Vista SP1 - The promised performance gains are there
Note: These are very preliminary results - I'll be carrying out more tests over the coming days.
People are very anxious to know whether Windows Vista SP1 brings with it any noticeable performance gains when carrying out day to day tasks. My preliminary results seem to indicate that SP1 offers measurable benefits when carrying out a variety of tasks.
I've carried out a number of tests:
Real world tests
- File copy (disk to disk)
- Copy files to compressed folder
- Extract files from compressed folder
Synthetic
- PassMark PerformanceTest 6.1
- PCMark Vantage
| Vista RTM | Vista SP1 | Change | |
| PassMark PerformanceTest 6.1 | 489.2 pts | 479.5 pts | -9.7 pts |
| Boot time (from BIOS handover to usable desktop) | 53 sec | 44 sec | 9 sec |
| PCMark Vantage | 2701 pts | 2683 pts | -18 pts |
| Copy 2477 files (20 folders) | 239 sec | 199 sec | 40 sec |
| Copy 55 files, 193MB to compressed file | 26 sec | 25 sec | 1 sec |
| Expand 55 files, 193MB from compressed folder | 28 sec | 27 sec | 1 sec |
| Copy 1 file, 664MB to compressed folder | 79 sec | 84 sec | -5 sec |
| Expand 1 file, 664MB from compressed folder | 80 sec | 57 sec | 23 sec |
Note: No system tweaking was carried out to improve the score (no disabling services and so on). I did make sure that all idle tasks were processed and defraged the drives prior to testing, but that was it. I've tried to keep these are real world as possible.
So, overall, real world tests show an improvement while synthetic benchmarks show a small drop in performance (I'm putting this down to drivers - and possibly even the bench software - not being SP1 ready yet). This is promising, but since this data is based on the results from a single test machine, it's too early to declare SP1 a winner.
More to come real soon ...
Over the next few days I’ll be sharing more of my experiences here. If you want to keep updated about SP1 and how it handles a variety of tasks on a selection of PCs, subscribe to the Hardware 2.0 RSS feed.
Thoughts?