So, I was thinking about doing the OS X 10.3.4 update when it occurred to me that it is entirely possible that the update could make my already slow 450mhz G4 Cube run even slower. "But how do I test such things," I asked myself. "Dude, benchmarks," was my reply, and I slapped my forehead for not having thought of it earlier.
There happens to be several benchmark tools available for OS X; I picked Haxial’s Benchmark for no particular reason, and ran it, making sure that nothing else was running. I then filed away the results and did the update.
Once the update was completed I ran the benchmark tool again and compared the before and after update results. Much to my relief, there was no big difference.
Benchmark: CPU & RAM | Benchmark: CPU & RAM |
Version: 1.000 | Version: 1.000 |
Date/Time: 27 May 2004, 06:33:21 PM | Date/Time: 27 May 2004, 11:10:08 PM |
Operating System: MacOS 10.3.3 | Operating System: MacOS 10.3.4 |
Total Time: 96,073 millisecs (1:36) | Total Time: 96,193 millisecs (1:36) |
Average Speed: 175.6M bytes/second | Average Speed: 175.4M bytes/second |
Time Speed | Time Speed |
Memory Copy: 8,874 126.2M/sec | Memory Copy: 9,443 118.6M/sec |
Memory Fill: 9,195 522.0M/sec | Memory Fill: 9,159 524.0M/sec |
Memory Clear: 9,656 497.0M/sec | Memory Clear: 9,573 501.4M/sec |
Memory Equal: 9,676 115.7M/sec | Memory Equal: 9,783 114.4M/sec |
CRC: 12,371 51.7M/sec | CRC: 12,274 52.1M/sec |
Quick Hash: 12,371 51.7M/sec | Quick Hash: 12,276 52.1M/sec |
Secure Hash: 9,447 33.8M/sec | Secure Hash: 9,365 34.1M/sec |
Encryption: 24,483 6.5M/sec | Encryption: 24,320 6.5M/sec |
Haxial Benckmark before update | Haxial Benckmark after 10.3.4 update |
---|
Benchmark tools are kinda interesting because they can give you an idea of how well your system is performing, which is especially useful around update time. What’s even better is to compare your results with those of other benchmarkers. Haxial’s Benchmark tool is pretty basic, and doesn’t have much backend support (it is free after all).
There are other free benchmark tools, including SpeedRun, which I didn’t try, and Xbench, which has a Web site where you can send your data and compare it to others. My Cube compared fairly well against other Cubes. I’m not endorsing Haxial’s Benchmark or Xbench, but I’ll use both the next time I update; I’m leaning towards Xbench because it has a more comprehensive battery of tests then Haxial’s product, including graphics and AltiVec tests.
What I like about these benchmark tools is that the average Joe or Jill can run it and see something meaningful from the data, yet, those of us who get a kick out of numbers labeled ‘Mops/sec’ can also get our jollies.
So, benchmarking; it’s fun, interesting, and useful. Give it a shot, see what you think.