Theoretical Peak FLOPS per instruction set on modern Intel CPUs
Theoretical Peak FLOPS per instruction set on modern Intel CPUs
复制标题
现代 Intel CPU 上每指令集的理论峰值 FLOPS
DOI：
--
复制
发表时间：
2016
期刊：
影响因子：
--
通讯作者：
Romain Dolbeau Bull
中科院分区：
文献类型：
--
作者：
Romain Dolbeau Bull
过去，用FLOPS（每秒浮点运算次数）来评估CPU的理论峰值性能仅仅是将频率乘以每个周期的浮点指令数。然而，今天的CPU具有诸如向量化、融合乘加、超线程或“turbo”模式等功能。在本文中，我们将研究最近全功能英特尔CPU的理论峰值。不仅考虑到简单的绝对峰值，还考虑到当前Intel CPU的相关指令集和编码以及频率缩放行为。修订版本1.41，2016/10/04 08：49：16
It used to be that evaluating the theoretical peak performance of a CPU in FLOPS (floating point operations per seconds) was merely a matter of multiplying the frequency by the number of floating-point instructions per cycles. Today however, CPUs have features such as vectorization, fused multiply-add, hyper-threading or “turbo” mode. In this paper, we look into this theoretical peak for recent full-featured Intel CPUs., taking into account not only the simple absolute peak, but also the relevant instruction sets and encoding and the frequency scaling behavior of current Intel CPUs. Revision 1.41, 2016/10/04 08:49:16