In this research, we address server platforms utilizing
multicore CPUs that can be frequency scaled, such as the
AMD Opteron or Intel Nehalem processors, running a
typical server load, which may include a heavily loaded
system running a task consisting of up to twenty-four
concurrently executing benchmarks. Also, in order to
exclude all other variables, we wanted to assess the
potential power savings and performance improvements
achieved by CPU scheduling and multicore CPU frequency
scaling alone, assuming all other platform resources upon
various server systems are equal, including CPUs or hard
drives. Hence, in our survey, we evaluate hardware
components that are identical or similar, which may be
the case in a typical server rack.
In this research, we address server platforms utilizingmulticore CPUs that can be frequency scaled, such as theAMD Opteron or Intel Nehalem processors, running atypical server load, which may include a heavily loadedsystem running a task consisting of up to twenty-fourconcurrently executing benchmarks. Also, in order toexclude all other variables, we wanted to assess thepotential power savings and performance improvementsachieved by CPU scheduling and multicore CPU frequencyscaling alone, assuming all other platform resources uponvarious server systems are equal, including CPUs or harddrives. Hence, in our survey, we evaluate hardwarecomponents that are identical or similar, which may bethe case in a typical server rack.
การแปล กรุณารอสักครู่..
