@crunklord420 have you tried memcpy'ing buffer data unconditionally? You should be able to get ~60FPS with linewise updates, and more with bigger chunks, if you skip the dirty pixel check and just let REP MOVS do its work