intel-gfx.lists.freedesktop.org archive mirror
 help / color / mirror / Atom feed
From: Chris Wilson <chris@chris-wilson.co.uk>
To: intel-gfx@lists.freedesktop.org
Subject: [PATCH 13/13] drm/i915: Use the LLC mode on gen6 for everything but display.
Date: Thu, 14 Apr 2011 10:03:47 +0100	[thread overview]
Message-ID: <1302771827-26112-14-git-send-email-chris@chris-wilson.co.uk> (raw)
In-Reply-To: <1302771827-26112-1-git-send-email-chris@chris-wilson.co.uk>

From: Eric Anholt <eric@anholt.net>

Improves full-screen openarena on my laptop 20.3% +/- 4.0% (n=3)
Improves 800x600 nexuiz on my laptop 12.3% +/- 0.1% (n=3)

We have more room to improve with doing LLC caching for display using
GFDT, and in doing LLC+MLC caching, but this was an easy performance
win and incremental improvement toward those two.

Signed-off-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
---
 drivers/gpu/drm/i915/i915_gem.c |   18 +++++++++++++++++-
 1 files changed, 17 insertions(+), 1 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 46b63c3..7992308 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -3704,7 +3704,23 @@ struct drm_i915_gem_object *i915_gem_alloc_object(struct drm_device *dev,
 	obj->base.write_domain = I915_GEM_DOMAIN_CPU;
 	obj->base.read_domains = I915_GEM_DOMAIN_CPU;
 
-	obj->cache_level = I915_CACHE_NONE;
+	if (IS_GEN6(dev)) {
+		/* On Gen6, we can have the GPU use the LLC (the CPU
+		 * cache) for about a 10% performance improvement
+		 * compared to uncached.  Graphics requests other than
+		 * display scanout are coherent with the CPU in
+		 * accessing this cache.  This means in this mode we
+		 * don't need to clflush on the CPU side, and on the
+		 * GPU side we only need to flush internal caches to
+		 * get data visible to the CPU.
+		 *
+		 * However, we maintain the display planes as UC, and so
+		 * need to rebind when first used as such.
+		 */
+		obj->cache_level = I915_CACHE_LLC;
+	} else
+		obj->cache_level = I915_CACHE_NONE;
+
 	obj->base.driver_private = NULL;
 	obj->fence_reg = I915_FENCE_REG_NONE;
 	INIT_LIST_HEAD(&obj->mm_list);
-- 
1.7.4.1

  parent reply	other threads:[~2011-04-14  9:04 UTC|newest]

Thread overview: 42+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-04-14  9:03 i915 llc for -next Chris Wilson
2011-04-14  9:03 ` [PATCH 01/13] drm/i915: Rename agp_type to cache_level Chris Wilson
2011-04-14 12:39   ` Keith Packard
2011-04-14 20:57     ` [PATCH] " Chris Wilson
2011-04-14  9:03 ` [PATCH 02/13] drm/i915: Do not clflush snooped objects Chris Wilson
2011-04-14  9:03 ` [PATCH 03/13] drm/i915: Introduce i915_gem_object_finish_gpu() Chris Wilson
2011-04-14 16:01   ` Daniel Vetter
2011-04-14  9:03 ` [PATCH 04/13] drm/i915: Introduce i915_gem_object_finish_gtt() Chris Wilson
2011-04-14 16:12   ` Daniel Vetter
2011-04-14 20:20     ` Chris Wilson
2011-05-04 16:47   ` Keith Packard
2011-04-14  9:03 ` [PATCH 05/13] drm/i915/gtt: Split out i915_gem_gtt_rebind_object() Chris Wilson
2011-04-14 16:52   ` Daniel Vetter
2011-04-14  9:03 ` [PATCH 06/13] drm/i915: Add an interface to dynamically change the cache level Chris Wilson
2011-04-14 16:54   ` Daniel Vetter
2011-04-14  9:03 ` [PATCH 07/13] drm/i915: Mark the cursor and the overlay as being part of the display planes Chris Wilson
2011-05-04 17:09   ` Keith Packard
2011-05-04 18:28     ` Chris Wilson
2011-05-04 18:46       ` Keith Packard
2011-05-04 19:47         ` Chris Wilson
2011-04-14  9:03 ` [PATCH 08/13] drm/i915: Pin after setting to the display plane Chris Wilson
2011-04-14 17:34   ` Daniel Vetter
2011-04-14 21:31     ` Chris Wilson
2011-04-15  6:04   ` [PATCH 1/2] drm/i915: Combine pinning " Chris Wilson
2011-04-15  6:04     ` [PATCH 2/2] drm/i915: Use the uncached domain for the display planes Chris Wilson
2011-04-16 10:54       ` Daniel Vetter
2011-04-15 12:11     ` [PATCH 1/2] drm/i915: Combine pinning after setting to the display plane Daniel Vetter
2011-04-16  6:26       ` Chris Wilson
2011-04-16  6:27       ` [PATCH] drm/i915: Combine pinning with " Chris Wilson
2011-04-16 10:52         ` Daniel Vetter
2011-04-16 11:00           ` Chris Wilson
2011-04-14  9:03 ` [PATCH 09/13] drm/i915: Use the uncached domain for the display planes Chris Wilson
2011-04-14  9:03 ` [PATCH 10/13] drm/i915: Use the CPU domain for snooped pwrites Chris Wilson
2011-04-14 17:40   ` Daniel Vetter
2011-04-14  9:03 ` [PATCH 11/13] drm/i915: Prevent mmap access through the GTT of snooped pages Chris Wilson
2011-05-04 17:30   ` Keith Packard
2011-04-14  9:03 ` [PATCH 12/13] drm/i915: Prevent mixing of snooped and tiling modes for old chipsets Chris Wilson
2011-04-14 17:43   ` Daniel Vetter
2011-04-14 20:26     ` Chris Wilson
2011-05-04 17:32   ` Keith Packard
2011-04-14  9:03 ` Chris Wilson [this message]
2011-04-15  6:12 ` i915 llc for -next Chris Wilson

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1302771827-26112-14-git-send-email-chris@chris-wilson.co.uk \
    --to=chris@chris-wilson.co.uk \
    --cc=intel-gfx@lists.freedesktop.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).