the biggest missed opportunity with SLI/NVLink was not using the high speed shared memory bus to enable the use of a large number of display outputs as if they were all connected to one GPU.
it wasn't until I started trying to daily-drive 5 or 6 monitors that I appreciated just how slow it is for the desktop compositor to move the framebuffer between cards, and how many stutters and other annoyances it causes when moving windows between screens.
@gsuberland I’m curious what you’re using that many monitors for. I have three, and quite frequently think it may be too many with the third just serving as a distraction.
@jon @gsuberland I daily-drive four and the answer is context! One is basically always chat/comms tools, one is whatever I'm keeping an eye on - monitoring etc - and two form my main working environment, often spilling onto the other two if I need more context (e.g. docs for component A, B, and some code all visible at once)
It's only four because my work desk is also my home desk and I use HDMI/USB switches to flip between, and more than four gets Complicated for either Mac or PC.
@jon @gsuberland fwiw it’s pretty common in the City to tile your entire visual field with monitors so that you can keep an eye on markets/chat/email/system logs/analytics/code/all your janky spreadsheets simultaneously and flick between them instantly. Accordingly, it turns out that a bit of UI work so that an app makes it obvious out of the corner of your eye when Something Has Happened but also makes it easy to pick up where you left off after getting distracted an hour ago really pays off.
There’s a trope about twitchy overcaffeinated distractible tech people but IME they have nothing on City folks.
@jon context separation. and ADHD meaning I have a lot of context open.
@jon @gsuberland I said *the* answer - I should have said it's my answer, others may vary!
