Designing for Performance
This technique breaks down instruction execution into sequential steps and allows for instruction-level parallelism.
What is Pipelining?
This approach uses multiple simpler processor cores on one chip instead of a single complex core.
What is Multicore?
This law formulates the maximum speedup possible by parallelizing a program.
What is Amdahl's Law?
This measures the frequency of the clock signal driving the processor.
What is Clock Speed?
This standardized set of programs evaluates CPU performance across different application domains.
What are SPEC CPU Benchmarks?
This performance optimization technique guesses the outcome of branches and executes instructions speculatively.
What is Branch Prediction?
These chips contain a large number of cores, typically over 50, aimed at highly parallel workloads.
What are Many Integrated Cores (MICs)?
This law extends the analysis of Amdahl's Law to any system enhancement.
What is Generalized Amdahl's Law?
This metric represents the average number of clock cycles needed to execute one instruction on a processor.
What are Cycles Per Instruction (CPI)?
This methodology normalizes runtimes against a reference machine and calculates the geometric mean of normalized ratios.
What is SPEC Evaluation Methodology?
This approach analyzes instruction sequences to optimize the scheduling of instructions in pipelines.
What is Data Flow Analysis?
This integration combines general CPU cores with GPU cores on the same chip.
What is CPU+GPU Integration?
This calculation determines the overall performance of a program based on clock speed, instruction count, and CPI.
What is Processor Time Calculation?
This metric reports rates in terms of throughput for parallel/multicore systems.
What is SPECrate?
This technique involves executing instructions along a predicted path before confirming if they are required.
What is Speculative Execution?
This term refers to using the parallel computing capabilities of GPUs for non-graphics applications.
What is General Purpose GPU(GPGPU)?
This measure represents the instruction execution rate calculated from clock speed and CPI.
What is Million Instructions Per Second (MIPS)?
This principle involves balancing throughputs of processor, memory, and I/O devices to prevent bottlenecks.
What is Performance Balance?