【LibJPEG专题系列十九】FPU可否加快JPEG？DSP芯片又如何？

席萌0209

                    Section - [19] Could an FPU speed up JPEG?
                                        How about a DSP chip?

Since JPEG is so compute-intensive, many people suggest that using an FPU chip (a math coprocessor) should
speed it up.  This is not so. Most production-quality JPEG programs use only integer arithmetic and so they are
unaffected by the presence or absence of floating-point hardware.


It is possible to save a few math operations by doing the DCT step in floating point.  On most PC-class machines,
FP operations are enough slower than integer operations that the overall speed is still much worse with FP.
Some high-priced workstations and supercomputers have fast enough FP hardware to make an FP DCT method
be a win.


DSP (digital signal processing) chips are ideally suited for fast repetitive integer arithmetic, so programming a DSP
to do JPEG can yield significant speedups.  DSPs are available as add-ons for some PCs and workstations; if you
have such hardware, look for a JPEG program that can exploit it.

席萌0209

                                         Section 19—
                    FPU可否加快JPEG？DSP芯片又如何？

        由于JPEG是计算密集型的压缩算法，很多人认为使用FPU芯片（算术协处理器）可以加快其步伐。
事实却并非如此。大多数注重质量的JPEG程序只使用整型运算，因此，他们不受硬件浮点存在与否的影响。


    通过执行浮点的DCT步骤，有可能节省一些数学运算。在一些PC级的机器中，FP运算比整型运算要慢
得多，以至于FP的整体运行速度很糟糕。一些高价位的工作平台以及超级计算机拥有足够快速的FP硬件
使得FP DCT这一运算方法成为赢家。


        DSP（数字信号处理）芯片非常适合快速重复的整型运算，所以在DSP芯片上实现JPEG解码可以明显
地加快运算速度。DSP芯片可附加在一些PC机和工作平台上；如果你有这样的硬件，你可以找一个能够
利用它的JPEG程序。