dGPU Output Question

Framework Laptop 16
1

I’m sure this has been answered already, so I apologize for repeating since I didn’t see this…

Is the only way to utilize the dGPU with an external monitor via the USB-C port in back? Or is it able to be routed to the modules on the sides?

I’m in batch 14, so I’ve got a while, but my brain is already trying to plan out how to rearrange my desk at home for this, so I figured why not ask.

1 Like
2

The USB-C on the back is explicitly video out only, although you don’t have to use that particular port for that. You can’t use it for anything else. Although you can use a DP or HDMI adapter to change to those interfaces.

AFAIK you can use any of the display out ports on the sides as well.

1 Like
3

Thanks for that. I figured the USB-C on back was exclusively video, but wasn’t sure if the swappable ports on the side would use the dGPU. I had read somewhere those would use the iGUP instead, but I read a lot of things and may be getting my information jumbled.

4

If you use a side port then the displays will be connected to the iGPU.

The dGPU will do most of the GPU work, however the dGPU’s output will be routed through the iGPU, incurring a ~5-10% performance penalty.

The port on the back is display and USB 2.0 data.

4 Likes
5

Ahh okay, thanks for clarifying! That makes sense.

1 Like
6

Ahhh, interesting. Good to know, I hadn’t seen that.

2 Likes
7

To elaborate more on why:

There are a couple reasons why the side ports don’t support the dGPU.

  1. The USB4 ports (the most capable ports) are controlled by the CPU, however modern CPUs only support connecting the USB4 ports to the iGPU. Some older CPUs (ex. the i9 11980HK and other CPUs on the FCBGA1787 socket) did support dGPU over the USB4 ports, however achieving that on modern CPUs would require separate USB4 controllers independent of the CPU, however in general I’ve heard that those controllers have additional performance overhead, additional power draw, and Windows’s built in USB4 support integrates better with integrated USB4 controller than separate (I don’t understand the last one but I’ve heard it, if anyone knows more about that please let me know).

  2. Because Framework has the GPU on a separate board (for the modularity) there has to be a connection between those two boards. The more things (power, PCIe, display signals, USB, etc) that Framework wants to send over that connector, the larger and more expensive the connector becomes. Framework crammed a lot in, however they did decide to only have 1 display signal (which it uses for the integrated display, the integrated display dynamically switches between the iGPU to save power and the dGPU to get more performance depending on what’s needed).

  3. The Framework Laptop 16 can be used without a dGPU. If it had a port connected only to the dGPU then the capabilities of that port would be reduced when the dGPU is removed. To avoid that Framework would need a MUX (a chip that can switch the port between the dGPU and iGPU), which adds additional cost and complexity to the board.

So the side ports are connected to the iGPU (although the system can route the dGPU through the iGPU to allow for ~90% as good performance as being connected to the dGPU ports), the rear port is connected to the dGPU (and only supports display and USB 2.0), and the display dynamically switches between the two.

3 Likes