Skip to contentSkip to conversation

drm/i915/hsw: Change default LLC age to 3

PATCH Superseded Ben Widawsky 3 participants 9 messages Jul 31, 2013 – Aug 5, 2013

Subsystems (2)

v1 of 4 Superseded Latest is v4 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 drm/i915/hsw: Change default LLC age to 3

The default LLC age was changed:
commit 0d8ff15e9a15f2b393e53337a107b7a1e5919b6d

Author: Ben Widawsky <[email protected]>
Date:   Thu Jul 4 11:02:03 2013 -0700

drm/i915/hsw: Set correct Haswell PTE encodings.

This caused a regression in performance on certain benchmarks. While I think a discussion still needs to happen about how the kernel should default for both eLLC, and LLC - just revert this behavior for now.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=67062

---
 drivers/gpu/drm/i915/i915_gem_gtt.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/i915_gem_gtt.c +3 -1
index e7b4204..4b1e6e3 100644
@@ -52,8 +52,10 @@
  */
 #define HSW_CACHEABILITY_CONTROL(bits)	((((bits) & 0x7) << 1) | \
 					 (((bits) & 0x8) << (11 - 3)))
+#define HSW_WB_LLC_AGE3			HSW_CACHEABILITY_CONTROL(0x2)
 #define HSW_WB_LLC_AGE0			HSW_CACHEABILITY_CONTROL(0x3)
 #define HSW_WB_ELLC_LLC_AGE0		HSW_CACHEABILITY_CONTROL(0xb)
+#define HSW_LLC				HSW_WB_LLC_AGE3
 
 static gen6_gtt_pte_t gen6_pte_encode(dma_addr_t addr,
 				      enum i915_cache_level level)
@@ -105,7 +107,7 @@ static gen6_gtt_pte_t hsw_pte_encode(dma_addr_t addr,
 	pte |= HSW_PTE_ADDR_ENCODE(addr);
 
 	if (level != I915_CACHE_NONE)
-		pte |= HSW_WB_LLC_AGE0;
+		pte |= HSW_LLC;
 
 	return pte;
 }
Signature · 1 line
1.8.3.4
Reviewed-by on v1 1/1
#1 lore raw
Quoted 34 lines · Ben Widawsky

The default LLC age was changed:
commit 0d8ff15e9a15f2b393e53337a107b7a1e5919b6d

Author: Ben Widawsky <[email protected]>
Date:   Thu Jul 4 11:02:03 2013 -0700

drm/i915/hsw: Set correct Haswell PTE encodings.

This caused a regression in performance on certain benchmarks. While I think a discussion still needs to happen about how the kernel should default for both eLLC, and LLC - just revert this behavior for now.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=67062

---
  drivers/gpu/drm/i915/i915_gem_gtt.c | 4 +++-
  1 file changed, 3 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/i915_gem_gtt.c +2 -0
index e7b4204..4b1e6e3 100644
@@ -52,8 +52,10 @@
   */
  #define HSW_CACHEABILITY_CONTROL(bits)	((((bits) & 0x7) << 1) | \
  					 (((bits) & 0x8) << (11 - 3)))
+#define HSW_WB_LLC_AGE3			HSW_CACHEABILITY_CONTROL(0x2)
  #define HSW_WB_LLC_AGE0			HSW_CACHEABILITY_CONTROL(0x3)
  #define HSW_WB_ELLC_LLC_AGE0		HSW_CACHEABILITY_CONTROL(0xb)
+#define HSW_LLC				HSW_WB_LLC_AGE3
 
  static gen6_gtt_pte_t gen6_pte_encode(dma_addr_t addr,
  				      enum i915_cache_level level)

drivers/gpu/drm/i915/i915_gem_gtt.c static gen6_gtt_pte_t hsw_pte_encode(dma_addr_t addr, · line 113

Quoted 8 lines
drivers/gpu/drm/i915/i915_gem_gtt.c +1 -1
@@ -105,7 +107,7 @@ static gen6_gtt_pte_t hsw_pte_encode(dma_addr_t addr,
  	pte |= HSW_PTE_ADDR_ENCODE(addr);
 
  	if (level != I915_CACHE_NONE)
-		pte |= HSW_WB_LLC_AGE0;
+		pte |= HSW_LLC;
 
  	return pte;
  }

I think the patch is clearer without the wrapper define HSW_LLC. But, anyways, with or without the wrapper,

#2 lore raw

PATCH v2 1/1 drm/i915/hsw: Change default LLC age to 3

The default LLC age was changed:
commit 0d8ff15e9a15f2b393e53337a107b7a1e5919b6d

Author: Ben Widawsky <[email protected]>
Date:   Thu Jul 4 11:02:03 2013 -0700

drm/i915/hsw: Set correct Haswell PTE encodings.

This caused a regression in performance on certain benchmarks. While I think a discussion still needs to happen about how the kernel should default for both eLLC, and LLC - just revert this behavior for now.

v2: Drop the extra #define (Chad)

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=67062

---
 drivers/gpu/drm/i915/i915_gem_gtt.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/i915_gem_gtt.c +3 -1
index e7b4204..4b1e6e3 100644
@@ -52,8 +52,10 @@
  */
 #define HSW_CACHEABILITY_CONTROL(bits)	((((bits) & 0x7) << 1) | \
 					 (((bits) & 0x8) << (11 - 3)))
+#define HSW_WB_LLC_AGE3			HSW_CACHEABILITY_CONTROL(0x2)
 #define HSW_WB_LLC_AGE0			HSW_CACHEABILITY_CONTROL(0x3)
 #define HSW_WB_ELLC_LLC_AGE0		HSW_CACHEABILITY_CONTROL(0xb)
+#define HSW_LLC				HSW_WB_LLC_AGE3
 
 static gen6_gtt_pte_t gen6_pte_encode(dma_addr_t addr,
 				      enum i915_cache_level level)
@@ -105,7 +107,7 @@ static gen6_gtt_pte_t hsw_pte_encode(dma_addr_t addr,
 	pte |= HSW_PTE_ADDR_ENCODE(addr);
 
 	if (level != I915_CACHE_NONE)
-		pte |= HSW_WB_LLC_AGE0;
+		pte |= HSW_LLC;
 
 	return pte;
 }
Signature · 1 line
1.8.3.4
#3 lore raw

PATCH v3 1/1 drm/i915/hsw: Change default LLC age to 3

The default LLC age was changed:
commit 0d8ff15e9a15f2b393e53337a107b7a1e5919b6d

Author: Ben Widawsky <[email protected]>
Date:   Thu Jul 4 11:02:03 2013 -0700

drm/i915/hsw: Set correct Haswell PTE encodings.

This caused a regression in performance on certain benchmarks. While I think a discussion still needs to happen about how the kernel should default for both eLLC, and LLC - just revert this behavior for now.

v2: Drop the extra #define (Chad)

v3: Actually git add

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=67062

---
 drivers/gpu/drm/i915/i915_gem_gtt.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/i915_gem_gtt.c +2 -1
index e7b4204..3e7f124 100644
@@ -52,6 +52,7 @@
  */
 #define HSW_CACHEABILITY_CONTROL(bits)	((((bits) & 0x7) << 1) | \
 					 (((bits) & 0x8) << (11 - 3)))
+#define HSW_WB_LLC_AGE3			HSW_CACHEABILITY_CONTROL(0x2)
 #define HSW_WB_LLC_AGE0			HSW_CACHEABILITY_CONTROL(0x3)
 #define HSW_WB_ELLC_LLC_AGE0		HSW_CACHEABILITY_CONTROL(0xb)
 
@@ -105,7 +106,7 @@ static gen6_gtt_pte_t hsw_pte_encode(dma_addr_t addr,
 	pte |= HSW_PTE_ADDR_ENCODE(addr);
 
 	if (level != I915_CACHE_NONE)
-		pte |= HSW_WB_LLC_AGE0;
+		pte |= HSW_WB_LLC_AGE3;
 
 	return pte;
 }
Signature · 1 line
1.8.3.4
#4 lore raw
Quoted 11 lines · Ben Widawsky

The default LLC age was changed:
commit 0d8ff15e9a15f2b393e53337a107b7a1e5919b6d

Author: Ben Widawsky <[email protected]>
Date:   Thu Jul 4 11:02:03 2013 -0700

drm/i915/hsw: Set correct Haswell PTE encodings.

This caused a regression in performance on certain benchmarks. While I think a discussion still needs to happen about how the kernel should default for both eLLC, and LLC - just revert this behavior for now.

Quoted 7 lines

v2: Drop the extra #define (Chad)

v3: Actually git add

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=67062

iris_pte_encode was split up after the regressing commit in

commit 4d15c145a6234d999c0452eec0d275c1fbf0688c

Author: Ben Widawsky <[email protected]>
Date:   Thu Jul 4 11:02:06 2013 -0700

drm/i915: Use eLLC/LLC by default when available

Hence I think that one should be patched, too (QA doesn't have an iris machine afaik, so no regression report expected). And since I have some open questions about Chris WT-for-Iris patch too I think it's best to spin a v4 of this one here.
-Daniel

Quoted 28 lines
---
 drivers/gpu/drm/i915/i915_gem_gtt.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/i915_gem_gtt.c +2 -2
index e7b4204..3e7f124 100644
@@ -52,6 +52,7 @@
  */
 #define HSW_CACHEABILITY_CONTROL(bits) ((((bits) & 0x7) << 1) | \
                                         (((bits) & 0x8) << (11 - 3)))
+#define HSW_WB_LLC_AGE3                        HSW_CACHEABILITY_CONTROL(0x2)
 #define HSW_WB_LLC_AGE0                        HSW_CACHEABILITY_CONTROL(0x3)
 #define HSW_WB_ELLC_LLC_AGE0           HSW_CACHEABILITY_CONTROL(0xb)
 
@@ -105,7 +106,7 @@ static gen6_gtt_pte_t hsw_pte_encode(dma_addr_t addr,
        pte |= HSW_PTE_ADDR_ENCODE(addr);
 
        if (level != I915_CACHE_NONE)
-               pte |= HSW_WB_LLC_AGE0;
+               pte |= HSW_WB_LLC_AGE3;
 
        return pte;
 }
--

1.8.3.4

Quoted 4 lines

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

--
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch

#5 lore raw
Quoted 26 lines · Daniel Vetter
Quoted 11 lines · Ben Widawsky

The default LLC age was changed:
commit 0d8ff15e9a15f2b393e53337a107b7a1e5919b6d

Author: Ben Widawsky <[email protected]>
Date:   Thu Jul 4 11:02:03 2013 -0700

drm/i915/hsw: Set correct Haswell PTE encodings.

This caused a regression in performance on certain benchmarks. While I think a discussion still needs to happen about how the kernel should default for both eLLC, and LLC - just revert this behavior for now.

Quoted 7 lines

v2: Drop the extra #define (Chad)

v3: Actually git add

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=67062

iris_pte_encode was split up after the regressing commit in

commit 4d15c145a6234d999c0452eec0d275c1fbf0688c

Author: Ben Widawsky <[email protected]>
Date:   Thu Jul 4 11:02:06 2013 -0700
Quoted 8 lines

drm/i915: Use eLLC/LLC by default when available

Hence I think that one should be patched, too (QA doesn't have an iris machine afaik, so no regression report expected). And since I have some open questions about Chris WT-for-Iris patch too I think it's best to spin a v4 of this one here.
-Daniel

We've come up with a theory as to why this was a regression (which I think nullifies your request for the eLLC version)

On the surface, one would think that aging all objects equally, whether they be oldest, or youngest, makes no difference. The key with LLC is that it is shared with the CPU, and presumably the CPU follows the normal cache aging FIFO rules (ie. everything is aged 3). On this logic, eLLC would not share the same flaw. I think it makes sense to have someone prove we need it for eLLC, as this does allow mesa/DDX to get 2 differently aged eLLC objects, which they cannot do if we default to 3. On the same note, even LLC age 0 is still ideal in my opinion, but was asked to not do this so that meas 9.2 doesn't have performance regressions.

[snip]

Signature · 1 line
Ben Widawsky, Intel Open Source Technology Center
#6 lore raw
Quoted 39 lines · Ben Widawsky
Quoted 26 lines · Daniel Vetter
Quoted 11 lines · Ben Widawsky

The default LLC age was changed:
commit 0d8ff15e9a15f2b393e53337a107b7a1e5919b6d

Author: Ben Widawsky <[email protected]>
Date:   Thu Jul 4 11:02:03 2013 -0700

drm/i915/hsw: Set correct Haswell PTE encodings.

This caused a regression in performance on certain benchmarks. While I think a discussion still needs to happen about how the kernel should default for both eLLC, and LLC - just revert this behavior for now.

Quoted 7 lines

v2: Drop the extra #define (Chad)

v3: Actually git add

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=67062

iris_pte_encode was split up after the regressing commit in

commit 4d15c145a6234d999c0452eec0d275c1fbf0688c

Author: Ben Widawsky <[email protected]>
Date:   Thu Jul 4 11:02:06 2013 -0700
Quoted 8 lines

drm/i915: Use eLLC/LLC by default when available

Hence I think that one should be patched, too (QA doesn't have an iris machine afaik, so no regression report expected). And since I have some open questions about Chris WT-for-Iris patch too I think it's best to spin a v4 of this one here.
-Daniel

We've come up with a theory as to why this was a regression (which I think nullifies your request for the eLLC version)

Quoted 10 lines

On the surface, one would think that aging all objects equally, whether they be oldest, or youngest, makes no difference. The key with LLC is that it is shared with the CPU, and presumably the CPU follows the normal cache aging FIFO rules (ie. everything is aged 3). On this logic, eLLC would not share the same flaw. I think it makes sense to have someone prove we need it for eLLC, as this does allow mesa/DDX to get 2 differently aged eLLC objects, which they cannot do if we default to 3. On the same note, even LLC age 0 is still ideal in my opinion, but was asked to not do this so that meas 9.2 doesn't have performance regressions.

Well we have a get/set_caching ioctl, explictly with room for stuff like aging. So this can be added in any case. But yeah if someone can check whether we have the same regression on iris or not would be good. Can you please poke QA (or whoever has one of these boxen)? -Daniel
--
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch

#7 lore raw
Quoted 51 lines · Daniel Vetter
Quoted 39 lines · Ben Widawsky
Quoted 26 lines · Daniel Vetter
Quoted 11 lines · Ben Widawsky

The default LLC age was changed:
commit 0d8ff15e9a15f2b393e53337a107b7a1e5919b6d

Author: Ben Widawsky <[email protected]>
Date:   Thu Jul 4 11:02:03 2013 -0700

drm/i915/hsw: Set correct Haswell PTE encodings.

This caused a regression in performance on certain benchmarks. While I think a discussion still needs to happen about how the kernel should default for both eLLC, and LLC - just revert this behavior for now.

Quoted 7 lines

v2: Drop the extra #define (Chad)

v3: Actually git add

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=67062

iris_pte_encode was split up after the regressing commit in

commit 4d15c145a6234d999c0452eec0d275c1fbf0688c

Author: Ben Widawsky <[email protected]>
Date:   Thu Jul 4 11:02:06 2013 -0700
Quoted 8 lines

drm/i915: Use eLLC/LLC by default when available

Hence I think that one should be patched, too (QA doesn't have an iris machine afaik, so no regression report expected). And since I have some open questions about Chris WT-for-Iris patch too I think it's best to spin a v4 of this one here.
-Daniel

We've come up with a theory as to why this was a regression (which I think nullifies your request for the eLLC version)

Quoted 10 lines

On the surface, one would think that aging all objects equally, whether they be oldest, or youngest, makes no difference. The key with LLC is that it is shared with the CPU, and presumably the CPU follows the normal cache aging FIFO rules (ie. everything is aged 3). On this logic, eLLC would not share the same flaw. I think it makes sense to have someone prove we need it for eLLC, as this does allow mesa/DDX to get 2 differently aged eLLC objects, which they cannot do if we default to 3. On the same note, even LLC age 0 is still ideal in my opinion, but was asked to not do this so that meas 9.2 doesn't have performance regressions.

Quoted 2 lines

Well we have a get/set_caching ioctl, explictly with room for stuff like aging.

It's like someone wrote such patches almost a year ago, and they were rejected as overly complicated by some maintainer.

Quoted 4 lines

So this can be added in any case. But yeah if someone can check whether we have the same regression on iris or not would be good. Can you please poke QA (or whoever has one of these boxen)? -Daniel

Well, forgetting regressions for just one sec, what do you think makes the most sense? If you think it should be 3, it's probably not even worth the effort since I barely care - and I'm happy to just change it..

Signature · 1 line
Ben Widawsky, Intel Open Source Technology Center
#8 lore raw
Quoted 6 lines · Ben Widawsky
Quoted 2 lines · Daniel Vetter

Well we have a get/set_caching ioctl, explictly with room for stuff like aging.

It's like someone wrote such patches almost a year ago, and they were rejected as overly complicated by some maintainer.

Rejected as "needs to be justified with performance data and a mesa/userspace use-case", not "rejected as too complicated, period".

Quoted 8 lines
Quoted 4 lines

So this can be added in any case. But yeah if someone can check whether we have the same regression on iris or not would be good. Can you please poke QA (or whoever has one of these boxen)? -Daniel

Well, forgetting regressions for just one sec, what do you think makes the most sense? If you think it should be 3, it's probably not even worth the effort since I barely care - and I'm happy to just change it..

Since the iris split happened after the regression and we've just argued that age3 seems to be the right default (as long as mesa doesn't do anything special) I think we should argue for age3 on iris systems, too. Imo argueing against simply propagating the copy&paste change needs performance data.
-Daniel
--
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch