SDL3 Enhances Logic For Selecting The Most Performant Vulkan GPU
The SDL3 library that is widely-used by cross-platform games has landed new code for revising how it selects the high performance GPU in multi-GPU systems.
The new code merged today for the Simple DirectMedia Layer project is picking the dedicated GPU with the highest amount of video RAM for the Vulkan renderer.
In cases beyond just having a single dedicated GPU to complement any integrated graphics, the SDL3 library is now set to pick the dedicated GPU with the most amount of vRAM available on the basis that it's likely to be the highest-end graphics card available on the system.
The logic isn't perfect but should pan out for the vast majority of use cases.
Those interested in all the details around this improved Vulkan device ranking for SDL can see this merge request for all the information on this newly-merged code.
The new code merged today for the Simple DirectMedia Layer project is picking the dedicated GPU with the highest amount of video RAM for the Vulkan renderer.
In cases beyond just having a single dedicated GPU to complement any integrated graphics, the SDL3 library is now set to pick the dedicated GPU with the most amount of vRAM available on the basis that it's likely to be the highest-end graphics card available on the system.
"If we prefer high performance, sum up all device local memory (rounded to megabytes) to deviceRank. In the niche case of someone having multiple dedicated GPUs in the same system, this theoretically picks the most powerful one (or at least the one with the most memory!)"
The logic isn't perfect but should pan out for the vast majority of use cases.
👁 SDL3 Vulkan device rank code merged
Those interested in all the details around this improved Vulkan device ranking for SDL can see this merge request for all the information on this newly-merged code.