OpenGL 4+ ?

[jjmafiae]

When will löve start supporting OpenGL 4 and up ?

[slime]

When a useful feature from that is needed.

LÖVE will run on systems that support OpenGL 4, it just doesn't need to take advantage of OpenGL 4 / DirectX 11 - specific features like tessellation shaders or indirect drawing or atomic counters in shaders.

Requiring support for those features would also bump up the minimum system requirements needed to run LÖVE significantly.

[jjmafiae]

Also OpenGL 5 is rumored to get low level hardware access(Similar to AMD's mantel and DX12) will löve ever get that?

[murks]

I'm glad no very recent version of OpenGL is required since I'm stuck at 3.1 for the time being.

[slime]

Also OpenGL 5 is rumored to get low level hardware access(Similar to AMD's mantel and DX12) will löve ever get that?

For what purpose?

[Automatik]

Mantel and DX12 are reputed to be faster, so I guess he's talking about optimizing löve.

[slime]

Having a Mantle/Metal/DX12 backend would make some things a little bit faster, but only for systems that support it - the same systems which already have no trouble running any LÖVE game. Older systems which are less powerful would not be able to use those backends. It would be a bad use of development time compared to the (not very significant) benefits it would give.

The newer lower-level APIs are at their best when used in modern 3D engines like Unreal and Unity, rather than 2D frameworks like LÖVE.

[jjmafiae]

How many people still run systems that are older than 7 years?

well new AMD laptop APUs might see some improvement with OpenGL 5 like less heat and power consumption.

[murks]

Intel graphics chipsets are generally a few years behind, so the machines don't need to be that old.

[Automatik]

How many people still run systems that are older than 7 years?

You'd be surprised.
Also, the drivers are sometimes also the problem : Currently, the Mesa driver(Linux) support OpenGL 3 but not 4. (It's just not implemented) And it's not the worse driver.