`mov al, 1` vs `or al, 1` which is faster?

There is no way to tell. You have to try it in your code to see how it performs.

But a major problem with your question is that they do different things. They are not equivalent instructions.

They are same if AL=0.

Only if you don't care about the FLAGS register.

BTW: Having the previous value of AL = 1 also works.

`mov al, 1` because it renames the register and has no input dependency on its previous value.

It might not always be faster, but it is objectively better, as long as it does the same thing (which I'm assuming you know it does in your specific case, since you're asking). It's also the same amount of bytes to encode: 2.

maybe the "faster" aspect of the question is not only related to the "time needed to execute the instruction", but also, "what amount of circuitry is needed to execute the instruction". means that in an extremely RISC system, the question is relevant cause to mov something, you just point to register, and store immediate value. and for a OR, the CPU should point to register, load the register, OR immediate, store in FLAGS and store in register.

but the time spended to ask this question, you missed a lot of fun in your real life...

There is a dependency upon all the other 56 bits of RAX.

Both instructions use previous bits of RAX. Both instructions rename. No difference there. Only one instruction updates the FLAGS. So MOV might be more efficient with power usage. Measure it in your system to see if there is a difference.

If the OP wants a guess, I suspect there will be no measurable difference in execution time in any code or in any CPU. These low level instructions are done in the core of the ALU and are practically identical with regard to how they pass through the execution stages.