Regressions In Catalyst, File-System Benchmarks

Here's a few benchmarks that didn't make the front-page of Phoronix this week but are worthy of some attention. These tests were done by the OpenBenchmarking.org community.

Here's a large file-system comparison. Tested configurations are Reiser4 with and without noatime, EXT4 with and without journaling, EXT4 with the data-ordered mode, Btrfs, ZFS via Debian GNU/kFreeBSD, and lastly is ZFS with compression enabled.

Here's a sample of the Wine benchmarking that CodeWeavers is doing on an automated basis now. This is with the 3DMark 2011, Half-Life 2, and 3DMark 2000 test profiles they have created for inclusion on OpenBenchmarking.org. There's also Unigine Heaven.

We're still working out various tweaks to OpenBenchmarking.org and the Phoromatic Tracker integration via the Ekofisk project, but you should be able to get the idea.

There's also this Wine test that shows a regression in Catalyst 11.3. This performance regression was spotted during their Wine tests between Catalyst 11.2 and 11.3. They've already opened this ATI bug report.

Plus there's many other benchmarks that you can look through from previous days via OpenBenchmarking.org.