Iris performance and multiple displays

[mattspace]

Hi Folks,

something i've been wondering - does the performance of integrated graphics change depending on how many monitors are plugged in?

The 2014 mini supports 2x 2560x1600 displays, but does that mean each one only performs at potentially half the capacity of the graphics hardware (edit: if you plug two of them in)?

For example, in a 3d application your major limiting factor in how large a model you can move around responsively in the workspace is the graphics hardware, and how much memory it has available to it. Would that be better on a system that only had a single screen plugged in?

 

[paulrbeers]

Hi Folks,

something i've been wondering - does the performance of integrated graphics change depending on how many monitors are plugged in?

The 2014 mini supports 2x 2560x1600 displays, but does that mean each one only performs at potentially half the capacity of the graphics hardware (edit: if you plug two of them in)?

For example, in a 3d application your major limiting factor in how large a model you can move around responsively in the workspace is the graphics hardware, and how much memory it has available to it. Would that be better on a system that only had a single screen plugged in?

For 2d applications it won't matter. For anything 3d (modeling, games, etc) yes obviously with two screens you are pushing twice the resolution and thus would lower overall GPU performance. However, this is assuming that the images are changing on both screens at any given time. This is also assuming that all monitors in both scenarios are the same (ie there would be little performance difference if you used one 2560x1600 vs 2x1920x1080 since the total resolution is roughly the same)

 

[mattspace]

For 2d applications it won't matter. For anything 3d (modeling, games, etc) yes obviously with two screens you are pushing twice the resolution and thus would lower overall GPU performance. However, this is assuming that the images are changing on both screens at any given time. This is also assuming that all monitors in both scenarios are the same (ie there would be little performance difference if you used one 2560x1600 vs 2x1920x1080 since the total resolution is roughly the same)

So there's a fixed total pool of GPU resources, I guess my question is then how dynamic that is - for example is there a minimum amount used by each screen to render the pixels, then an amount that's used for 3d computation which is shifted dynamically wherever it's needed, or is there a fixed ratio based on screen size / proportion etc, so having a second screen would set a lower ceiling on the potential maximum resources any screen could have.

Then taking into account something like a USB display adapter as an option for extra screens - would that free up more gpu for the primary screen, at the expense of, what, CPU based rendering?