Hi All,
I am about to refresh some of the computers in my lab and was wondering if anybody had any thoughts on the current line of Apple computers with regard to stimulus presentation from PsychoPy. Specifically …
-
I understand the triple-buffering in the more recent OSes has created a delay in visual stimulus presentation. How reliable is this delay? That is, is it always a 1 frame delay that could be accounted for by setting onset to be desired onset - 1 frame rather than desire onset?
-
Does anybody have experience or a strong opinion as to whether the Mac Pro, the iMac, or the MacMini is best for stimulus presentation? Obviously, the Mac Pro is the most powerful (and expensive!), but it seems like overkill. However, my concern with the iMac is whether the integrated display plays nicely with a mirrored external display. I worry that the timing of the second external display would be unreliable. Is that a valid concern, or would the external display have the same priority as the integrated display? Finally, the Mini is certainly most cost-effective, but it does not have an independent GPU and therefore probably sub-optimal.
Any and all thoughts will be greatly appreciated.
Happy New Year!
-Andrew
Hi,
I’m a long-time Mac user, so don’t take this as an anti-Mac thing, but is there a particular reason why you want to use a Mac in the lab (rather than for personal or development use, which I certainly do understand)?
We are working on a publication which will exhaustively compare all three platforms on a variety of stimulus and response measures, across a comprehensive range of stimulus software (not just PsychoPy). I’m sad to say that Macs in a lab setting really do come a distant third to Windows and Linux. If you put the same money you’d spend on Mac with a decent graphics card into a PC (Windows or Linux), you will really be able to come up with a no-compromises system that will perform much better for you. Compared to an iMac, you get to dissociate the computer from the display system (with the current issues you note re using an external display). With a Mac Mini, you aren’t stuck with a given display, but, as you mention, you also don’t get a discrete graphics card built-in. And a Mac Pro is well, really just far beyond what is reasonable, price-wise, for this task. Despite its massive hardware capability, it will still be limited in experimental performance terms by certain macOS factors, in the same way as the lowly Mac mini is.
But with PsychoPy at least, you can continue to develop experiments on your personal Mac (of whatever flavour), for relatively seamless deployment in the lab cross-platform.
Hi Michael,
Thanks for you thoughts. My inclination was to go with Mac for the sake of ease, familiarity, and consistency with other lab computers. But it certainly gives me pause if you find Macs placing a “distant third.” About a year ago, I did intend on switching our stim presentation computer to Linux and bought a Linux machine with a high-end graphics card. But, though not completely new to Linux, I still struggled to get things up and running reliably. Maybe I’ll give it another go with PsychoPy 3.
I look forward to reading your publication on the topic!
So a good middle ground in terms of usability might simply be installing Windows 10 on that PC and using the stand-alone PsychoPy app to avoid difficulties with dependencies and configuration.
Probably still worth installing drivers direct from the manufacturer of your graphics card, though, rather than relying on What windows provides automatically.
I always feel a little dirty using Windows but if it is just for a lab PC (which shouldn’t be running anything much else other than PsychoPy), then it is probably tolerable…
1 Like