内核编译参数对内核性能影响的测试报告

d00m3d

Well, in this case it is rather hard to judge.

I suggest you run few rounds of SCIMARK2 and compare the results.

nonplus

毁灭老大突然间来了个中英对话。。。

话说平时应用中这点差距肉眼分辨不出。。。

d00m3d

Those Chinese words are copy-and-paste from LZ.

I am on trip to the USA at the moment.  Computers available here do not have Chinese input method.  I cannot post anything in Chinese right now.

糊涂

Post by d00m3d;2064449
Those Chinese words are copy-and-paste from LZ.

I am on trip to the USA at the moment.  Computers available here do not have Chinese input method.  I cannot post anything in Chinese right now.

d00m3d版主，这个http://www.uni-ulm.de/acssu/ime/可用

楼主，我觉得如果方便的话，从技术角度讲，让大家多了解一些你自己的那个测试程序的细节比较好一点。

糊涂

（误操作，重复发帖了，有劳版主删之）

d00m3d

Post by 糊涂;2064458
d00m3d版主，这个http://www.uni-ulm.de/acssu/ime/可用

楼主，我觉得如果方便的话，从技术角度讲，让大家多了解一些你自己的那个测试程序的细节比较好一点。

Thank you very much for your hint.

It is still not applicable to me because I do not know pinyin.  

fender010

我那个程序的细节就不说了，用C语言写的，主要的数据类型是double，

有空在找点专业的测试工具测试一下，毕竟一个程序没有什么说服力。

d00m3d

Post by fender010;2064578
我那个程序的细节就不说了，用C语言写的，主要的数据类型是double，

有空在找点专业的测试工具测试一下，毕竟一个程序没有什么说服力。

It looks to me it may not be a matter of your program, I suspect there might have overhead here that could introduce errors.

If a program is too short or need very short processing time, then the overhead may dominate and then the overall time of execution becomes misleading.

fender010

刚刚安装scimark2，一会有结果，哈哈。

fender010

结果在顶楼。