RC6 To Be Flipped On For Sandy Bridge, Ivy Bridge
Keith Packard on Friday evening fired off an email and patch to enable RC6 power-savings support on Intel Sandy Bridge and Ivy Bridge hardware where possible. This hardware feature can decrease power usage while also increasing performance in certain workloads.
For Sandy Bridge (Gen6) hardware, RC6 will be enabled when IOMMU is not enabled while for the next-generation Ivy Bridge (Gen7) hardware, RC6 can safely be enabled always. For Ironlake and previous generations of Intel hardware, RC6 isn't being used by default.
The RC6 support state can still be manually toggled via the i915.i915_enable_rc6=1 kernel module parameter.
RC6 allows the Intel hardware to rest in a deeper power-savings state. The feature hasn't been on by default yet since it's caused havoc for some systems with stability issues and other problems.
Back in October I did some Intel Sandy Bridge RC6 benchmarks to show the increased power-savings. That's also when I discovered that RC6 can actually increase the OpenGL performance too. The increase is often quite significant. The Intel explanation about RC6 increasing the graphics performance can be found in this posting.
Recently a patch was also floated that can make RC6 go deeper, but that's not queued up for this kernel cycle.
Keith Packard & Co are hoping to push this RC6-by-default into the Linux 3.2 kernel pending any new issues coming up. However, with Linux already being concerned this cycle and already ranting about one DRM pull, I wouldn't be surprised if he rejects this patch for Linux 3.2. RC6 has been known to cause problems in the past with earlier revisions of the Intel DRM, and since this isn't a regression fix, he may try to postpone it until the Linux 3.3 kernel but we will see in the coming days what happens.
The "final" RC6 patch can be found in this mailing list message.
For Sandy Bridge (Gen6) hardware, RC6 will be enabled when IOMMU is not enabled while for the next-generation Ivy Bridge (Gen7) hardware, RC6 can safely be enabled always. For Ironlake and previous generations of Intel hardware, RC6 isn't being used by default.
The RC6 support state can still be manually toggled via the i915.i915_enable_rc6=1 kernel module parameter.
RC6 allows the Intel hardware to rest in a deeper power-savings state. The feature hasn't been on by default yet since it's caused havoc for some systems with stability issues and other problems.
Back in October I did some Intel Sandy Bridge RC6 benchmarks to show the increased power-savings. That's also when I discovered that RC6 can actually increase the OpenGL performance too. The increase is often quite significant. The Intel explanation about RC6 increasing the graphics performance can be found in this posting.
Recently a patch was also floated that can make RC6 go deeper, but that's not queued up for this kernel cycle.
Keith Packard & Co are hoping to push this RC6-by-default into the Linux 3.2 kernel pending any new issues coming up. However, with Linux already being concerned this cycle and already ranting about one DRM pull, I wouldn't be surprised if he rejects this patch for Linux 3.2. RC6 has been known to cause problems in the past with earlier revisions of the Intel DRM, and since this isn't a regression fix, he may try to postpone it until the Linux 3.3 kernel but we will see in the coming days what happens.
The "final" RC6 patch can be found in this mailing list message.
Add A Comment