@Atirut @kasdeya they can both be used to get similar effects (having a flat polygon respond to lighting direction differently across its surface) but they're encoding different data, and are incorporated into rendering in different ways!

a bump map is (as it sounds like) single-channel, storing effectively a height-map of the surface detail. it doesn't actually contain any directional information, but this can be reconstructed reasonably well from the bump data

a normal map is three channels, and it stores directional data — it's quite literally storing for each pixel the direction of the "normal" vector (normal here meaning the same thing as the normal force in physics: it's the vector that extends perpendicular to the surface tangent). a normal map doesn't actually have any direct height or displacement information, but is much better at capturing subtle surface details that still affect lighting (like there being a tiny seam between two panels so they reflect light at slightly different angles)

whether a render pipeline is set up for bump maps vs normal maps depends on a lot of things, and they can even be used together in certain scenarios as the kind of detail they capture is more complementary than equivalent, but yeah they're two different ways of trying to enhance a surface's interaction with how light is rendered, and they can often be used to achieve similar effects