scripts/bench_cpu_mem_bwscript to runcpu_mem_bwand process the results ofcpu_mem_bwscripts/rooflinescript and document to draw graph of roofline model.
cpu_info_test.cppget cpu information(Eg. number of big-core/freq)cpu_inst_gflops_latency.cppmeasure instruction throughput/latencycpu_mem_bw.cppmeasure CPU hierarchical memory bandwidths/latency of micro-kernelscpu_stream.cppmperf version of John McCalpin's STREAM benchmarkcpu_spec_dram_bw.cppmeasure dram bandwidthcpu_pmu_transpose.cppcollect data of cpu pmu eventscpu_tma_transpose.cppARM TMA examplegpu_march_probe.cppget gpu micro-arch parameters(number of register/warp size/Cache Line size)gpu_spec_dram_bw.cppmeasure GPU DRAM Bandwidthgpu_mem_bw.cppmeasure Bandwidth of GPU multi-level cachesgpu_adreno_pmu_test.cppcollect data of Adreno GPU pmu eventsgpu_mali_pmu_test.cppcollect data of Mali GPU pmu eventsgpu_inst_gflops_latency.cppmeasure gpu/OpenCL instruction throughput/latency
cpu_pmu_analysis/- store some study cases on arm cpu platform, keep adding.
mali_pmu_analysis/- store some study cases on mali gpu platform, keep adding.
adreno_pmu_analysis/- store some study cases on adreno gpu platform, keep adding.