[Intel-gfx] [PATCH 31/32] drm/i915: Add background commentary to "waitboosting"

Chris Wilson chris at chris-wilson.co.uk
Fri Dec 11 03:33:27 PST 2015

Describe the intent of boosting the GPU frequency to maximum before
waiting on the GPU.

RPS waitboosting was introduced with

commit b29c19b645287f7062e17d70fa4e9781a01a5d88
Author: Chris Wilson <chris at chris-wilson.co.uk>
Date:   Wed Sep 25 17:34:56 2013 +0100

    drm/i915: Boost RPS frequency for CPU stalls

but lacked a concise comment in the code to explain itself.

Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
 drivers/gpu/drm/i915/i915_gem.c | 16 ++++++++++++++++
 1 file changed, 16 insertions(+)

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index a0584cffa7cd..56b00bf69d89 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -1246,6 +1246,22 @@ int __i915_wait_request(struct drm_i915_gem_request *req,
+	/* This client is about to stall waiting for the GPU. In many cases
+	 * this is undesirable and limits the throughput of the system, as
+	 * many clients cannot continue processing user input/output whilst
+	 * asleep. RPS autotuning may take tens of milliseconds to respond
+	 * to the GPU load and thus incurs additional latency for the client.
+	 * We can circumvent that promoting the GPU frequency to maximum
+	 * before we wait. This makes GPU throttle up much more quickly
+	 * (good for benchmarks), but at a cost of spending more power
+	 * processing the workload (bad for battery). Not all clients even
+	 * want their results immediately and for them we should just let
+	 * the GPU select its own frequency to maximise efficiency.
+	 * To prevent a single client from forcing the clocks too high for
+	 * the whole system, we only allow each client to waitboost once
+	 * in a busy period.
+	 */
 	if (INTEL_INFO(req->i915)->gen >= 6)
 		gen6_rps_boost(req->i915, rps, req->emitted_jiffies);

More information about the Intel-gfx mailing list