Computer Architecture – Difference Between Embedded Controller and Northbridge


I found the terms 'Embedded Controller' 'Northbridge' 'southbridge' 'Platform Controller Hub' 'super I-O' and I am not sure to understand the difference.

Northbridge and southbride difference is clear as it's explained on wikipedia

EC is also explained on wikipedia.

However no mention of each other is on neither page. And I have a feeling that they might be the same?

I am trying to write on an EC register and got confused by the North/southbridge notions. (note this EC is IT8570 for reference).

Best Answer

Northbridge and "Embedded Controller" have absolutely nothing in common, but they have a distant relationship. The situation is fully explained on the referenced Wikipedia page, on the second block diagram.

In short, Northbridge provides a very high-speed level of connectivity to high-speed components of PC like memory and graphics, plus to a somewhat narrower path to Southbridge.

The Southbridge provides connectivity and interfaces to less-high-speed peripherals like USB, SATA, PATA, Networking, PCI bus etc, and includes so-called LPC - Low Pin Count bus, which is 4-bit wide.

The LPC bus connects to "Super-I/O" controller, which provides connectivity to low-speed peripherals as UART, PS/2 keyboard/mouse, parallel port, floppy, BIOS flash, power management functions, fan control, clock control, etc etc.

The confusion arises because the Wikipedia forgot to mention that the "SuperIO" is also called "Embedded Controller". Notable manufacturers are SMSC/MCHP, ITE, Winbond, etc. Usually the chip has a rectangular shape with 128 pins, easy recognizable on PC mainboards.

ADDITION: There are also some beefier versions of "embedded controllers", mostly for small server platforms, like Winbond WPCM450, dubbed as "Baseboard Management controller", which also includes a simple VGA graphics core, so a user can see full-scale boot and system messages without hassles to add a video card. The chip has a lot of other useful features.