.png){kind=icon}
返回
Energy-Efficient Floating-Point Unit Design
.png){kind=icon}
PRE
AI
DOI：10.1109/TC.2010.121.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Energy-efficient computation is critical if we are going to continue to scale performance in power-limited systems. For floating-point applications that have large amounts of data parallelism, one should optimize the throughput/mm(2) given a power density constraint. We present a method for creating a trade-off curve that can be used to estimate the maximum floating-point performance given a set of area and power constraints. Looking at FP multiply-add units and ignoring register and memory overheads, we find that in a 90 nm CMOS technology at 1 W/mm(2), one can achieve a performance of 27 GFlops/mm(2) single precision, and 7.5 GFlops/mm(2) double precision. Adding register file overheads reduces the throughput by less than 50 percent if the compute intensity is high. Since the energy of the basic gates is no longer scaling rapidly, to maintain constant power density with scaling requires moving the overall FP architecture to a lower energy/performance point. A 1 W/mm(2) design at 90 nm is a high-energy design, so scaling it to a lower energy design in 45 nm still yields a 7 x performance gain, while a more balanced 0.1 W/mm(2) design only speeds up by 3.5 x when scaled to 45 nm. Performance scaling below 45 nm rapidly decreases, with a projected improvement of only similar to 3 x for both power densities when scaling to a 22 nm technology.
Keyword:
Arithmetic and logic structures
high-speed arithmetic
floating point
fused multiply-add
throughput/mm(2) optimization
期刊
IF:
3.8
论文数:
5.4K
被引数:
9.8K
机构
引用论文
.png){kind=icon}
分享收藏
没有更多内容
没有更多内容