As　CPU　processing　speed　has　slowed　down　year-on-year,　heterogeneous　＇CPU-GPU＇　architectures　combining　multi-core　CPU　and　GPU　accelerators　have　become　increasingly　attractive.　Under　this　backdrop,　the　Heterogeneous　System　Architecture　(HSA)　standard　was　released　in　2012.　New　Accelerated　Processing　Unit　(APU)　architectures　–　AMD　Kaveri　and　Carrizo　–　were　released　in　2014　and　2015　respectively,　and　are　compliant　with　HSA.　These　architectures　incorporate　two　technologies　central　to　HSA,　hUMA　(heterogeneous　Unified　Memory　Access)　and　hQ　(heterogeneous　Queuing).　This　paper　realizes　radix　sort　and　matrix-vector　multiplication　–　two　data-parallel　applications　on　Kaveri　platform.　By　analyzing　the　performance,　a　dynamic　task　scheduling　stratgy　is　proposed.　The　experimental　results　show　that　the　running　efficiency　of　algorithm　can　be　greatly　improved　by　using　APU　with　reasonable　task　scheduling.　In　the　same　way,　the　other　data-parallel　algorithm　would　also　be　optimized　on　these　heterogeneous　multi-core　architecture.　©　Springer　Nature　Singapore　Pte　Ltd.　2018.