180楼#
发布于:2014-02-05 15:32
|
|
181楼#
发布于:2014-02-05 19:29
pcxfirefox:我需要在整合完28.0mini版本的补丁后分离27.0的补丁。以后我直接在hg上操作应该分离会容易些,现在diff -ruNa太费时间了回到原帖帮我看,这个中文问题,我好像打开方式不对! 图片:QQ图片20140205192848.jpg ![]() 哪里我做不对?难道还有哪个代码? |
|
182楼#
发布于:2014-02-05 20:21
|
|
|
183楼#
发布于:2014-02-05 20:30
pcxfirefox:啥东西?看不懂回到原帖缺了一句话 |
|
184楼#
发布于:2014-02-05 20:51
pcxfirefox:啥东西?看不懂回到原帖缺了一句话嘛。。。英文有! 貌似我弄错了,但哪个代码不清楚。。。 |
|
185楼#
发布于:2014-02-07 15:26
js跑分跟原版貌似没啥区别,还稍微低一点。From是原版,TO是pcx sse2版,同一台机器同一个profile。我记得以前版本的pcx跑分要比原版高个10%的样子
TEST COMPARISON FROM TO DETAILS =============================================================================== ** TOTAL **: ?? 205.9ms +/- 1.7% 211.0ms +/- 2.8% not conclusive: might be *1.025x as slow* =============================================================================== 3d: ?? 32.8ms +/- 5.0% 35.3ms +/- 8.5% not conclusive: might be *1.076x as slow* cube: ?? 12.0ms +/- 6.9% 12.7ms +/- 11.9% not conclusive: might be *1.058x as slow* morph: - 6.3ms +/- 5.5% 6.3ms +/- 5.5% raytrace: *1.124x as slow* 14.5ms +/- 6.3% 16.3ms +/- 9.5% significant access: - 18.9ms +/- 7.4% 18.0ms +/- 4.2% binary-trees: - 2.6ms +/- 14.2% 2.6ms +/- 14.2% fannkuch: - 8.3ms +/- 4.2% 8.3ms +/- 4.2% nbody: - 4.4ms +/- 32.7% 3.9ms +/- 5.8% nsieve: - 3.6ms +/- 10.3% 3.2ms +/- 9.4% bitops: - 11.4ms +/- 9.4% 10.7ms +/- 7.1% 3bit-bits-in-byte: ?? 1.0ms +/- 0.0% 1.2ms +/- 37.7% not conclusive: might be *1.20x as slow* bits-in-byte: - 3.9ms +/- 16.0% 3.4ms +/- 10.9% bitwise-and: - 2.5ms +/- 36.3% 2.1ms +/- 10.8% nsieve-bits: - 4.0ms +/- 8.4% 4.0ms +/- 8.4% controlflow: - 2.1ms +/- 10.8% 2.1ms +/- 10.8% recursive: - 2.1ms +/- 10.8% 2.1ms +/- 10.8% crypto: ?? 16.6ms +/- 5.1% 17.1ms +/- 5.7% not conclusive: might be *1.030x as slow* aes: - 8.3ms +/- 8.2% 8.2ms +/- 12.2% md5: ?? 4.6ms +/- 8.0% 5.0ms +/- 11.7% not conclusive: might be *1.087x as slow* sha1: ?? 3.7ms +/- 9.3% 3.9ms +/- 5.8% not conclusive: might be *1.054x as slow* date: ?? 24.6ms +/- 6.2% 26.0ms +/- 10.3% not conclusive: might be *1.057x as slow* format-tofte: *1.059x as slow* 11.8ms +/- 3.8% 12.5ms +/- 4.0% significant format-xparb: ?? 12.8ms +/- 12.6% 13.5ms +/- 17.0% not conclusive: might be *1.055x as slow* math: ?? 18.1ms +/- 4.3% 18.3ms +/- 1.9% not conclusive: might be *1.011x as slow* cordic: - 3.6ms +/- 10.3% 3.3ms +/- 10.5% partial-sums: ?? 12.7ms +/- 2.7% 13.0ms +/- 0.0% not conclusive: might be *1.024x as slow* spectral-norm: ?? 1.8ms +/- 16.7% 2.0ms +/- 0.0% not conclusive: might be *1.111x as slow* regexp: - 15.3ms +/- 3.8% 14.7ms +/- 3.3% dna: - 15.3ms +/- 3.8% 14.7ms +/- 3.3% string: ?? 66.1ms +/- 4.1% 68.8ms +/- 5.3% not conclusive: might be *1.041x as slow* base64: ?? 7.1ms +/- 3.2% 7.6ms +/- 10.1% not conclusive: might be *1.070x as slow* fasta: *1.051x as slow* 7.9ms +/- 2.9% 8.3ms +/- 4.2% significant tagcloud: *1.106x as slow* 17.9ms +/- 1.3% 19.8ms +/- 4.4% significant unpack-code: - 24.8ms +/- 10.3% 24.7ms +/- 9.6% validate-input: - 8.4ms +/- 7.2% 8.4ms +/- 5.9% |
|
186楼#
发布于:2014-02-07 15:37
|
|
|
187楼#
发布于:2014-02-07 15:40
pcxfirefox:不同CPU不同系统有差别感谢作者回复!用你的编译版已经好久了,一直都很好用:) |
|
188楼#
发布于:2014-02-12 15:59
那个硬件加速别去掉啊
|
|
189楼#
发布于:2014-02-12 17:58
|
|
|
190楼#
发布于:2014-02-12 18:40
pcxfirefox:暂时不准备去 只是先做关乎浏览器性能的就不要去掉了,WebRTC这个砍掉没有 |
|
191楼#
发布于:2014-02-12 19:33
|
|
|
192楼#
发布于:2014-02-12 21:09
pcxfirefox:WebRTC肯定去掉了GM要sync,那你打算怎么办sync实在是想精简掉啊 |
|
193楼#
发布于:2014-02-12 23:11
|
|
|
194楼#
发布于:2014-02-13 11:07
pcxfirefox:我有个Workaround的解决方案,保留sync的一部分,然后把其他部分去掉,再把选项隐藏掉能否反馈GM作者让他后期开发调整不依赖sync,这样以后就能吧sync去除干净了 |
|