I've been seeing tons of people cheering that DX 12 is going to suddenly make single-thread programs into multi-thread programs.
My understanding is that DX 12 is just incredibly more efficient at sending instructions to the GPU, making the load on the CPU lighter.
I mean, if a CPU has to process instructions A, B, C, in order, there is no way to send instruction B to another core until instruction A is done.
Am I missing something? Or are lots of people still just misinformed?
I know that Windows already spreads instructions out to all the available cores, but the program still can't run faster than the slowest thread of instructions.
My understanding is that DX 12 is just incredibly more efficient at sending instructions to the GPU, making the load on the CPU lighter.
I mean, if a CPU has to process instructions A, B, C, in order, there is no way to send instruction B to another core until instruction A is done.
Am I missing something? Or are lots of people still just misinformed?
I know that Windows already spreads instructions out to all the available cores, but the program still can't run faster than the slowest thread of instructions.