First up for testing is Xonotic, one of the most visually impressive and promising open-source first person shooter games.
When comparing the performance with and without Radeon DPM, the Radeon HD 6770 and HD 4890 graphics cards see no change in performance. This should be expected for all three graphics cards since their boot frequencies are their top-rated frequencies supported by their respective video BIOS -- i.e. Radeon DPM can't increase the frequencies any higher than they already were at, unless the open-source driver supported GPU overclocking. The HD 6870 interesting saw a very minor drop in frame-rate with DPM enabled.
The margin is quite small for how the Radeon HD 6870 performance was impacted with DPM enabled, plus the performance was already 200+ FPS, so I'm really not too worried. It's possible the performance could be negatively impacted if the Radeon DPM code is too conservative with its frequency scaling and takes too long to ramp up the clock frequencies, but that's hard to tell as there isn't any convenient sysfs/debugfs interface right now for being able to poll the core/memory frequencies in a real-time manner.
With this first test we see the GPU temperatures are indeed lower with dynamic power management enabled. The Radeon HD 6870 had an average GPU core temperature of 58.8 degrees (Celsius) without DPM and then the average went down to 52.3 degrees when re-clocking was enabled, since the frequencies/voltages were able to be reduced when the GPU wasn't under full load. The Radeon HD 6770 graphics card went from an average of 46.5 degrees to 43.4 degrees when DPM was enabled. The Radeon HD 4890 graphics card actually saw a rise in temperature from 65.7 to 68.8 degrees, but that may just be incidental.
Enabling Radeon DPM also leads to reduced power usage. The HD 6870 went from an average of 174 Watts for Xonotic to 130 Watts with DPM. The HD 6870 dropped from 131 Watts to 121 Watts. The Radeon HD 4890 went from 187 Watts to 178 Watts. These are the power consumption changes namely when the GPU is under load, but will be more noticeable when the GPU is idling. The GPU power consumption is still reduced under load due to PCI-E ASPM now being enabled and other changes in the Radeon DRM code.
While the frame-rates didn't change for the three graphics cards during testing, the lower power consumption led to noticeably higher performance-per-Watt values during testing. (This data is also auto-generated by the Phoronix Test Suite by setting the PERFORMANCE_PER_WATT=1 environment variable.)