Gaming performance on rMBP using only external display

[itsygeek]

Hey guys,

I've be scouring the interwebz for a decent answer and haven't found it. So posting a question here.

I've been trying to wait for broadwell refresh but may as well pull the trigger on 13" base model in couple of months. Probably the only game I will play if at all will be Dota 2 and some CS.

Now I've heard you can get decent performance OSx and bootcamp. My question is, if I run the game on external display (@1080p) (using it as my main display with the mbp screen off), will it help push more fps?

 

[Freyqq]

1080p is lower res than the native res on the internal display, but it is unlikely that you'd want to run a game at native res on the internal display anyway.

Light gaming is probably fine, but don't expect to run Crysis 3 at max settings or anything. Dota 2 and CS shouldn't be a problem. Both are designed to run on a wide range of systems.

 

[leman]

If you run a game in Windows using the build-in LCD on a lower resolution, the system has to upscale the image somehow so that it can be displayed on the panel. I am not sure how modern hardware does this. Either there is a hardware scaling unit (then scaling is be free and using an external monitor does not matter at all) or the scaling is done on the GPU by the driver, in which case there will be a small performance hit. Someone should test this out, actually. I just deleted my bootcamp partition so I can't do this... A proper test would be running a fill-limited benchmark on a internal LCD at 1080p and then on the external 1080p monitor in clamshell mode. Both has to be done under Windows, of course. Under OS X, you WILL have a performance hit because of the additional HiDPI rescaling stuff, which IS done on the GPU.

 

[Freyqq]

If you run a game in Windows using the build-in LCD on a lower resolution, the system has to upscale the image somehow so that it can be displayed on the panel. I am not sure how modern hardware does this. Either there is a hardware scaling unit (then scaling is be free and using an external monitor does not matter at all) or the scaling is done on the GPU by the driver, in which case there will be a small performance hit. Someone should test this out, actually. I just deleted my bootcamp partition so I can't do this... A proper test would be running a fill-limited benchmark on a internal LCD at 1080p and then on the external 1080p monitor in clamshell mode. Both has to be done under Windows, of course. Under OS X, you WILL have a performance hit because of the additional HiDPI rescaling stuff, which IS done on the GPU.

it isn't upscaled. It is just stretched.

 

[leman]

it isn't upscaled. It is just stretched.

It's the same thing. You need to map a low-resolution discrete grid to a high-resolution one. Whether you call it stretching, upscaling or resampling, the nature of the process is the same.

 

[Freyqq]

It's the same thing. You need to map a low-resolution discrete grid to a high-resolution one. Whether you call it stretching, upscaling or resampling, the nature of the process is the same.

stretching doesn't affect performance though. I would be highly surprised if performance was affected if you set the resolution at 1920x1080 and tried two different monitors - one a higher potential res than 1920x1080 and one at 1080p. The computer sends the signal to the screen, and the screen does all the work to stretch the picture.

Upscaling is different. It means that the computer is taking the lower resolution image and converts it into a higher resolution image. For instance, the xbox 360 had a custom upscaling chip that took a 540p render and make it look sharp at 1080p. Some DVD players also have an upscaling chip to make the 480p dvd look sharp at 1080p.

 

[leman]

stretching doesn't affect performance though. I would be highly surprised if performance was affected if you set the resolution at 1920x1080 and tried two different monitors - one a higher potential res than 1920x1080 and one at 1080p. The computer sends the signal to the screen, and the screen does all the work to stretch the picture.

Exactly, in your example the upscaling is done by the monitor and it occurs for 'free' (for the sake of terminological quality — upscaling for me is any kind of signal modification that maps a low-res source to a higher-res destination). I am sorry, if I was't clear enough in my post, but I was talking about the internal LCD. And benchmarking in Windows should show whether custom resolutions incur additional cost (i.e., the GPU has to do some extra work), or whether its free (as in the case of an external LCD).