Vector simplification using shared scratch pad memory

• Main Authors: NORRIE THOMAS, PHELPS ANDREW EVERETT, RAJAMANI, GURUSHANKAR, HEDLUND, MATTHEW, LIVER, JOUPPI NORMAN PAUL
• Format: Patent
• Language: Chinese; English
• Published: 19.01.2024
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS

Methods, systems, and apparatus are described for performing vector simplification using a shared scratchpad memory of a hardware circuit having a processor core in communication with the shared memory. For each processor core, a respective vector of values is generated based on calculations performed at the processor core. The shared memory receives respective vectors of values from respective resources of the processor cores using a direct memory access (DMA) data path of the shared memory. The shared memory performs an accumulation operation on each vector of values using an operator unit coupled to the shared memory. The operator unit is configured to accumulate values based on an arithmetic operation encoded at the operator unit. A result vector is generated based on performing an accumulation operation using the respective value vectors. 描述了用于使用硬件电路的共享便笺式存储器来执行向量简化的方法、系统和装置，该硬件电路具有与共享存储器进行通信的处理器核。对于每个处理器核，基于在处理器核处执行的计算来生成相应的值向量。共享存储器使用共享存储器的直接存储器存取(DMA)数据路径从处理器核的相应资源接收值的相应向量。共享存储器使用耦合到共享存储器的运算符单元对值的各个向量