You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I want to use clustering to assign different GPUs to different instances of headless gl called in different threads. Is this possible in any way shape or form?
The text was updated successfully, but these errors were encountered:
Unfortunately, as of right now, there is no way. The closest parameter is powerPreference but it isn't implemented by this library, and is only a hint (you don't get fine-grained control over GPUs directly).
The reason it isn't implemented by this library is that we are using an older version of Angle, and even then, it comes with platform- and device-specific caveats. For example, on Windows, depending on the graphics card, the card's control panel defines which card is used for which process (it's not even as granular as per-thread).
Furthermore, this library likely isn't fully compatible with Node's worker threads (see #151, #182).
I want to use clustering to assign different GPUs to different instances of headless gl called in different threads. Is this possible in any way shape or form?
The text was updated successfully, but these errors were encountered: