3-D Game Play
- 1. Introduction
- 2. Our Test And Comparison Systems, Continued
- 3. Detailed Configuration: Test And Comparison Systems
- 4. Pictures Of The Turion 64/Mobile Sempron Test System
- 5. Mobile Sempron Test System, Continued
- 6. Mobile Sempron Test System, Continued
- 7. Turion 64 And Centrino Notebook Cooling Systems Are Barely Different
- 8. Turion 64 Or Mobile Sempron: Which CPU For What Uses?
- 9. Representative Office Application
- 10. 3-D Game Play
- 11. A Warning To Upgraders: Careful When Upgrading The CPU!
- 12. Battery Tests
- 13. Battery Lifetime
- 14. Battery Charge Times
- 15. Doom 3
- 16. Unreal Tournament 2004
- 17. SiSoftware Sandra Pro (Synthetic Benchmark)
- 18. WinRAR (Data Compression)
- 19. Windows Media Encoder (Video Encoding And Processing)
- 20. Studio 9 Plus (Video Encoding And Processing)
- 21. Conclusion: Goals May Be Met, But There's Still Room For Improvement
10. 3-D Game Play
This last item is the most telling usage scenario for a notebook. For this test we ran through the game demos in 3D Mark 2005 in a continuous loop. As before, we again logged both core voltage and core clock rates, as well as CPU loads.



Under these conditions, all three CPUs occasionally ran at 100% utilization. As a consequence, we observed nearly constant operation at maximum core clock rates and high core voltage levels. Only when one iteration of the loop ended and the next iteration began to load could the CPU take a breather. During that interval, core clock rates and voltage levels would return to their minimum levels for a few seconds. Playing games also shortens a notebook's battery life most, because both CPU and GPU are maxed out nearly all the time. It's also easy to see that because the Mobile Sempron maxes out at a lower clock rate, it also delivers lower frame rates for games when working at full load in concert with the GPU. But a lower frame rate means less work for the GPU, and thus a lower overall load on the notebook as a whole. Somewhat paradoxically, this results in a longer battery life for the slower machine.
Otherwise, it's easy to observe that the curve paths for the Mobile Sempron 2800+ and the Turion MT-34 strongly resemble one another. In sharp contrast, the graph for the Turion 64 ML-34 shows larger, more frequent transitions between minimum and maximum core voltage and core clock rates than the others. This does not imply a CPU defect, as you might presume from this data. The problem lies more in the system, or perhaps, even more with our test design. That's because we didn't ask the vendor before undertaking this series of tests if the notebook's cooling system could handle long-term, sustained peak loads without causing problems for the CPU. In a Turion 64 ML-34 system peak loads can lead to extended periods of power consumption of 34 Watts, but the notebook's cooling system isn't designed to handle the resulting heat dissipation. Whenever the processor temperature exceeds its maximum permissible temperature value, the CPU automatically lowers the core clock rate and core voltage levels, to avoid overheating or perhaps even frying the processor itself. This rapid switching between maximum and minimum core clock and voltage levels results in lower mean heat output. Naturally, this also causes mean performance to diminish over time as well. Because we couldn't determine exactly when this throttle effect kicked in, we decided to conduct our comparison benchmarks only with the Turion 64 MT-34 and the Mobile Sempron in the future.
- Previous page Representative Office Application
- Next page A Warning To Upgraders: Careful When...





please try to keep the detail clear the pictures above are not explained clearly