录音/制作/创作 吉他 扩声技术 视频技术 作品展示 生活 信息 更多... | 音频应用专卖店

pc vs mac 速度測試 , 結果

( 20 )
12
 
[收藏]
-  第 1 页  -

13670
#1 03-1-2 15:28

pc vs mac 速度測試 , 結果

pc vs mac 速度測試 , 結果..pc 比 mac 快

Apple has been advertising its G4 processors as faster than Intel Pentium 4 chips -- and that megahertz don't matter -- often citing higher performance on certain Photoshop filters. But we were wondering how the fastest Mac stacks up against an equivalently equipped PC in real world situations regularly encountered by digital video compositors. So, with Adobe After Effects 5.5 installed on the mightiest Mac available at this writing, a Power Mac G4 with dual 1GHz processors, and the PC version of After Effects 5.5 installed on a dual Athlon 1800+MP by Polywell, we did a series of six After Effects 5.5 projects that tested a variety of aspects of the rendering speed of each platform. Here are our test results.

First, let's take a look at the two machines we used for testing. We took every opportunity to make the machines as evenly matched as possible. The dual AMD Athlon MP machine, running Windows XP Professional, is a Polywell 890E (click here for a detailed review), which originally had 1GB of RAM, but we removed two of the four 256 MB RAM sticks to equip it with the same amount of RAM as the Mac G4, 512MB. An Athlon 1800+ chip runs at 1.533 GHz, which AMD says is roughly equivalent to a 1.8GHz Pentium 4 chip. This is significant, since currently the fastest Intel chip is now at 2.4GHz, so this isn't the fastest PC available at this date. Against the Athlon MP entry, we have a dual G4 Power Mac running OS X 10.1.3, with two 1GHz Motorola G4 processors. At first glance, this would seem to be a mismatch, but since the G4 chips have an Altivec processor and can do more operations per clock cycle than the Athlon MP, we expected the G4 to be able to easily keep up with its Athlon counterpart in After Effects rendering.

How We Tested
For our benchmarks, we used six projects featured in the excellent new book, After Effects 5.5 Magic, a tutorial tome by New Riders Publishing. Included in the book is a CD with all the QuickTime, Photoshop and Illustrator files for the projects for both Macs and PCs, along with the After Effects 5.5 project files for each platform. We transferred all the project and media files from the CD onto each workstation, and began our testing. Render settings for all the tests in this article were identical on both platforms, set at Best and rendered lossless at 720x486. Each file was output to its platform's native format with no compression added -- on the Windows platform, they were uncompressed Video for Windows files (.avi) and on the Mac OS X platform, they were rendered to the QuickTime animation codec (.mov). We also tried rendering both the Mac and the PC files to the same QuickTime animation codec, but that made no difference in our results. So we decided to stick with each OS's native video format because that would be the most likely method to be used in the real world. Also, even though we discovered it made no difference whatsoever, we used a single IBM 75GB ATA-100 disk on the Polywell PC instead of its original RAID array for the testing.



After the testing was done, the files for each comp were compressed for QuickTime streaming using our custom settings for Sorenson 3 via Discreet Cleaner. Since the streaming QuickTime movies for Mac and PC were identical, we've included just one movie for each test. The times on each graph are represented in minutes and seconds. The Polywell machine, as configured, currently sells for $2564, while the similarly configured Mac G4 sells for $2870.

Simple Animation
The first test involves a simple cel-style animation. It consists of a pict file created in Photoshop that was used as a background for tracing paths, which were then animated and filled with another pict file making the pink color of the face. It's a goofy, simple animation, but it shows how well the processors can calculate the paths and render the final composite. In this test, the Athlon machine handily beat the Mac, taking 9 seconds to render the project, while the Mac took 16 seconds.



Video Composite
Here's a project that's commonplace for After Effects users. It's a composite that uses a number of effects in one shot. This comp uses a green screen layer, a CGI layer and then an element that was shot on film. The background layer is the castle hall behind the actors. The two men were shot against a green screen, and a QuickTime movie of moving smoke was also layered into the shot. All three layers are combined to result in a realistic final composite. This rendered into a 55MB QuickTime movie on the Mac and a 57 MB .avi on the PC even though the shot was only two seconds long. As you can see, in this test the Mac came close to matching the PC, taking only 12 seconds longer to render the composite



Data Project
This test concentrates on animating layers from Photoshop and Illustrator. It animates randomly sequenced 1s and 0s traveling across the screen at different rates. Green lights were also added, making this the kind of graphic that might be used as a background for statistics in a production. The text elements were imported into After Effects as a three-layered Illustrator file. Since these are vector graphics, they could be scaled to any resolution within After Effects without diminishing their resolution. After the text layers were skewed using 3D effects, lighting effects were added to give the numbers a green cast. This kind of project is indicative of many text animation sequences that are commonly created in compositing software.


Gambler
In the book After Effects 5.5 Magic, this project was originally intended to be exported to the SWF (Shockwave Flash) format, but we processed it in 720x486 (QuickTime and AVI) pixels to see how our contestants could render this composite of Illustrator files in a more broadcast-bound situation. It's a good example of a 2D graphic composite that originated in Illustrator, which is many compositors' bread-and-butter. The Mac lost this one by nearly 33%.

12588
#2 03-1-2 17:04
呵呵,这种测试,没有意义,我用G4533能轻易跑48轨音频,用PC就比较吃力,到底谁更快呢?

4680
#3 03-1-3 23:08
Motolola 的芯片的时钟频率,明明比 INTEL 的慢多了,可为啥还那么快呀???

不解!

13670
#4 03-1-2 19:14
48轨音频是mono 還是stereo? 如果用PT , 應該是mono,
但以現在電腦的配置 , 跑多少都不是問題啊呀

12588
#5 03-1-2 19:20
mono,我用P41.5也可以跑,但是加不了什么插件了。在MAC上可以加20多个WAVES插件呢。

9211
#6 03-1-3 00:42
不要看CPU的主频。

还要看是几位的!

我们用的Intel X86是32位的。

我们用的MAC G4是128位的

如果哪个软件商基于128位来开发音频软件,这将是非常快的

13670
#7 03-1-3 12:03
那為什麼視頻會輸給pc呢? 不要忘記video editing / rendering 需求更高的運算能力!
還有一個問題, mac的cpu刻心也是32bit的

12588
#8 03-1-3 12:28
最初由 himhui 发布
[B]……mac的cpu刻心也是32bit的 [/B]


啊?真的?

13670
#9 03-1-3 13:11
現在只有某些SERVER用64 BIT CPU (例如 alpha / itanium), 你要知道 由 32bit 到 64bit 的過程是痛苦和艱巨,
因為沒有軟件商願意投資重寫軟件 ,歷史已有很多失敗過案, AMD快將推出 的 HAMMER (32/64bit兼容) cpu , 也會遇到相同問題 ,因為若沒軟件支持,64bit就等於沒有, INTEL 就比較聰明 , 技術己有32/64bit兼容的cpu (代號Yanhill,半年前已停止 , 一旦AMD成功, 它便可馬上推出 , 再者, INTEL之所以放棄(32/64bit兼容) cpu , 是因為這個效能是底於 純64bit cpu的 , mac 所謂128bit 是不一樣的東西 , 不然, apple 為什麼常常有轉用X86 CPU的念頭?

13670
#10 03-1-3 16:15
這個link是有關IBM 新硏發的POWER 4 64bit cpu ,說明暫不會用於 MAC


http://www.eweek.com/article2/0,3959,449725,00.asp

http://www.eweek.com/article2/0,3959,449725,00.asp

74
#11 03-1-3 19:31
最初由 himhui 发布
[B]這個link是有關IBM 新硏發的POWER 4 64bit cpu ,說明暫不會用於 MAC


http://www.eweek.com/article2/0,3959,449725,00.asp

http://www.eweek.com/article2/0,3959,449725,00.asp [/B]
谁家的64位cpu都很小心的推出
一个不小心就前途尽毁啊。

13670
#12 03-1-3 23:24
所以上面就用了 dual athlon xp 1800 (=P4 1.5G) VS G4 dual 1G 造了個測試 , 証明G4 并非如廣告中說的一樣 , 反而輸給了PC

2035
#13 03-1-4 11:03
最初由 js10th 发布
[B]Motolola 的芯片的时钟频率,明明比 INTEL 的慢多了,可为啥还那么快呀???

不解! [/B]

通用处理器和专用嵌入式的处理器是不一样的。涉及到周边的环境支持,和运作指令的编写!

2035
#14 03-1-4 11:11
最初由 himhui 发布
[B]那為什麼視頻會輸給pc呢? 不要忘記video editing / rendering 需求更高的運算能力!
還有一個問題, mac的cpu刻心也是32bit的 [/B]


能否提供MAC是32位核心的说明资料??

13670
#15 03-1-4 12:33
[QUOTE]最初由 himhui 发布
[B]這個link是有關IBM 新硏發的POWER 4 64bit cpu ,說明暫不會用於 MAC


http://www.eweek.com/article2/0,3959,449725,00.asp


在文中:-
The current bleak economic climate, which has undermined Apple's sales and profits, and the likely high costs of introducing such a new platform may be deterring the company from making the leap from 32-bit to 64-bit computing.

在現時深深損害著Apple之銷售以及嬴利的黯淡經濟氣候中 , Apple或會決定放棄這個可能以高成本引入運算由32bit跳躍到64bit的新系統
您需要登录后才可以回帖 登录 | 注册

本版积分规则

搜索