本文介绍了gcc 64bit编译器不提供任何速度优势的处理方法，对大家解决问题具有一定的参考价值，需要的朋友们下面随着小编来一起学习吧！ 问题描述 29岁程序员，3月因学历无情被辞！ 我必须说我没想到这一点。我刚刚在FreeBSD上做了一些措施 6.2和gcc 3.4.6并且在32位和64位模式之间没有显着差异 - 无论是编译速度还是编译速度都没有 编译结果的速度。 作为基准测试我采用了SmallEiffel编译器的引导程序（ 有38个文件约100 kloc）。我得到的时间非常可靠。可靠。 bootstraping（从最后的eiffel 编译器阶段编译c文件，然后使用它重新生成eiffel编译器 和所有工具）采取 32位：2,58（墙），2,57（用户），0,3（系统）= exe大小： 1,757,177字节 64位：2,48（墙），2,42（用户），0,5（系统）= exe大小： 2,179,326字节 这是衡量质量的标准生成的代码 仅汇编 13,9（32位）< -16,2（64位）sec -O0 33,5（32位）< -31,4（64位）sec -02 55,7（32位）< -51,9（64位）sec -03 所以新的寄存器只是没有给出额外的内存开销所消耗的任何好处。 我很惊讶gcc的代码生成部分（编译器的速度为）也是一样的，因为这是最可怕的 弱点我的任何发展（它与eiffel 编译器的工作原理有关。 MSVC快16倍（使用预编译头） gcc（没有预编译头）。我无法再次比较它们因为操作系统现在在不同的机器上但是预编译的标题是 在gcc上几乎无用 - 以及Eiffel生成头文件的方式 已经很完美（每个c文件只有一个标题，每个文件的每个文件都是）。我得到的结果是-8％（gcc 4.x）高达28％（gcc 3.4.6）。 所以我一直认为问题是gcc的代码生成器， 这是有道理的，因为SPARC上的Sun Studio 11每Mhz快4倍 然后是Intel86代码（因为 的简单调用约定编译器）和tinycc比gcc快9倍。两者都没有使用 预编译的头文件。 但额外的寄存器数量似乎并没有减少 的复杂性代码生成器。我真的不明白这一点。我认为这是一个一对一的过度设计的架构，它会减慢整个gnu系列的价值。 不幸的是只有其他可用的c编译器在linux上我可以测试 几乎和gcc一样慢。似乎英特尔只关心速度 的可执行文件。 那里有一些可用的东西（不幸的是，tinycc已经满了 $ b然后是gcc。这将使我的开发变得更加容易。 啊是的，在我的iMac PPC上我也试过了，codewarrior< -gcc只是 给了一点点比小于2倍的性能。但是这个编译器不再存在了。 OpenWatcom也没有开发，lcc- linux32 / lcc-linux64似乎还没有。 解决方案 我不希望别的什么。出于什么原因你会认为 64位更好？ $ b $你在用什么机器？没有它，绝对时间是没有意义的。 是的。代码膨胀和更大的内存占用消耗了新寄存器的好处 。 lcc-win32比gcc快20倍。我认为大多数编译器比gcc更快 ，因为gcc从未关心过这个问题。该团队显然已经实现了其他目标。 确切地说。这种情况的一个症状是-O2生成 更好的代码而不是-O3 ... -O9。 此类软件的问题这是唯一让人感兴趣的是b / b ，因为 认可归于将功能xxx添加到gcc的人。并且 永远不会从从gcc中删除不需要的功能xxx的人。 这意味着外行人对他/她的宠物感兴趣/> 项目，而不是很多人一般关心这个项目。 这意味着很多功能都会添加到软件中，但从来没有 a对所有优化进行了修改，以确定它是否很糟糕 来保持它们。 gcc项目的另一个重大难点是它是一个编译器 应该在任何地方运行，这意味着很多后端，因此很多问题都是b $ b。微软有更少的问题，因为它基本上在x86上运行，而且只在一个操作系统中运行 ... 您可以使用lcc-lin64，但它在Linux下不是免费的，你有 买它。 是的，使用lcc-lin64。 它们可用但需要一些调整。 仅适用于更好的寄存器模型。对于今天的计算机加速CPU来说，ISA似乎真的不那么重要了。很高兴看到 大地址空间没有给予很高的压力。但在进行最后的比较之前，我必须检查一下Boehm-Weisser-GC的速度。 你用的是哪台机器？没有它，绝对时间是没有意义的。无意义。 它是一个AthlonX2 4400，但没有并行执行。数字本身 是 无关紧要我只是想指出数字是多么相等 （小于1％）甚至是exe大小只增加了20％。 起初我甚至预计程序编译错误所以 i添加了一个assert（sizeof（void *）= = 8）但是没关系。 的确如此。我认为这是一个很大的问题。 只要价格合理，就不会有问题。但我需要 更多 信息。我购买了Code-Warriors Linux和Kylix。两者都是 可怕的错误 我从来没有让Kylix编译一个简单的hello世界。但是我跟 弗里德里希谈了一下这个，而且看起来你的方式更好了。有没有网站， i看了他的Q网站上的并且找不到任何东西。 I must say i didn''t expect this. I just did some measures on FreeBSD6.2 with gcc 3.4.6 and there is absolutely no significant differencebetween 32 and 64 bit mode - neither in compilation speed, nor inspeed of the compiled result.As a benchmark i took a bootstrap of the SmallEiffel Compiler (whichhas 38 files with around 100 kloc). The times i got were veryreliable. The bootstraping (compile the c files from last eiffelcompiler stage, and then use this to regenerated the eiffel compilerand all tools) took32Bit: 2,58 (wall), 2,57 (user), 0,3 (system) = exe size:1,757,177 byte64Bit: 2,48 (wall), 2,42 (user) , 0,5 (system) = exe size:2,179,326 byteThis is a measure of the quality of the generated codeThe compilation only13,9 (32bit) <-16,2 (64 bit) sec -O033,5 (32bit) <-31,4 (64 bit) sec -0255,7 (32bit) <-51,9 (64 bit) sec -03So the new registers are just not giving any benefits that are notconsumed by the additional memory overhead.I''m mostly surprised that the code generation part of gcc (speed ofthe compiler) is also the same, because this is the most terribleweakness for any of my development (it has to do with how the eiffelcompiler works). MSVC is 16x faster then (using precompiled headers)gcc (without precompiled headers). I can''t compare them again becausethe OS are now on different machines but precompiled headers arealmost useless on gcc - and the way Eiffel generates the header fileis already perfect (only one header per c file, and first statement ineach file). I got results from -8% (gcc 4.x) upto 28% (gcc 3.4.6) .So i always believed that the problem is the code generator of gcc,which makes sense because Sun Studio 11 on SPARC is 4x faster per Mhzthen the Intel86 code (because of the easy calling conventions for thecompiler) and the tinycc is 9x faster then gcc. Both aren''t usingprecompiled headers.But the additional number of registers seems to not reduce thecomplexity of the code generator. I really don''t understand this. Iguess it is the one-for-all overengineered architecture that slowsdown the whole gnu collection.Unfortunately the only other useable c compiler on linux i could testwas almost as slow as gcc. Seems that Intel does only care about speedof the executable.So is there something useable out there (tinycc is unfortunately fullof bugs and not anymore maintained) that is at least 3-5 times fasterthen gcc. It would make my development much easier.Ah yes, on my iMac PPC i also tried it, codewarrior <-gcc was onlygiving a little bit less then 2x performance. But this compiler doesnot exist anymore. OpenWatcom is also not developed and the lcc-linux32/lcc-linux64 seem to be not yet available. 解决方案I would not expect anything else. For what reason you would think64 bit is better?What machine were you using? Without that, absolute times aremeaningless.Yes. Code bloat, and larger memory footprint eat away the benefitsof the new registers.lcc-win32 is 20 x faster than gcc. I think most compilers are fasterthan gcc since gcc has never cared about this. The team has apparentlyother objectives.Exactly. One of the symptoms for this situation is that -O2 generatesBETTER code thanh -O3 ... -O9.The problem with this type of software is that the only thing thatpeople are interested in is to put some more code, since therecognition goes to "the guy that added feature xxx to gcc" andnever to the "guy that erased unneeded feature xxx from gcc".This means that evertbody has his/her pet interest with theproject, and not a lot of people care about the project in general.This means that many features get added to the software but nevera reviw of all the optimizations is done to see if it is wortwhileto KEEP them.Another big difficulty of the gcc project is that it is a compilerthat should run anywhere, what means a lot of back ends, andtherefore a lot of problems. Microsoft has less problems since it runsessentially on x86 and then only in one operating system...You could use lcc-lin64 but it is NOT free under linux, you haveto buy it.Yes, use lcc-lin64.They are available but need some tweaking.Only for the better register model. For todays computer speed the CPUISA seems really not so important any more. It''s good to see thatthe large address space is not giving a high penality. But i have tocheck thespeed of the Boehm-Weisser-GC before i make a final comparision.What machine were you using? Without that, absolute times aremeaningless.It''s an AthlonX2 4400 but no parallel execution. The numbers itselfareirrelevant I just wanted to point out how equal the numbers are(less then 1%) and even the exe size is only 20% larger.At first i even expected the program was wrongly compiled soi added a "assert(sizeof(void*) == 8)" but it was okay.Indeed. I see this as a huge problem.Don''t have a problem as long as the price is reasonable. But i needsome moreinformation. I purchased Code-Warriors Linux and Kylix. Both wereterrible mistakesand i never got Kylix compile a simple hello world. But i talked withFriedrich aboutit and it seems that you are going a better way. Is there any website,i looked athis Q website and couldn''t find anything. 这篇关于gcc 64bit编译器不提供任何速度优势的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持！ 上岸，阿里云！
