Retrieved 2018-05-02.
- "Cryptographic module in Snapdragon 805 is FIPS 140-2 certified". Qualcomm.
- "RK3128 - Rockchip Wiki". Rockchip wiki. Archived from the original on 2019-01-28. Retrieved 2018-05-02.
- "The Samsung Exynos 7420 Deep Dive - Inside A Modern 14nm SoC". AnandTech.
- "Sipeed M1 Datasheet v1.1" (PDF). kamami.pl. 2019-03-06. Retrieved 2021-05-03.
- "ESP32 Series Datasheet" (PDF). www.espressif.com. 2021-03-19. Retrieved 2021-05-03.
- "ESP32-C3 WiFi & BLE RISC-V processor is pin-to-pin compatible with ESP8266". CNX-Software. Retrieved 2020-11-22.
- "BL602-Bouffalo Lab (Nanjing) Co., Ltd". www.bouffalolab.com. Retrieved 2021-05-03.
- "Using the XMEGA built-in AES accelerator" (PDF). Retrieved 2014-12-03.
- "Cavium Networks Launches Industry's Broadest Line of Single and Dual Core MIPS64®-based OCTEON™ Processors Targeting Intelligent Next Generation Networks". Archived from the original on 2017-12-07. Retrieved 2016-09-17.
- P. Schmid and A. Roos (2010). "AES-NI Performance Analyzed". Tom's Hardware. Retrieved 2010-08-10.
- T. Krovetz, W. Dai (2010). "How to get fast AES calls?". Crypto++ user group. Retrieved 2010-08-11.
- "Crypto++ 5.6.0 Pentium 4 Benchmarks". Crypto++ Website. 2009. Archived from the original on 19 September 2010. Retrieved 2010-08-10.
- "NonStop SSH Reference Manual". Retrieved 2020-04-09.
- "NonStop cF SSL Library Reference Manual". Retrieved 2020-04-09.
- "BackBox H4.08Tape Encryption Option". Retrieved 2020-04-09.
- "Intel Advanced Encryption Standard Instructions (AES-NI)". Intel. March 2, 2010. Archived from the original on 7 July 2010. Retrieved 2010-07-11.
- "AES-NI enhancements to NSS on Sandy Bridge systems". 2012-05-02. Retrieved 2012-11-25.
- "System Administration Guide: Security Services, Chapter 13 Solaris Cryptographic Framework (Overview)". Oracle. September 2010. Retrieved 2012-11-27.
- "FreeBSD 8.2 Release Notes". FreeBSD.org. 2011-02-24. Archived from the original on 2011-04-12. Retrieved 2011-12-18.
- OpenSSL: CVS Web Interface
- "Cryptographic Backend (GnuTLS 3.6.14)". gnutls.org. Retrieved 2020-06-26.
- "AES-GCM in libsodium". libsodium.org.
- "Hardware Acceleration". www.veracrypt.fr.
- "aes - The Go Programming Language". golang.org. Retrieved 2020-06-26.
- Shimpi, Anand Lal. "The Clarkdale Review: Intel's Core i5 661, i3 540 & i3 530". www.anandtech.com. Retrieved 2020-06-26.
- Saarinen, Markku-Juhani O. (17 April 2020). "mjosaarinen/sm4ni: Demonstration that AES-NI instructions can be used to implement the Chinese Encryption Standard SM4". GitHub.
- Kivilinna, Jussi (2013). Block Ciphers: Fast Implementations on x86-64 Architecture (PDF) (M.Sc.). University of Oulu. pp. 33, 42. Retrieved 2017-06-22.
- Intel Advanced Encryption Standard Instructions (AES-NI)
- AES instruction set whitepaper (2.93 MiB, PDF) from Intel