Tegra

Tegra is a system on a chip (SoC) series developed by Nvidia for mobile devices such as smartphones, personal digital assistants, and mobile Internet devices. The Tegra integrates an ARM architecture central processing unit (CPU), graphics processing unit (GPU), northbridge, southbridge, and memory controller onto one package. Early Tegra SoCs are designed as efficient multimedia processors, while more recent models emphasize gaming performance without sacrificing power efficiency.

History

The Tegra APX 2500 was announced on February 12, 2008. The Tegra 6xx product line was revealed on June 2, 2008,^[1] and the APX 2600 was announced in February 2009. The APX chips were designed for smartphones, while the Tegra 600 and 650 chips were intended for smartbooks and mobile Internet devices (MID).^[2]

The first product to use the Tegra was Microsoft's Zune HD media player in September 2009, followed by the Samsung M1.^[3] Microsoft's KIN was the first cellular phone to use the Tegra;^[4] however, the phone did not have an app store, so the Tegra's power did not provide much advantage. In September 2008, Nvidia and Opera Software announced that they would produce a version of the Opera 9.5 browser optimised for the Tegra on Windows Mobile and Windows CE.^[5][6] At Mobile World Congress 2009, Nvidia introduced its port of Google's Android to the Tegra.

On January 7, 2010, Nvidia officially announced and demonstrated its next generation Tegra system-on-a-chip, the Nvidia Tegra 250, at Consumer Electronics Show 2010.^[7] Nvidia primarily supports Android on Tegra 2, but booting other ARM-supporting operating systems is possible on devices where the bootloader is accessible. Tegra 2 support for the Ubuntu GNU/Linux distribution was also announced on the Nvidia developer forum.^[8]

Nvidia announced the first quad-core SoC at the February 2011 Mobile World Congress event in Barcelona. Though the chip was codenamed Kal-El, it is now branded as Tegra 3. Early benchmark results show impressive gains over Tegra 2,^[9][10] and the chip was used in many of the tablets released in the second half of 2011.

In January 2012, Nvidia announced that Audi had selected the Tegra 3 processor for its in-vehicle infotainment systems and digital instruments display.^[11] The processor will be integrated into Audi's entire line of vehicles worldwide, beginning in 2013. The process is ISO 26262-certified.^[12]

In summer of 2012 Tesla Motors began shipping the Model S all electric, high performance sedan, which contains two NVIDIA Tegra 3D Visual Computing Modules (VCM). One VCM powers the 17-inch touchscreen infotainment system, and one drives the 12.3-inch all digital instrument cluster."^[13]

In March 2015, Nvidia announced the Tegra X1, the first SoC to have a graphics performance of 1 teraflop. At the announcement event, Nvidia showed off Epic Games' Unreal Engine 4 "Elemental" demo, running on a Tegra X1.

On October 20, 2016, Nvidia announced that Nintendo's upcoming Switch hybrid home/portable game console will be powered by Tegra hardware.^[14]

On March 15, 2017, TechInsights revealed the Nintendo Switch is powered by the Tegra X1 (model T210).^[15]

Specifications

Tegra APX

Tegra APX 2500

• Processor: ARM11 600 MHz MPCore (originally GeForce ULV)
  • Suffix: APX (formerly CSX)
• Memory: NOR or NAND flash, Mobile DDR
• Graphics: Image processor (FWVGA 854×480 pixels)
  • Up to 12 megapixels camera support
  • LCD controller supports resolutions up to 1280×1024
• Storage: IDE for SSD
• Video codecs: up to 720p MPEG-4 AVC/h.264 and VC-1 decoding
• Includes GeForce ULV support for OpenGL ES 2.0, Direct3D Mobile, and programmable shaders
• Output: HDMI, VGA, composite video, S-Video, stereo jack, USB
• USB On-The-Go

Tegra APX 2600

• Enhanced NAND flash
• Video codecs:^[16]
  • 720p H.264 Baseline Profile encode or decode
  • 720p VC-1/WMV9 Advanced Profile decode
  • D-1 MPEG-4 Simple Profile encode or decode

Tegra 6xx

Tegra 600

• Targeted for GPS segment and automotive
• Processor: ARM11 700 MHz MPCore
• Memory: low-power DDR (DDR-333, 166 MHz)
• SXGA, HDMI, USB, stereo jack
• HD camera 720p

Tegra 650

• Targeted for GTX of handheld and notebook
• Processor: ARM11 800 MHz MPCore
• Low power DDR (DDR-400, 200 MHz)
• Less than 1 watt envelope
• HD image processing for advanced digital still camera and HD camcorder functions
• Display supports 1080p at 24 frame/s, HDMI v1.3, WSXGA+ LCD and CRT, and NTSC/PAL TV output
• Direct support for Wi-Fi, disk drives, keyboard, mouse, and other peripherals
• A complete board support package (BSP) to enable fast time to market for Windows Mobile-based designs

Tegra 2

The second generation Tegra SoC has a dual-core ARM Cortex-A9 CPU, an ultra low power (ULP) GeForce GPU,^[17] a 32-bit memory controller with either LPDDR2-600 or DDR2-667 memory, a 32KB/32KB L1 cache per core and a shared 1MB L2 cache.^[18] Tegra 2's Cortex A9 implementation does not include ARM's SIMD extension, NEON. There is a version of the Tegra 2 SoC supporting 3D displays; this SoC uses a higher clocked CPU and GPU.

The Tegra 2 video decoder is largely unchanged from the original Tegra and has limited support for HD formats.^[19] The lack of support for high-profile H.264 is particularly troublesome when using online video streaming services.

Common features:

• CPU cache: L1: 32 KB instruction + 32 KB data, L2: 1 MB
• 40 nm semiconductor technology

Model number	CPU			GPU			Memory				Adoption
	Processor	Cores	Frequency	Micro- architecture	Core configuration1	Frequency	Type	Amount	Bus width	Band- width	Availability
AP20H (Ventana/Unknown)	Cortex-A9	2	1.0 GHz	VLIW-based VEC4 units[20]	4:4:?:?	300 MHz	LPDDR2 300 MHz DDR2 333 MHz	?	32 bit single-channel	2.4 GB/s 2.7 GB/s	Q1 2010
T20 (Harmony/Ventana)						333 MHz
AP25			1.2 GHz			400 MHz					Q1 2011
T25

Devices

Model	Devices
AP20H	Motorola Atrix 4G, Motorola Droid X2, Motorola Photon, LG Optimus 2X / LG Optimus Dual P990 / Optimus 2x SU660 (?), Samsung Galaxy R, Samsung Captivate Glide, ZTE Mimosa X, Micromax Superfone A85, T-Mobile G2X P999, Acer Iconia Tab A200 and A500, LG Optimus Pad, Motorola Xoom,[21] Sony Tablet S, Dell Streak 7, Dell Streak Pro,[22] Asus Slider, Toshiba Thrive[23] tablet, T-Mobile G-Slate
T20	Avionic Design Tamonten Processor Board,[24] Exper EasyPad, Notion Ink Adam tablet, Olivetti OliPad 100, Point of View Mobii 10.1, ViewSonic G Tablet, ViewSonic ViewPad 10s, ASUS Eee Pad Transformer, Samsung Galaxy Tab 10.1, Toshiba AC100, Toshiba Folio 100, Advent Vega, Hannspree Hannspad, Aigo n700, CompuLab Trim-Slice nettop, E-Noa Interpad, Malata Tablet Zpad, MSI 10-inch (250 mm) tablet, Toradex Colibri T20, Lenovo IdeaPad Tablet K1, Lenovo ThinkPad Tablet, Velocity Micro Cruz Tablet L510, Zyrex Onepad SP1110, Zyrex Onepad SP1113G, Acer Iconia Tab A100
AP25	Fusion Garage Grid 10

Tegra 3

The Tegra 3 (codenamed "Kal-El")^[25] is functionally a SoC with a quad-core ARM Cortex-A9 MPCore CPU, but includes a fifth "companion" core in what Nvidia refers to as a "variable SMP architecture".^[26] While all cores are Cortex-A9s, the companion core is manufactured with a low-power silicon process. This core operates transparently to applications and is used to reduce power consumption when processing load is minimal. The main quad-core portion of the CPU powers off in these situations.

Tegra 3 is the first Tegra release to support ARM's SIMD extension, NEON.

The GPU in Tegra 3 is an evolution of the Tegra 2 GPU, with 4 additional pixel shader units and higher clock frequency. It can also output video up to 2560×1600 resolution and supports 1080p MPEG-4 AVC/h.264 40 Mbit/s High-Profile, VC1-AP, and simpler forms of MPEG-4 such as DivX and Xvid.^[27]

The Tegra 3 was released on November 9, 2011.^[28]

Common features:

• CPU cache: L1: 32 KB instruction + 32 KB data, L2: 1 MB
• 40 nm LPG semiconductor technology by TSMC

Model number	CPU			GPU			Memory				Adoption
	Processor	Cores	Frequency (multi- / single-core mode)	Micro- architecture	Core configuration1	Frequency	Type	Amount	Bus width	Band- width	Availability
T30L	Cortex-A9	4+1	1.2 GHz / up to 1.3 GHz	VLIW-based VEC4 units[20]	8:4:?:?	416 MHz	DDR3-1333	?	32 bit single-channel	5.3 GB/s[29]	Q1 2012
T30			1.4 GHz / up to 1.5 GHz			520 MHz	LPDDR2-1066 DDR3-L-1500	?		4.3 GB/s 6.0 GB/s[30]	Q4 2011
AP33
T33			1.6 GHz / up to 1.7 GHz[29]				DDR3-1600	?		6.4 GB/s[29]	Q2 2012

Devices

Model	Devices
T30L	Asus Transformer Pad TF300T, ASUS MeMO Pad Smart ME301T, Microsoft Surface, Nexus 7 (2012),[31] Sony Xperia Tablet S, Acer Iconia Tab A210, WEXLER.TAB 7t, Lenovo IdeaTab A2109, Toshiba AT300 (Excite 10),[32] Toshiba AT10-A (Excite Pure), BLU Quattro 4.5,[33] BLU Quattro 4.5 HD[34] HP Slate 7 Plus
T30	Asus Eee Pad Transformer Prime,[35] IdeaTab K2 / LePad K2,[36] Acer Iconia Tab A510, Acer Iconia Tab A700, ZTE Era, ZTE PF 100, ZTE T98, Toshiba AT270, Asus VivoTab RT, Fuhu Inc. nabi 2 Tablet,[37] Tesla Model S, Kungfu K3,[38] Goophone I5, Olivetti Olipad 3,[39] Microsoft Surface,[40] Lenovo IdeaPad Yoga 11[41][42], Toradex Colibri T30, Toradex Apalis T30, Nvidia Cardhu developer tablet, Realpad Bunaken (RL-P700-QC)
AP33	LG Optimus 4X HD, Fujitsu Arrows X F-10D, HTC One X, XOLO Play T1000,[43]
T33	Asus Transformer Pad Infinity (TF700T), Fujitsu ARROWS X F-02E, Ouya, HTC One X+, Fujitsu Arrows Tab F-05E

Tegra 4

The Tegra 4 (codenamed "Wayne") was announced on January 6, 2013 and is a SoC with a quad-core CPU, but includes a fifth low-power Cortex A15 companion core which is invisible to the OS and performs background tasks to save power. This power-saving configuration is referred to as "variable SMP architecture" and operates like the similar configuration in Tegra 3.^[44]

The GeForce GPU in Tegra 4 is again an evolution of its predecessors. However, numerous feature additions and efficiency improvements were implemented. The number of processing resources was dramatically increased, and clock rate increased as well. In 3D tests, the Tegra 4 GPU is typically several times faster than that of Tegra 3.^[45] Additionally, the Tegra 4 video processor has full support for hardware decoding and encoding of WebM video (up to 1080p 60Mbit/s @ 60fps).^[46]

Along with Tegra 4, Nvidia also introduced i500, an optional software modem based on Nvidia's acquisition of Icera, which can be reprogrammed to support new network standards. It supports category 3 (100Mbit/s) LTE but will later be updated to Category 4 (150Mbit/s).

Common features:

• CPU cache: L1: 32 KB instruction + 32 KB data, L2: 2 MB
• 28 nm HPL semiconductor technology

Model number	CPU			GPU			Memory				Adoption
	Processor	Cores	Frequency	Microarchitecture	Core configuration1	Frequency	Type	Amount	Bus width	Band- width	Availability
T114[47]	Cortex-A15	4+1	up to 1.9 GHz	VLIW-based VEC4 units[48]	72 (48:24:4)[20][48]	672 MHz[49]	DDR3L or LPDDR3	?	32 bit dual-channel	up to 14.9 GB/s (1866 MT/s data rate)[50][51]	Q2 2013[52]

¹ Pixel shaders : Vertex shaders : Pixel pipelines

Devices

Model

Devices

T114

Nvidia Shield Portable, Tegra Note 7, Microsoft Surface 2, HP Slate 7 Extreme,[53] HP Slate 7 Beats Special Edition,[54] HP Slate 8 Pro,[55] HP SlateBook x2,[56] HP SlateBook 14,[57] HP Slate 21,[58] ZTE N988S, nabi Big Tab, Nuvola NP-1, Project Mojo, Asus Transformer Pad TF701T, Toshiba AT10-LE-A (Excite Pro), Vizio 10" tablet, Wexler.Terra 7, Wexler.Terra 10, Acer TA272HUL AIO, Xiaomi Phone 3,[59] Coolpad 大观 4,[60] Audi Tablet,[61] Le Pan TC1020 10.1",[62] Matrimax iPLAY 7[63], Kobo Arc 10HD[64]

Tegra 4i

The Tegra 4i (codenamed "Grey") was announced on February 19, 2013. With hardware support for the same audio and video formats,^[46] but using Cortex-A9 cores instead of Cortex-A15, the Tegra 4i is a low-power variant of the Tegra 4 and is designed for phones and tablets. Unlike its Tegra 4 counterpart, the Tegra 4i also integrates the Icera i500 LTE/HSPA+ baseband processor onto the same die.

Common features:

• 28 nm HPM semiconductor technology
• CPU cache: L1: 32 KB instruction + 32 KB data, L2: 1 MB

Model number	CPU			GPU			Memory				Adoption
	Processor	Cores	Frequency	Microarchitecture	Core configuration1	Frequency	Type	Amount	Bus width	Band- width	Availability
T148?[65]	Cortex-A9 "R4"	4+1	up to 2.0 GHz	VLIW-based VEC4 units[48]	60 (48:12:2)[48]	660 MHz[49]	LPDDR3		32 bit single-channel	6.4–7.5 GB/s (800–933 MHz)[51]	Q1 2014

¹ Pixel shaders : Vertex shaders : Pixel pipelines

Devices

Model	Devices
T148?	Blackphone, LG G2 mini LTE, Wiko Highway 4G,[66] Explay 4Game,[67] Wiko Wax[68][69] QMobile Noir LT-250[70]

Tegra K1

Nvidia’s Tegra K1 (codenamed "Logan") features ARM Cortex-A15 cores in a 4+1 configuration similar to Tegra 4, or Nvidia's 64-bit Project Denver dual-core processor as well as a Kepler graphics processing unit with support for Direct3D 12, OpenGL ES 3.1, CUDA 6.5, OpenGL 4.4/OpenGL 4.5, and Vulkan.^[71][72] Nvidia claims that it outperforms both the Xbox 360 and the PS3, whilst consuming significantly less power.^[73]

Support Adaptive Scalable Texture Compression.^[74]

In late April 2014, Nvidia shipped the "Jetson TK1" development board containing a Tegra K1 SoC and running Ubuntu Linux.^[75]

• Processor:
  • 32-bit variant quad-core ARM Cortex-A15 MPCore R3 + low power companion core
  • or 64-bit variant with dual-core Project Denver^[76] (variant once codenamed "Stark"^[77])
• GPU consisting of 192 ALUs using Kepler technology
• 28 nm HPM process
• Released in Q2 2014
• Power consumption: 8 watts^[73]

Model number	CPU			GPU				Memory				Adoption
	Processor	Cores	Frequency	Micro- architecture	Core configuration1	Frequency	GFLOPS (FP32)	Type	Amount	Bus width	Band- width	Availability
T124[78]	Cortex-A15 R3 (32-bit)	4+1	up to 2.3 GHz[79]	GK20A (Kepler)	192:8:4[80]	756-951 MHz	290–365	DDR3L LPDDR3[80]	max 8 GB with 40-bit address extension2	64 bit	17 GB/s[80]	Q2 2014
T132	Denver (64-bit)	2[80]	up to 2.5 GHz[79]						max 8 GB	?	?	Q3 2014

¹ Unified Shaders : Texture mapping units : Render output units

² ARM Large Physical Page Extension (LPAE) supports 1 TiB (2⁴⁰ bytes). The 8 GiB limitation is part-specific.

Devices

Model	Devices
T124	Jetson TK1 development board,[81] Nvidia Shield Tablet,[82] Acer Chromebook 13,[83] Lenovo ThinkVision 28, Xiaomi MiPad,[84] Snail Games OBox, UTStarcom MC8718, Google Project Tango tablet,[85] Apalis TK1 System on Module,[86] Fuze Tomahawk F1,[87] JXD Singularity S192[88]
T132	HTC Nexus 9[89][90]

Tegra X1

Nvidia's Tegra X1 (codenamed "Erista") features four ARM Cortex-A57 cores and four ARM Cortex-A53 cores, as well as a Maxwell-based graphics processing unit.^[91][92] It supports Adaptive Scalable Texture Compression.^[74] Contrary to initial belief, Nvidia does not use the eight cores in ARM big.LITTLE configuration. Instead devices utilizing the Tegra X1 always show themselves as having only four ARM Cortex-A57 cores available. The other four ARM Cortex-A53 cores can not be accessed by the operating system and are used automatically in very low power scenarios determined by the CPU itself.^[93][94]

• CPU: ARMv8 ARM Cortex-A57 quad-core + ARM Cortex-A53 quad-core (64-bit)
• GPU: Maxwell-based 256 core GPU
• MPEG-4 HEVC & VP9 encoding/decoding support^[95]
• TSMC 20 nm process
• TDP 15 watts,^[96] with average power consumption less than 10 watts^[95]

Model number	CPU			GPU						Memory				Adoption
	Processor	Cores	Frequency	Micro- architecture	Core configuration1	Frequency		GFLOPS (FP32)	GFLOPS (FP16)	Type	Amount	Bus width	Band- width	Availability
T210	Cortex-A57 + Cortex-A53	4 + 4	A57: 1.9 GHz A53: 1.3 GHz	GM20B (Maxwell)	256:16:16	1000 MHz		512	1024	LPDDR4	3 GB	64 bit	25.6 GB/s	Q2 2015
NX-SoC (T210) [15][97]			A57: 1.02 GHz A53: 1.02 GHz			Docked	768-921 MHz[98]	393	786		4 GB			Q1 2017
						Undocked	307.2–384 MHz[98]	157–196	314–392

Devices

Model	Devices
T210	Nvidia Shield Android TV, Nvidia Jetson TX1 development board,[99] Nvidia Drive CX & PX, Google Pixel C, Nintendo Switch[100][15]

Tegra X2

Nvidia's Tegra X2^[101] (codenamed "Parker") features Nvidia’s own custom general-purpose ARMv8-compatible core Denver 2 as well as code-named Pascal graphics processing core with GPGPU support.^[102] The chips are made using FinFET process technology using TSMC's 16 nm FinFET+ manufacturing process.^{[103][104][105]}

• CPU: Nvidia Denver2 ARMv8 (64-bit) dual-core + ARMv8 ARM Cortex-A57 quad-core (64-bit)
• RAM: up to 16GB LPDDR4
• GPU: Pascal-based, 256 CUDA cores
• TSMC 16 nm FinFET process
• TDP: 7.5-15W^[106]

Model number	CPU			GPU				Memory				Adoption
	Processor	Cores	Frequency	Micro- architecture	Core configuration1	Frequency	GFLOPS	Type	Amount	Bus width	Band- width	Availability
T186	Denver2 + Cortex-A57	2 + 4	Denver2: 1.4–2.0 GHz A57: 1.2–2.0 GHz	GP10B (Pascal)[107]	256:?:?	854–1465 MHz	437–750 (FP32) 874–1500 (FP16)	LPDDR4	8 GB	128 bit	58.4 GB/s

Devices

Model	Devices
T186	Nvidia Drive PX2 (variants)
	Nvidia Jetson TX2[106]

Xavier

The most recent Tegra SoC, named Xavier after the comics character Professor X, was announced on 28 September 2016. It is to contain 7 billion transistors and 8 custom ARMv8 cores, a Volta GPU with 512 CUDA cores and is to be made on TSMC 16 nm FinFET+ process. It is able to encode and decode 8K Ultra HD (7680x4320). Nvidia aims for a 20-30 watt TDP and the die size is estimated to be about 300mm².^{[108][109][110]}

• CPU: Nvidia custom ARMv8 (64-bit), 8 cores
• GPU: Volta-based, 512 CUDA cores
• TSMC 16 nm FinFET process
• 20 TOPS DL and 160 SPECint @ 20 W;^[108] 30 TOPS DL @ 30 W^[110] (TOPS DL = Deep Learning Tera-Ops)

Model number	CPU			GPU				Memory				Adoption
	Processor	Cores	Frequency	Micro- architecture	Core configuration1	Frequency	GFLOPS	Type	Amount	Bus width	Band- width	Availability
		8		? (Volta)	512:?:?

Devices

Model	Devices
Unknown	Xavier AI Car Supercomputer
Unknown	Bosch and Nvidia designed Self Driving System[111]

Comparison

Series/ Generation		Tegra 2	Tegra 3	Tegra 4	Tegra K1		Tegra X1		Tegra X2	Xavier
Model/Device		all	all	all	T124	T132	T210	NX-SoC (T210)[15]	all	?
CPU	Instruction set	ARMv7				ARMv8
	Cores	2 A9	4+1 A9	4+1 A15	4+1 A15	2 Denver	4 A53 + 4 A57		2 Denver2 + 4 A57	8 customized
	L1 Cache (I / D)	32 / 32 KB				128 / 64 KB	32 / 32 KB + 64 / 32 KB		128 / 64 KB + 48 / 32 KB	?
	L2 Cache	1MB		2MB			128KB + 2MB		2MB + 2MB	?
GPU	Architecture	Vec4			Kepler		Maxwell		Pascal	Volta
	CUDA Cores	4+4*	8+4*	48+24*	192		256			512
RAM	Protocol	DDR2	DDR3 (+DDR2)	DDR3			LPDDR4			?
	Max. size	1GB	2GB	4GB	8GB				16GB	?
Process		40 nm		28 nm			20 nm		16 nm FF

* Pixel shaders + Vertex shaders. Since Kepler, Unified shaders are used.

Software support

Linux

Nvidia distributes proprietary device drivers for Tegra through OEMs and as part of its "Linux for Tegra" (formerly "L4T") development kit. The newer and more powerful devices of the Tegra family are now supported by Nvidia's own Vibrante Linux distribution. Vibrante comes with a larger set of GNU/Linux tools plus several Nvidia provided libraries for acceleration in the area of data processing and especially image processing for driving safety and automated driving up to the level of deep learning and neuronal networks that make e.g. heavy use of the CUDA capable accelerator blocks, and via OpenCV can make use of the NEON vector extensions of the ARM cores.

As of April 2012, due to different "business needs" from that of their GeForce line of graphics cards, Nvidia and one of their Embedded Partners, Avionic Design GmbH from Germany, are also working on submitting open source drivers for Tegra upstream to the mainline Linux kernel.^[112][113] Nvidia co-founder & CEO laid out the Tegra processor roadmap using Ubuntu Unity in GPU Technology Conference 2013.^[114]

QNX

The DrivePX2 board was announced with QNX RTOS support at the April 2016 GPU Technology Converence.^[115]

Similar platforms

See also

• Project Denver
• Imageon
• Nomadik
• PXA
• Renesas Electronics
• ZiiLABS

References

^[116]

External links

• Official website
• Nvidia's Tegra APX website
• Nvidia's Tegra FAQ
• Tegra X1 Whitepaper
• Tegra K1 Whitepaper
• Tegra 4 CPU Whitepaper
• Tegra 4 GPU Whitepaper
• Tegra 3 Whitepaper
• Tegra 2 Whitepaper