|
talk show ghost posted:Couldn't this be solved by clever motherboard design having both the discrete and integrated in use but only use the discrete to actually display anything? It's not that simple. You'd have to also be aware there are 2 GPUs and data would have to flow well from one to the other. There's a host of annoyances. Similar to why no one has a good solution for dynamically switching between an integrated GPU and a discrete one on the fly based on workload - you have to be able to assume the rest of the system is playing nice, which it most definitely is not.
|
# ¿ Jan 3, 2011 08:16 |
|
|
# ¿ Apr 29, 2024 07:00 |
|
DuckConference posted:Don't some of the macbooks do exactly this? Not quite. The most recent Macbook pros will switch from an integrated intel part to the discrete nvidia chip. System apps that are "aware" will run integrated until a switch occurs. But they have to be coded with the assumption that their graphics context can and will be yanked out from underneath them at any time. Apps that aren't aware will always power up the discrete part, even if what they are doing doesn't need that power. For example, a simple Core Animation app will switch EVERYTHING over to discrete. It's a whitelist, not based on computational need.
|
# ¿ Jan 3, 2011 08:41 |