Difference between revisions of "Tegra/SoCs"

From eLinux.org
Jump to: navigation, search
(Added reference to tegra132 as from Android source)
 
(4 intermediate revisions by 2 users not shown)
Line 3: Line 3:
 
Each SoC has an engineering name (typically TegraNN where NN is some number) and a marketing name (typically "Tegra XX" where XX is some name or number). Software typically uses the engineering name for the chip in order to be isolated from marketing decisions. This wiki primarily uses the engineering names , but this page lists both in order to allow mapping between the two.
 
Each SoC has an engineering name (typically TegraNN where NN is some number) and a marketing name (typically "Tegra XX" where XX is some name or number). Software typically uses the engineering name for the chip in order to be isolated from marketing decisions. This wiki primarily uses the engineering names , but this page lists both in order to allow mapping between the two.
  
See the [[Tegra#Resources|resources]] section of the main [[Tegra]] page for links to Soc documentation etc.
+
See the [[Tegra/Resources|Tegra resources]] page for links to Soc documentation etc.
  
 
= Tegra20 =
 
= Tegra20 =
  
Marketing name: Tegra 2
+
Marketing name: Tegra 2. Other names: T20, AP20, T25, AP25.
Other names: T20, AP20, T25, AP25.
 
  
 
http://www.nvidia.com/object/tegra-superchip.html
 
http://www.nvidia.com/object/tegra-superchip.html
Line 22: Line 21:
 
= Tegra30 =
 
= Tegra30 =
  
Marketing name: Tegra 3, ...
+
Marketing name: Tegra 3. Other names: T30, AP30, (also 33, 37, ...)
Other names: T30, AP30, (also 33, 37, ...)
 
  
 
http://www.nvidia.com/object/tegra-3-processor.html
 
http://www.nvidia.com/object/tegra-3-processor.html
Line 37: Line 35:
 
= Tegra114 =
 
= Tegra114 =
  
Marketing name: Tegra 4, ...
+
Marketing name: Tegra 4.
  
 
http://www.nvidia.com/object/tegra-4-processor.html
 
http://www.nvidia.com/object/tegra-4-processor.html
Line 50: Line 48:
 
= Tegra124 =
 
= Tegra124 =
  
Marketing name: Tegra K1. This is the 32-bit version of Tegra K1.
+
Marketing name: Tegra K1 32-bit. This is the 32-bit version of Tegra K1, as opposed to the later 64-bit version of Tegra K1.
  
http://www.nvidia.com/object/tegra-k1-processor.html
+
General info: [http://www.nvidia.com/object/tegra-k1-processor.html http://www.nvidia.com/object/tegra-k1-processor.html]
 +
 
 +
Detailed Technical Reference Manual: [https://developer.nvidia.com/tegra-k1-technical-reference-manual https://developer.nvidia.com/tegra-k1-technical-reference-manual] (> 2300 pages, requires free registration)
  
 
* Quad-core ARM Cortex-A15 CPU, with 5th battery-saver core
 
* Quad-core ARM Cortex-A15 CPU, with 5th battery-saver core
Line 62: Line 62:
 
* LVDS, HDMI, eDP DSI display
 
* LVDS, HDMI, eDP DSI display
 
* ...
 
* ...
 +
 +
= Tegra132 =
 +
Marketing name: Tegra K1 64-bit ("Denver"). This is a dual-core ARMv8 compatible SoC including the very same GPU as Tegra124.
 +
 +
* Dual-core ARM ARMv8
 +
* GK20A Kepler-based GPU with 192 CUDA cores
 +
* Up to 8GB of RAM
 +
* PCIe (5 lanes, Gen2 capable)
 +
* USB
 +
* MMC/SD
 +
* LVDS, HDMI, eDP DSI display
 +
* ...
 +
 +
[[Category:NVIDIA]]
 +
[[Category:Tegra]]

Latest revision as of 01:29, 4 November 2014

Many NVIDIA Tegra SoCs are supported by mainline OSS software. This page provides a few details of each SoC.

Each SoC has an engineering name (typically TegraNN where NN is some number) and a marketing name (typically "Tegra XX" where XX is some name or number). Software typically uses the engineering name for the chip in order to be isolated from marketing decisions. This wiki primarily uses the engineering names , but this page lists both in order to allow mapping between the two.

See the Tegra resources page for links to Soc documentation etc.

Tegra20

Marketing name: Tegra 2. Other names: T20, AP20, T25, AP25.

http://www.nvidia.com/object/tegra-superchip.html

  • Dual-core ARM Cortex-A9 CPU
  • Up to 1GB of RAM
  • PCIe
  • USB
  • MMC/SD
  • LVDS, VGA, HDMI, DSI display
  • ...

Tegra30

Marketing name: Tegra 3. Other names: T30, AP30, (also 33, 37, ...)

http://www.nvidia.com/object/tegra-3-processor.html

  • Quad-core ARM Cortex-A9 CPU, with 5th battery-saver core
  • Up to 2GB of RAM
  • USB
  • MMC/SD
  • PCIe
  • LVDS, VGA, HDMI, DSI display
  • ...

Tegra114

Marketing name: Tegra 4.

http://www.nvidia.com/object/tegra-4-processor.html

  • Quad-core ARM Cortex-A15 CPU, with 5th battery-saver core
  • Up to 4GB of RAM (1 or 2 is much more typical on mainline boards)
  • USB
  • MMC/SD
  • LVDS, HDMI, DSI display
  • ...

Tegra124

Marketing name: Tegra K1 32-bit. This is the 32-bit version of Tegra K1, as opposed to the later 64-bit version of Tegra K1.

General info: http://www.nvidia.com/object/tegra-k1-processor.html

Detailed Technical Reference Manual: https://developer.nvidia.com/tegra-k1-technical-reference-manual (> 2300 pages, requires free registration)

  • Quad-core ARM Cortex-A15 CPU, with 5th battery-saver core
  • gk20a Kepler-based GPU with 192 CUDA cores
  • Up to 8GB of RAM (1 or 2 is much more typical on mainline boards)
  • PCIe (5 lanes, Gen2 capable)
  • USB
  • MMC/SD
  • LVDS, HDMI, eDP DSI display
  • ...

Tegra132

Marketing name: Tegra K1 64-bit ("Denver"). This is a dual-core ARMv8 compatible SoC including the very same GPU as Tegra124.

  • Dual-core ARM ARMv8
  • GK20A Kepler-based GPU with 192 CUDA cores
  • Up to 8GB of RAM
  • PCIe (5 lanes, Gen2 capable)
  • USB
  • MMC/SD
  • LVDS, HDMI, eDP DSI display
  • ...