New P-State Patches Could Boost Intel Graphics Performance Under Some Conditions

Francisco Jerez of Intel's open-source Linux graphics driver team has posted not some new DRM driver patches today but rather CPUFreq/P-State driver patches that could really help Intel integrated graphics performance under some conditions and especially on the lower-power Intel platforms.

The set of nine patches aim to provide GPU-bound energy efficiency improvements to the Intel P-State CPU frequency scaling driver. When using the non-hardware P-State governor of Intel's CPU frequency scaling driver on lower-power platforms like Broxton, an energy efficiency problem was uncovered. Francisco explained, " Under heavy IO load the
current controller tends to increase frequencies to the maximum turbo P-state, partly due to IO wait boosting, partly due to the roughly flat frequency response curve of the current controller, which causes it to ramp frequencies up and down repeatedly for any oscillating workload (think of graphics, audio or disk IO when any of them becomes a bottleneck), severely increasing energy usage relative to a (throughput-wise equivalent) controller able to provide the same average frequency without fluctuation. The core energy efficiency improvement has been observed to be of the order of 20% via RAPL, but it's expected to vary substantially between workloads."

After a round of tuning and a new low-pass filtering controller for Intel CPUs like Broxton, the results appear fairly positive:

The most obvious impact of this series will likely be the overall improvement in graphics performance on systems with an IGP integrated into the processor package (though for the moment this is only enabled on BXT+), because the TDP budget shared among CPU and GPU can frequently become a limiting factor in low-power devices. On heavily TDP-bound devices this series improves performance of virtually any non-trivial graphics rendering by a significant amount (of the order of the energy efficiency improvement for that workload assuming the optimization didn't cause it to become non-TDP-bound).

See [1]-[5] for detailed numbers including various graphics benchmarks and a sample of the Phoronix daily-system-tracker. Some popular graphics benchmarks like GfxBench gl_manhattan31 and gl_4 improve between 5% and 11% on our systems.

More details on these in-development P-State patches via the mailing list. It will certainly be interesting to see how much further they can push/tune this and how well it will work out on other Intel platforms too.