AuroraMiddleware/skia2
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity
9e94812bef
skia2/tests/sksl/blend/golden/BlendSaturation.asm.frag

503 lines
17 KiB
GLSL
Raw Normal View History

Enable SPIR-V disassembly output from skslc. We now have SPIR-V golden outputs for `blend` and `shared` tests. This exposes a handful of SPIR-V limitations for us to address. Change-Id: Ie5278889b8a61432403d06231b17765885bee0ac Bug: skia:10694 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337182 Commit-Queue: John Stiles <johnstiles@google.com> Commit-Queue: Brian Osman <brianosman@google.com> Auto-Submit: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com>
2020-11-20 21:28:50 +00:00
OpCapability Shader
%1 = OpExtInstImport "GLSL.std.450"
OpMemoryModel Logical GLSL450
OpEntryPoint Fragment %main "main" %sk_FragColor %sk_Clockwise %src %dst
OpExecutionMode %main OriginUpperLeft
OpName %sk_FragColor "sk_FragColor"
OpName %sk_Clockwise "sk_Clockwise"
OpName %src "src"
OpName %dst "dst"
Avoid inlining functions that are called repeatedly. Previously, we'd gauge suitability for inlining by counting the nodes in a function; past a certain limit, the function was considered "too big." Now, we also incorporate the number of times that function is called. So if a function is called three times, and its size is 20 nodes, it would be considered to have an inlining cost of 60 (3 * 20) instead of 20. This should tamp down the aggressive nature of the inliner in cases like gaussian convolution or complicated blends, and will hopefully satisfy Pinpoint. No change visible in Nanobench (which doesn't test any of these sorts of patterns, but certainly inlines things): http://screen/AwD5hkgkEfjVx4g Change-Id: Ie5e32898245ac854adb9ddd52d87001df6a67125 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337676 Commit-Queue: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com> Auto-Submit: John Stiles <johnstiles@google.com>
2020-11-23 19:48:06 +00:00
OpName %_blend_set_color_saturation_helper "_blend_set_color_saturation_helper"
Enable SPIR-V disassembly output from skslc. We now have SPIR-V golden outputs for `blend` and `shared` tests. This exposes a handful of SPIR-V limitations for us to address. Change-Id: Ie5278889b8a61432403d06231b17765885bee0ac Bug: skia:10694 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337182 Commit-Queue: John Stiles <johnstiles@google.com> Commit-Queue: Brian Osman <brianosman@google.com> Auto-Submit: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com>
2020-11-20 21:28:50 +00:00
OpName %main "main"
Revert "Eliminate inliner temporary variables for functions with a single exit." This reverts commit 345d72124d35b7c434be7bd9b6b16a66fbd980e6. Reason for revert: can break ES2 for-loop rules Original change's description: > Eliminate inliner temporary variables for functions with a single exit. > > When we determine that a function only contains a single return > statement, there is no need to create a temporary variable and store the > result expression into a variable. Instead, we can directly replace the > function-call expression with the return-statement's expression. > > This dramatically simplifies the final optimized output from chains of > very simple inlined functions, which is a very common pattern for trees > of Skia fragment processors. > > Change-Id: I6789064a321daf43db2e1cef4915f25ed74d6131 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/344665 > Commit-Queue: John Stiles <johnstiles@google.com> > Reviewed-by: Brian Osman <brianosman@google.com> > Auto-Submit: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,ethannicholas@google.com,johnstiles@google.com Change-Id: I60845f22159605a06047b030e2686a769121a35a No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://skia-review.googlesource.com/c/skia/+/345120 Reviewed-by: John Stiles <johnstiles@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-12-16 18:24:48 +00:00
OpName %_0_blend_saturation "_0_blend_saturation"
Enable SPIR-V disassembly output from skslc. We now have SPIR-V golden outputs for `blend` and `shared` tests. This exposes a handful of SPIR-V limitations for us to address. Change-Id: Ie5278889b8a61432403d06231b17765885bee0ac Bug: skia:10694 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337182 Commit-Queue: John Stiles <johnstiles@google.com> Commit-Queue: Brian Osman <brianosman@google.com> Auto-Submit: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com>
2020-11-20 21:28:50 +00:00
OpName %_1_alpha "_1_alpha"
OpName %_2_sda "_2_sda"
OpName %_3_dsa "_3_dsa"
OpName %_4_blend_set_color_saturation "_4_blend_set_color_saturation"
Revert "Eliminate inliner temporary variables for functions with a single exit." This reverts commit 345d72124d35b7c434be7bd9b6b16a66fbd980e6. Reason for revert: can break ES2 for-loop rules Original change's description: > Eliminate inliner temporary variables for functions with a single exit. > > When we determine that a function only contains a single return > statement, there is no need to create a temporary variable and store the > result expression into a variable. Instead, we can directly replace the > function-call expression with the return-statement's expression. > > This dramatically simplifies the final optimized output from chains of > very simple inlined functions, which is a very common pattern for trees > of Skia fragment processors. > > Change-Id: I6789064a321daf43db2e1cef4915f25ed74d6131 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/344665 > Commit-Queue: John Stiles <johnstiles@google.com> > Reviewed-by: Brian Osman <brianosman@google.com> > Auto-Submit: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,ethannicholas@google.com,johnstiles@google.com Change-Id: I60845f22159605a06047b030e2686a769121a35a No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://skia-review.googlesource.com/c/skia/+/345120 Reviewed-by: John Stiles <johnstiles@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-12-16 18:24:48 +00:00
OpName %_5_13_blend_color_saturation "_5_13_blend_color_saturation"
OpName %_6_sat "_6_sat"
OpName %_7_blend_set_color_luminance "_7_blend_set_color_luminance"
OpName %_8_11_blend_color_luminance "_8_11_blend_color_luminance"
OpName %_9_lum "_9_lum"
OpName %_10_12_blend_color_luminance "_10_12_blend_color_luminance"
OpName %_11_result "_11_result"
OpName %_12_minComp "_12_minComp"
OpName %_13_maxComp "_13_maxComp"
Enable SPIR-V disassembly output from skslc. We now have SPIR-V golden outputs for `blend` and `shared` tests. This exposes a handful of SPIR-V limitations for us to address. Change-Id: Ie5278889b8a61432403d06231b17765885bee0ac Bug: skia:10694 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337182 Commit-Queue: John Stiles <johnstiles@google.com> Commit-Queue: Brian Osman <brianosman@google.com> Auto-Submit: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com>
2020-11-20 21:28:50 +00:00
OpDecorate %sk_FragColor RelaxedPrecision
OpDecorate %sk_FragColor Location 0
OpDecorate %sk_FragColor Index 0
OpDecorate %sk_Clockwise RelaxedPrecision
OpDecorate %sk_Clockwise BuiltIn FrontFacing
OpDecorate %src RelaxedPrecision
OpDecorate %dst RelaxedPrecision
Avoid inlining functions that are called repeatedly. Previously, we'd gauge suitability for inlining by counting the nodes in a function; past a certain limit, the function was considered "too big." Now, we also incorporate the number of times that function is called. So if a function is called three times, and its size is 20 nodes, it would be considered to have an inlining cost of 60 (3 * 20) instead of 20. This should tamp down the aggressive nature of the inliner in cases like gaussian convolution or complicated blends, and will hopefully satisfy Pinpoint. No change visible in Nanobench (which doesn't test any of these sorts of patterns, but certainly inlines things): http://screen/AwD5hkgkEfjVx4g Change-Id: Ie5e32898245ac854adb9ddd52d87001df6a67125 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337676 Commit-Queue: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com> Auto-Submit: John Stiles <johnstiles@google.com>
2020-11-23 19:48:06 +00:00
OpDecorate %21 RelaxedPrecision
OpDecorate %23 RelaxedPrecision
OpDecorate %31 RelaxedPrecision
OpDecorate %32 RelaxedPrecision
Enable SPIR-V disassembly output from skslc. We now have SPIR-V golden outputs for `blend` and `shared` tests. This exposes a handful of SPIR-V limitations for us to address. Change-Id: Ie5278889b8a61432403d06231b17765885bee0ac Bug: skia:10694 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337182 Commit-Queue: John Stiles <johnstiles@google.com> Commit-Queue: Brian Osman <brianosman@google.com> Auto-Submit: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com>
2020-11-20 21:28:50 +00:00
OpDecorate %34 RelaxedPrecision
OpDecorate %36 RelaxedPrecision
Avoid inlining functions that are called repeatedly. Previously, we'd gauge suitability for inlining by counting the nodes in a function; past a certain limit, the function was considered "too big." Now, we also incorporate the number of times that function is called. So if a function is called three times, and its size is 20 nodes, it would be considered to have an inlining cost of 60 (3 * 20) instead of 20. This should tamp down the aggressive nature of the inliner in cases like gaussian convolution or complicated blends, and will hopefully satisfy Pinpoint. No change visible in Nanobench (which doesn't test any of these sorts of patterns, but certainly inlines things): http://screen/AwD5hkgkEfjVx4g Change-Id: Ie5e32898245ac854adb9ddd52d87001df6a67125 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337676 Commit-Queue: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com> Auto-Submit: John Stiles <johnstiles@google.com>
2020-11-23 19:48:06 +00:00
OpDecorate %37 RelaxedPrecision
OpDecorate %38 RelaxedPrecision
OpDecorate %40 RelaxedPrecision
OpDecorate %42 RelaxedPrecision
OpDecorate %43 RelaxedPrecision
OpDecorate %44 RelaxedPrecision
Enable SPIR-V disassembly output from skslc. We now have SPIR-V golden outputs for `blend` and `shared` tests. This exposes a handful of SPIR-V limitations for us to address. Change-Id: Ie5278889b8a61432403d06231b17765885bee0ac Bug: skia:10694 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337182 Commit-Queue: John Stiles <johnstiles@google.com> Commit-Queue: Brian Osman <brianosman@google.com> Auto-Submit: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com>
2020-11-20 21:28:50 +00:00
OpDecorate %47 RelaxedPrecision
Avoid inlining functions that are called repeatedly. Previously, we'd gauge suitability for inlining by counting the nodes in a function; past a certain limit, the function was considered "too big." Now, we also incorporate the number of times that function is called. So if a function is called three times, and its size is 20 nodes, it would be considered to have an inlining cost of 60 (3 * 20) instead of 20. This should tamp down the aggressive nature of the inliner in cases like gaussian convolution or complicated blends, and will hopefully satisfy Pinpoint. No change visible in Nanobench (which doesn't test any of these sorts of patterns, but certainly inlines things): http://screen/AwD5hkgkEfjVx4g Change-Id: Ie5e32898245ac854adb9ddd52d87001df6a67125 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337676 Commit-Queue: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com> Auto-Submit: John Stiles <johnstiles@google.com>
2020-11-23 19:48:06 +00:00
OpDecorate %56 RelaxedPrecision
OpDecorate %58 RelaxedPrecision
Revert "Eliminate inliner temporary variables for functions with a single exit." This reverts commit 345d72124d35b7c434be7bd9b6b16a66fbd980e6. Reason for revert: can break ES2 for-loop rules Original change's description: > Eliminate inliner temporary variables for functions with a single exit. > > When we determine that a function only contains a single return > statement, there is no need to create a temporary variable and store the > result expression into a variable. Instead, we can directly replace the > function-call expression with the return-statement's expression. > > This dramatically simplifies the final optimized output from chains of > very simple inlined functions, which is a very common pattern for trees > of Skia fragment processors. > > Change-Id: I6789064a321daf43db2e1cef4915f25ed74d6131 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/344665 > Commit-Queue: John Stiles <johnstiles@google.com> > Reviewed-by: Brian Osman <brianosman@google.com> > Auto-Submit: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,ethannicholas@google.com,johnstiles@google.com Change-Id: I60845f22159605a06047b030e2686a769121a35a No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://skia-review.googlesource.com/c/skia/+/345120 Reviewed-by: John Stiles <johnstiles@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-12-16 18:24:48 +00:00
OpDecorate %60 RelaxedPrecision
Declare all inlined variables at the topmost scope possible. By itself, this is uninteresting and even perhaps slightly counterproductive (as it separates vardecl from its initializer, increasing LOC). However, this enables a followup CL (http://review.skia.org/344665) which allows single-return functions to be inlined without the creation of a temporary variable at all. This applies to the majority of fragment processors in a typical Ganesh hierarchy. This change will greatly reduce the number of inliner-created temporary copies when compiling a typical tree of FPs. Change-Id: I03423a13cf35050637dabace4a32973a08a4ed0a Reviewed-on: https://skia-review.googlesource.com/c/skia/+/344764 Reviewed-by: Brian Osman <brianosman@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-12-15 22:50:21 +00:00
OpDecorate %61 RelaxedPrecision
Revert "Eliminate inliner temporary variables for functions with a single exit." This reverts commit 345d72124d35b7c434be7bd9b6b16a66fbd980e6. Reason for revert: can break ES2 for-loop rules Original change's description: > Eliminate inliner temporary variables for functions with a single exit. > > When we determine that a function only contains a single return > statement, there is no need to create a temporary variable and store the > result expression into a variable. Instead, we can directly replace the > function-call expression with the return-statement's expression. > > This dramatically simplifies the final optimized output from chains of > very simple inlined functions, which is a very common pattern for trees > of Skia fragment processors. > > Change-Id: I6789064a321daf43db2e1cef4915f25ed74d6131 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/344665 > Commit-Queue: John Stiles <johnstiles@google.com> > Reviewed-by: Brian Osman <brianosman@google.com> > Auto-Submit: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,ethannicholas@google.com,johnstiles@google.com Change-Id: I60845f22159605a06047b030e2686a769121a35a No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://skia-review.googlesource.com/c/skia/+/345120 Reviewed-by: John Stiles <johnstiles@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-12-16 18:24:48 +00:00
OpDecorate %63 RelaxedPrecision
Avoid inlining functions that are called repeatedly. Previously, we'd gauge suitability for inlining by counting the nodes in a function; past a certain limit, the function was considered "too big." Now, we also incorporate the number of times that function is called. So if a function is called three times, and its size is 20 nodes, it would be considered to have an inlining cost of 60 (3 * 20) instead of 20. This should tamp down the aggressive nature of the inliner in cases like gaussian convolution or complicated blends, and will hopefully satisfy Pinpoint. No change visible in Nanobench (which doesn't test any of these sorts of patterns, but certainly inlines things): http://screen/AwD5hkgkEfjVx4g Change-Id: Ie5e32898245ac854adb9ddd52d87001df6a67125 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337676 Commit-Queue: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com> Auto-Submit: John Stiles <johnstiles@google.com>
2020-11-23 19:48:06 +00:00
OpDecorate %66 RelaxedPrecision
Revert "Eliminate inliner temporary variables for functions with a single exit." This reverts commit 345d72124d35b7c434be7bd9b6b16a66fbd980e6. Reason for revert: can break ES2 for-loop rules Original change's description: > Eliminate inliner temporary variables for functions with a single exit. > > When we determine that a function only contains a single return > statement, there is no need to create a temporary variable and store the > result expression into a variable. Instead, we can directly replace the > function-call expression with the return-statement's expression. > > This dramatically simplifies the final optimized output from chains of > very simple inlined functions, which is a very common pattern for trees > of Skia fragment processors. > > Change-Id: I6789064a321daf43db2e1cef4915f25ed74d6131 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/344665 > Commit-Queue: John Stiles <johnstiles@google.com> > Reviewed-by: Brian Osman <brianosman@google.com> > Auto-Submit: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,ethannicholas@google.com,johnstiles@google.com Change-Id: I60845f22159605a06047b030e2686a769121a35a No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://skia-review.googlesource.com/c/skia/+/345120 Reviewed-by: John Stiles <johnstiles@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-12-16 18:24:48 +00:00
OpDecorate %68 RelaxedPrecision
OpDecorate %76 RelaxedPrecision
OpDecorate %78 RelaxedPrecision
OpDecorate %80 RelaxedPrecision
OpDecorate %84 RelaxedPrecision
OpDecorate %86 RelaxedPrecision
Declare all inlined variables at the topmost scope possible. By itself, this is uninteresting and even perhaps slightly counterproductive (as it separates vardecl from its initializer, increasing LOC). However, this enables a followup CL (http://review.skia.org/344665) which allows single-return functions to be inlined without the creation of a temporary variable at all. This applies to the majority of fragment processors in a typical Ganesh hierarchy. This change will greatly reduce the number of inliner-created temporary copies when compiling a typical tree of FPs. Change-Id: I03423a13cf35050637dabace4a32973a08a4ed0a Reviewed-on: https://skia-review.googlesource.com/c/skia/+/344764 Reviewed-by: Brian Osman <brianosman@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-12-15 22:50:21 +00:00
OpDecorate %88 RelaxedPrecision
OpDecorate %90 RelaxedPrecision
Revert "Eliminate inliner temporary variables for functions with a single exit." This reverts commit 345d72124d35b7c434be7bd9b6b16a66fbd980e6. Reason for revert: can break ES2 for-loop rules Original change's description: > Eliminate inliner temporary variables for functions with a single exit. > > When we determine that a function only contains a single return > statement, there is no need to create a temporary variable and store the > result expression into a variable. Instead, we can directly replace the > function-call expression with the return-statement's expression. > > This dramatically simplifies the final optimized output from chains of > very simple inlined functions, which is a very common pattern for trees > of Skia fragment processors. > > Change-Id: I6789064a321daf43db2e1cef4915f25ed74d6131 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/344665 > Commit-Queue: John Stiles <johnstiles@google.com> > Reviewed-by: Brian Osman <brianosman@google.com> > Auto-Submit: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,ethannicholas@google.com,johnstiles@google.com Change-Id: I60845f22159605a06047b030e2686a769121a35a No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://skia-review.googlesource.com/c/skia/+/345120 Reviewed-by: John Stiles <johnstiles@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-12-16 18:24:48 +00:00
OpDecorate %91 RelaxedPrecision
OpDecorate %92 RelaxedPrecision
OpDecorate %94 RelaxedPrecision
OpDecorate %100 RelaxedPrecision
OpDecorate %102 RelaxedPrecision
OpDecorate %108 RelaxedPrecision
OpDecorate %110 RelaxedPrecision
OpDecorate %113 RelaxedPrecision
OpDecorate %115 RelaxedPrecision
OpDecorate %121 RelaxedPrecision
Reland "Simplify _blend_set_color_saturation, removing an instruction." This reverts commit e81fb87bb444c7f1dd087479b0a868f365a465ae. Reason for revert: checking results with less-aggressive inliner Original change's description: > Revert "Simplify _blend_set_color_saturation, removing an instruction." > > This reverts commit ed289e777cfa41c059b94e64e0c749fe47db563b. > > Reason for revert: causing strange artifacts, only on Adreno > > Original change's description: > > Simplify _blend_set_color_saturation, removing an instruction. > > > > This tightens up our intrinsics slightly; after inlining, it eliminates > > one scratch variable. (We no longer need to copy `sda` into `hueColor` > > as hueColor is now unchanged.) > > > > Change-Id: Iece5ba2fe11cde54481704a1787114a2c2a66d9b > > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/336599 > > Commit-Queue: John Stiles <johnstiles@google.com> > > Commit-Queue: Brian Osman <brianosman@google.com> > > Auto-Submit: John Stiles <johnstiles@google.com> > > Reviewed-by: Brian Osman <brianosman@google.com> > > TBR=brianosman@google.com,johnstiles@google.com > > Change-Id: Ica506467b0a4e03d0cbe482034acfa2d9f8d2c16 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337560 > Reviewed-by: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,johnstiles@google.com Change-Id: Ia93263f3269c057e7eaa69ca2b05e783d18c0199 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337944 Auto-Submit: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-11-24 03:19:04 +00:00
OpDecorate %124 RelaxedPrecision
Revert "Eliminate inliner temporary variables for functions with a single exit." This reverts commit 345d72124d35b7c434be7bd9b6b16a66fbd980e6. Reason for revert: can break ES2 for-loop rules Original change's description: > Eliminate inliner temporary variables for functions with a single exit. > > When we determine that a function only contains a single return > statement, there is no need to create a temporary variable and store the > result expression into a variable. Instead, we can directly replace the > function-call expression with the return-statement's expression. > > This dramatically simplifies the final optimized output from chains of > very simple inlined functions, which is a very common pattern for trees > of Skia fragment processors. > > Change-Id: I6789064a321daf43db2e1cef4915f25ed74d6131 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/344665 > Commit-Queue: John Stiles <johnstiles@google.com> > Reviewed-by: Brian Osman <brianosman@google.com> > Auto-Submit: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,ethannicholas@google.com,johnstiles@google.com Change-Id: I60845f22159605a06047b030e2686a769121a35a No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://skia-review.googlesource.com/c/skia/+/345120 Reviewed-by: John Stiles <johnstiles@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-12-16 18:24:48 +00:00
OpDecorate %128 RelaxedPrecision
Reland "Simplify _blend_set_color_saturation, removing an instruction." This reverts commit e81fb87bb444c7f1dd087479b0a868f365a465ae. Reason for revert: checking results with less-aggressive inliner Original change's description: > Revert "Simplify _blend_set_color_saturation, removing an instruction." > > This reverts commit ed289e777cfa41c059b94e64e0c749fe47db563b. > > Reason for revert: causing strange artifacts, only on Adreno > > Original change's description: > > Simplify _blend_set_color_saturation, removing an instruction. > > > > This tightens up our intrinsics slightly; after inlining, it eliminates > > one scratch variable. (We no longer need to copy `sda` into `hueColor` > > as hueColor is now unchanged.) > > > > Change-Id: Iece5ba2fe11cde54481704a1787114a2c2a66d9b > > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/336599 > > Commit-Queue: John Stiles <johnstiles@google.com> > > Commit-Queue: Brian Osman <brianosman@google.com> > > Auto-Submit: John Stiles <johnstiles@google.com> > > Reviewed-by: Brian Osman <brianosman@google.com> > > TBR=brianosman@google.com,johnstiles@google.com > > Change-Id: Ica506467b0a4e03d0cbe482034acfa2d9f8d2c16 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337560 > Reviewed-by: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,johnstiles@google.com Change-Id: Ia93263f3269c057e7eaa69ca2b05e783d18c0199 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337944 Auto-Submit: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-11-24 03:19:04 +00:00
OpDecorate %131 RelaxedPrecision
Revert "Eliminate inliner temporary variables for functions with a single exit." This reverts commit 345d72124d35b7c434be7bd9b6b16a66fbd980e6. Reason for revert: can break ES2 for-loop rules Original change's description: > Eliminate inliner temporary variables for functions with a single exit. > > When we determine that a function only contains a single return > statement, there is no need to create a temporary variable and store the > result expression into a variable. Instead, we can directly replace the > function-call expression with the return-statement's expression. > > This dramatically simplifies the final optimized output from chains of > very simple inlined functions, which is a very common pattern for trees > of Skia fragment processors. > > Change-Id: I6789064a321daf43db2e1cef4915f25ed74d6131 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/344665 > Commit-Queue: John Stiles <johnstiles@google.com> > Reviewed-by: Brian Osman <brianosman@google.com> > Auto-Submit: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,ethannicholas@google.com,johnstiles@google.com Change-Id: I60845f22159605a06047b030e2686a769121a35a No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://skia-review.googlesource.com/c/skia/+/345120 Reviewed-by: John Stiles <johnstiles@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-12-16 18:24:48 +00:00
OpDecorate %135 RelaxedPrecision
OpDecorate %137 RelaxedPrecision
OpDecorate %143 RelaxedPrecision
Reland "Simplify _blend_set_color_saturation, removing an instruction." This reverts commit e81fb87bb444c7f1dd087479b0a868f365a465ae. Reason for revert: checking results with less-aggressive inliner Original change's description: > Revert "Simplify _blend_set_color_saturation, removing an instruction." > > This reverts commit ed289e777cfa41c059b94e64e0c749fe47db563b. > > Reason for revert: causing strange artifacts, only on Adreno > > Original change's description: > > Simplify _blend_set_color_saturation, removing an instruction. > > > > This tightens up our intrinsics slightly; after inlining, it eliminates > > one scratch variable. (We no longer need to copy `sda` into `hueColor` > > as hueColor is now unchanged.) > > > > Change-Id: Iece5ba2fe11cde54481704a1787114a2c2a66d9b > > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/336599 > > Commit-Queue: John Stiles <johnstiles@google.com> > > Commit-Queue: Brian Osman <brianosman@google.com> > > Auto-Submit: John Stiles <johnstiles@google.com> > > Reviewed-by: Brian Osman <brianosman@google.com> > > TBR=brianosman@google.com,johnstiles@google.com > > Change-Id: Ica506467b0a4e03d0cbe482034acfa2d9f8d2c16 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337560 > Reviewed-by: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,johnstiles@google.com Change-Id: Ia93263f3269c057e7eaa69ca2b05e783d18c0199 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337944 Auto-Submit: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-11-24 03:19:04 +00:00
OpDecorate %146 RelaxedPrecision
Revert "Eliminate inliner temporary variables for functions with a single exit." This reverts commit 345d72124d35b7c434be7bd9b6b16a66fbd980e6. Reason for revert: can break ES2 for-loop rules Original change's description: > Eliminate inliner temporary variables for functions with a single exit. > > When we determine that a function only contains a single return > statement, there is no need to create a temporary variable and store the > result expression into a variable. Instead, we can directly replace the > function-call expression with the return-statement's expression. > > This dramatically simplifies the final optimized output from chains of > very simple inlined functions, which is a very common pattern for trees > of Skia fragment processors. > > Change-Id: I6789064a321daf43db2e1cef4915f25ed74d6131 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/344665 > Commit-Queue: John Stiles <johnstiles@google.com> > Reviewed-by: Brian Osman <brianosman@google.com> > Auto-Submit: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,ethannicholas@google.com,johnstiles@google.com Change-Id: I60845f22159605a06047b030e2686a769121a35a No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://skia-review.googlesource.com/c/skia/+/345120 Reviewed-by: John Stiles <johnstiles@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-12-16 18:24:48 +00:00
OpDecorate %150 RelaxedPrecision
OpDecorate %152 RelaxedPrecision
OpDecorate %158 RelaxedPrecision
Reland "Simplify _blend_set_color_saturation, removing an instruction." This reverts commit e81fb87bb444c7f1dd087479b0a868f365a465ae. Reason for revert: checking results with less-aggressive inliner Original change's description: > Revert "Simplify _blend_set_color_saturation, removing an instruction." > > This reverts commit ed289e777cfa41c059b94e64e0c749fe47db563b. > > Reason for revert: causing strange artifacts, only on Adreno > > Original change's description: > > Simplify _blend_set_color_saturation, removing an instruction. > > > > This tightens up our intrinsics slightly; after inlining, it eliminates > > one scratch variable. (We no longer need to copy `sda` into `hueColor` > > as hueColor is now unchanged.) > > > > Change-Id: Iece5ba2fe11cde54481704a1787114a2c2a66d9b > > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/336599 > > Commit-Queue: John Stiles <johnstiles@google.com> > > Commit-Queue: Brian Osman <brianosman@google.com> > > Auto-Submit: John Stiles <johnstiles@google.com> > > Reviewed-by: Brian Osman <brianosman@google.com> > > TBR=brianosman@google.com,johnstiles@google.com > > Change-Id: Ica506467b0a4e03d0cbe482034acfa2d9f8d2c16 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337560 > Reviewed-by: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,johnstiles@google.com Change-Id: Ia93263f3269c057e7eaa69ca2b05e783d18c0199 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337944 Auto-Submit: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-11-24 03:19:04 +00:00
OpDecorate %161 RelaxedPrecision
Revert "Eliminate inliner temporary variables for functions with a single exit." This reverts commit 345d72124d35b7c434be7bd9b6b16a66fbd980e6. Reason for revert: can break ES2 for-loop rules Original change's description: > Eliminate inliner temporary variables for functions with a single exit. > > When we determine that a function only contains a single return > statement, there is no need to create a temporary variable and store the > result expression into a variable. Instead, we can directly replace the > function-call expression with the return-statement's expression. > > This dramatically simplifies the final optimized output from chains of > very simple inlined functions, which is a very common pattern for trees > of Skia fragment processors. > > Change-Id: I6789064a321daf43db2e1cef4915f25ed74d6131 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/344665 > Commit-Queue: John Stiles <johnstiles@google.com> > Reviewed-by: Brian Osman <brianosman@google.com> > Auto-Submit: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,ethannicholas@google.com,johnstiles@google.com Change-Id: I60845f22159605a06047b030e2686a769121a35a No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://skia-review.googlesource.com/c/skia/+/345120 Reviewed-by: John Stiles <johnstiles@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-12-16 18:24:48 +00:00
OpDecorate %165 RelaxedPrecision
OpDecorate %168 RelaxedPrecision
OpDecorate %184 RelaxedPrecision
OpDecorate %185 RelaxedPrecision
Reland "Simplify _blend_set_color_saturation, removing an instruction." This reverts commit e81fb87bb444c7f1dd087479b0a868f365a465ae. Reason for revert: checking results with less-aggressive inliner Original change's description: > Revert "Simplify _blend_set_color_saturation, removing an instruction." > > This reverts commit ed289e777cfa41c059b94e64e0c749fe47db563b. > > Reason for revert: causing strange artifacts, only on Adreno > > Original change's description: > > Simplify _blend_set_color_saturation, removing an instruction. > > > > This tightens up our intrinsics slightly; after inlining, it eliminates > > one scratch variable. (We no longer need to copy `sda` into `hueColor` > > as hueColor is now unchanged.) > > > > Change-Id: Iece5ba2fe11cde54481704a1787114a2c2a66d9b > > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/336599 > > Commit-Queue: John Stiles <johnstiles@google.com> > > Commit-Queue: Brian Osman <brianosman@google.com> > > Auto-Submit: John Stiles <johnstiles@google.com> > > Reviewed-by: Brian Osman <brianosman@google.com> > > TBR=brianosman@google.com,johnstiles@google.com > > Change-Id: Ica506467b0a4e03d0cbe482034acfa2d9f8d2c16 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337560 > Reviewed-by: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,johnstiles@google.com Change-Id: Ia93263f3269c057e7eaa69ca2b05e783d18c0199 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337944 Auto-Submit: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-11-24 03:19:04 +00:00
OpDecorate %188 RelaxedPrecision
Revert "Eliminate inliner temporary variables for functions with a single exit." This reverts commit 345d72124d35b7c434be7bd9b6b16a66fbd980e6. Reason for revert: can break ES2 for-loop rules Original change's description: > Eliminate inliner temporary variables for functions with a single exit. > > When we determine that a function only contains a single return > statement, there is no need to create a temporary variable and store the > result expression into a variable. Instead, we can directly replace the > function-call expression with the return-statement's expression. > > This dramatically simplifies the final optimized output from chains of > very simple inlined functions, which is a very common pattern for trees > of Skia fragment processors. > > Change-Id: I6789064a321daf43db2e1cef4915f25ed74d6131 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/344665 > Commit-Queue: John Stiles <johnstiles@google.com> > Reviewed-by: Brian Osman <brianosman@google.com> > Auto-Submit: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,ethannicholas@google.com,johnstiles@google.com Change-Id: I60845f22159605a06047b030e2686a769121a35a No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://skia-review.googlesource.com/c/skia/+/345120 Reviewed-by: John Stiles <johnstiles@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-12-16 18:24:48 +00:00
OpDecorate %189 RelaxedPrecision
Reland "Simplify _blend_set_color_saturation, removing an instruction." This reverts commit e81fb87bb444c7f1dd087479b0a868f365a465ae. Reason for revert: checking results with less-aggressive inliner Original change's description: > Revert "Simplify _blend_set_color_saturation, removing an instruction." > > This reverts commit ed289e777cfa41c059b94e64e0c749fe47db563b. > > Reason for revert: causing strange artifacts, only on Adreno > > Original change's description: > > Simplify _blend_set_color_saturation, removing an instruction. > > > > This tightens up our intrinsics slightly; after inlining, it eliminates > > one scratch variable. (We no longer need to copy `sda` into `hueColor` > > as hueColor is now unchanged.) > > > > Change-Id: Iece5ba2fe11cde54481704a1787114a2c2a66d9b > > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/336599 > > Commit-Queue: John Stiles <johnstiles@google.com> > > Commit-Queue: Brian Osman <brianosman@google.com> > > Auto-Submit: John Stiles <johnstiles@google.com> > > Reviewed-by: Brian Osman <brianosman@google.com> > > TBR=brianosman@google.com,johnstiles@google.com > > Change-Id: Ica506467b0a4e03d0cbe482034acfa2d9f8d2c16 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337560 > Reviewed-by: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,johnstiles@google.com Change-Id: Ia93263f3269c057e7eaa69ca2b05e783d18c0199 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337944 Auto-Submit: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-11-24 03:19:04 +00:00
OpDecorate %190 RelaxedPrecision
Revert "Eliminate inliner temporary variables for functions with a single exit." This reverts commit 345d72124d35b7c434be7bd9b6b16a66fbd980e6. Reason for revert: can break ES2 for-loop rules Original change's description: > Eliminate inliner temporary variables for functions with a single exit. > > When we determine that a function only contains a single return > statement, there is no need to create a temporary variable and store the > result expression into a variable. Instead, we can directly replace the > function-call expression with the return-statement's expression. > > This dramatically simplifies the final optimized output from chains of > very simple inlined functions, which is a very common pattern for trees > of Skia fragment processors. > > Change-Id: I6789064a321daf43db2e1cef4915f25ed74d6131 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/344665 > Commit-Queue: John Stiles <johnstiles@google.com> > Reviewed-by: Brian Osman <brianosman@google.com> > Auto-Submit: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,ethannicholas@google.com,johnstiles@google.com Change-Id: I60845f22159605a06047b030e2686a769121a35a No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://skia-review.googlesource.com/c/skia/+/345120 Reviewed-by: John Stiles <johnstiles@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-12-16 18:24:48 +00:00
OpDecorate %191 RelaxedPrecision
Declare all inlined variables at the topmost scope possible. By itself, this is uninteresting and even perhaps slightly counterproductive (as it separates vardecl from its initializer, increasing LOC). However, this enables a followup CL (http://review.skia.org/344665) which allows single-return functions to be inlined without the creation of a temporary variable at all. This applies to the majority of fragment processors in a typical Ganesh hierarchy. This change will greatly reduce the number of inliner-created temporary copies when compiling a typical tree of FPs. Change-Id: I03423a13cf35050637dabace4a32973a08a4ed0a Reviewed-on: https://skia-review.googlesource.com/c/skia/+/344764 Reviewed-by: Brian Osman <brianosman@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-12-15 22:50:21 +00:00
OpDecorate %192 RelaxedPrecision
Revert "Eliminate inliner temporary variables for functions with a single exit." This reverts commit 345d72124d35b7c434be7bd9b6b16a66fbd980e6. Reason for revert: can break ES2 for-loop rules Original change's description: > Eliminate inliner temporary variables for functions with a single exit. > > When we determine that a function only contains a single return > statement, there is no need to create a temporary variable and store the > result expression into a variable. Instead, we can directly replace the > function-call expression with the return-statement's expression. > > This dramatically simplifies the final optimized output from chains of > very simple inlined functions, which is a very common pattern for trees > of Skia fragment processors. > > Change-Id: I6789064a321daf43db2e1cef4915f25ed74d6131 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/344665 > Commit-Queue: John Stiles <johnstiles@google.com> > Reviewed-by: Brian Osman <brianosman@google.com> > Auto-Submit: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,ethannicholas@google.com,johnstiles@google.com Change-Id: I60845f22159605a06047b030e2686a769121a35a No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://skia-review.googlesource.com/c/skia/+/345120 Reviewed-by: John Stiles <johnstiles@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-12-16 18:24:48 +00:00
OpDecorate %197 RelaxedPrecision
OpDecorate %199 RelaxedPrecision
OpDecorate %201 RelaxedPrecision
OpDecorate %205 RelaxedPrecision
Reland "Simplify _blend_set_color_saturation, removing an instruction." This reverts commit e81fb87bb444c7f1dd087479b0a868f365a465ae. Reason for revert: checking results with less-aggressive inliner Original change's description: > Revert "Simplify _blend_set_color_saturation, removing an instruction." > > This reverts commit ed289e777cfa41c059b94e64e0c749fe47db563b. > > Reason for revert: causing strange artifacts, only on Adreno > > Original change's description: > > Simplify _blend_set_color_saturation, removing an instruction. > > > > This tightens up our intrinsics slightly; after inlining, it eliminates > > one scratch variable. (We no longer need to copy `sda` into `hueColor` > > as hueColor is now unchanged.) > > > > Change-Id: Iece5ba2fe11cde54481704a1787114a2c2a66d9b > > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/336599 > > Commit-Queue: John Stiles <johnstiles@google.com> > > Commit-Queue: Brian Osman <brianosman@google.com> > > Auto-Submit: John Stiles <johnstiles@google.com> > > Reviewed-by: Brian Osman <brianosman@google.com> > > TBR=brianosman@google.com,johnstiles@google.com > > Change-Id: Ica506467b0a4e03d0cbe482034acfa2d9f8d2c16 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337560 > Reviewed-by: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,johnstiles@google.com Change-Id: Ia93263f3269c057e7eaa69ca2b05e783d18c0199 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337944 Auto-Submit: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-11-24 03:19:04 +00:00
OpDecorate %207 RelaxedPrecision
Revert "Eliminate inliner temporary variables for functions with a single exit." This reverts commit 345d72124d35b7c434be7bd9b6b16a66fbd980e6. Reason for revert: can break ES2 for-loop rules Original change's description: > Eliminate inliner temporary variables for functions with a single exit. > > When we determine that a function only contains a single return > statement, there is no need to create a temporary variable and store the > result expression into a variable. Instead, we can directly replace the > function-call expression with the return-statement's expression. > > This dramatically simplifies the final optimized output from chains of > very simple inlined functions, which is a very common pattern for trees > of Skia fragment processors. > > Change-Id: I6789064a321daf43db2e1cef4915f25ed74d6131 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/344665 > Commit-Queue: John Stiles <johnstiles@google.com> > Reviewed-by: Brian Osman <brianosman@google.com> > Auto-Submit: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,ethannicholas@google.com,johnstiles@google.com Change-Id: I60845f22159605a06047b030e2686a769121a35a No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://skia-review.googlesource.com/c/skia/+/345120 Reviewed-by: John Stiles <johnstiles@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-12-16 18:24:48 +00:00
OpDecorate %209 RelaxedPrecision
OpDecorate %212 RelaxedPrecision
OpDecorate %216 RelaxedPrecision
OpDecorate %217 RelaxedPrecision
Avoid inlining functions that are called repeatedly. Previously, we'd gauge suitability for inlining by counting the nodes in a function; past a certain limit, the function was considered "too big." Now, we also incorporate the number of times that function is called. So if a function is called three times, and its size is 20 nodes, it would be considered to have an inlining cost of 60 (3 * 20) instead of 20. This should tamp down the aggressive nature of the inliner in cases like gaussian convolution or complicated blends, and will hopefully satisfy Pinpoint. No change visible in Nanobench (which doesn't test any of these sorts of patterns, but certainly inlines things): http://screen/AwD5hkgkEfjVx4g Change-Id: Ie5e32898245ac854adb9ddd52d87001df6a67125 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337676 Commit-Queue: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com> Auto-Submit: John Stiles <johnstiles@google.com>
2020-11-23 19:48:06 +00:00
OpDecorate %222 RelaxedPrecision
Revert "Eliminate inliner temporary variables for functions with a single exit." This reverts commit 345d72124d35b7c434be7bd9b6b16a66fbd980e6. Reason for revert: can break ES2 for-loop rules Original change's description: > Eliminate inliner temporary variables for functions with a single exit. > > When we determine that a function only contains a single return > statement, there is no need to create a temporary variable and store the > result expression into a variable. Instead, we can directly replace the > function-call expression with the return-statement's expression. > > This dramatically simplifies the final optimized output from chains of > very simple inlined functions, which is a very common pattern for trees > of Skia fragment processors. > > Change-Id: I6789064a321daf43db2e1cef4915f25ed74d6131 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/344665 > Commit-Queue: John Stiles <johnstiles@google.com> > Reviewed-by: Brian Osman <brianosman@google.com> > Auto-Submit: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,ethannicholas@google.com,johnstiles@google.com Change-Id: I60845f22159605a06047b030e2686a769121a35a No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://skia-review.googlesource.com/c/skia/+/345120 Reviewed-by: John Stiles <johnstiles@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-12-16 18:24:48 +00:00
OpDecorate %223 RelaxedPrecision
OpDecorate %224 RelaxedPrecision
OpDecorate %227 RelaxedPrecision
Reland "Simplify _blend_set_color_saturation, removing an instruction." This reverts commit e81fb87bb444c7f1dd087479b0a868f365a465ae. Reason for revert: checking results with less-aggressive inliner Original change's description: > Revert "Simplify _blend_set_color_saturation, removing an instruction." > > This reverts commit ed289e777cfa41c059b94e64e0c749fe47db563b. > > Reason for revert: causing strange artifacts, only on Adreno > > Original change's description: > > Simplify _blend_set_color_saturation, removing an instruction. > > > > This tightens up our intrinsics slightly; after inlining, it eliminates > > one scratch variable. (We no longer need to copy `sda` into `hueColor` > > as hueColor is now unchanged.) > > > > Change-Id: Iece5ba2fe11cde54481704a1787114a2c2a66d9b > > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/336599 > > Commit-Queue: John Stiles <johnstiles@google.com> > > Commit-Queue: Brian Osman <brianosman@google.com> > > Auto-Submit: John Stiles <johnstiles@google.com> > > Reviewed-by: Brian Osman <brianosman@google.com> > > TBR=brianosman@google.com,johnstiles@google.com > > Change-Id: Ica506467b0a4e03d0cbe482034acfa2d9f8d2c16 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337560 > Reviewed-by: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,johnstiles@google.com Change-Id: Ia93263f3269c057e7eaa69ca2b05e783d18c0199 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337944 Auto-Submit: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-11-24 03:19:04 +00:00
OpDecorate %229 RelaxedPrecision
Revert "Eliminate inliner temporary variables for functions with a single exit." This reverts commit 345d72124d35b7c434be7bd9b6b16a66fbd980e6. Reason for revert: can break ES2 for-loop rules Original change's description: > Eliminate inliner temporary variables for functions with a single exit. > > When we determine that a function only contains a single return > statement, there is no need to create a temporary variable and store the > result expression into a variable. Instead, we can directly replace the > function-call expression with the return-statement's expression. > > This dramatically simplifies the final optimized output from chains of > very simple inlined functions, which is a very common pattern for trees > of Skia fragment processors. > > Change-Id: I6789064a321daf43db2e1cef4915f25ed74d6131 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/344665 > Commit-Queue: John Stiles <johnstiles@google.com> > Reviewed-by: Brian Osman <brianosman@google.com> > Auto-Submit: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,ethannicholas@google.com,johnstiles@google.com Change-Id: I60845f22159605a06047b030e2686a769121a35a No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://skia-review.googlesource.com/c/skia/+/345120 Reviewed-by: John Stiles <johnstiles@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-12-16 18:24:48 +00:00
OpDecorate %230 RelaxedPrecision
OpDecorate %231 RelaxedPrecision
OpDecorate %237 RelaxedPrecision
OpDecorate %238 RelaxedPrecision
Reland "Simplify _blend_set_color_saturation, removing an instruction." This reverts commit e81fb87bb444c7f1dd087479b0a868f365a465ae. Reason for revert: checking results with less-aggressive inliner Original change's description: > Revert "Simplify _blend_set_color_saturation, removing an instruction." > > This reverts commit ed289e777cfa41c059b94e64e0c749fe47db563b. > > Reason for revert: causing strange artifacts, only on Adreno > > Original change's description: > > Simplify _blend_set_color_saturation, removing an instruction. > > > > This tightens up our intrinsics slightly; after inlining, it eliminates > > one scratch variable. (We no longer need to copy `sda` into `hueColor` > > as hueColor is now unchanged.) > > > > Change-Id: Iece5ba2fe11cde54481704a1787114a2c2a66d9b > > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/336599 > > Commit-Queue: John Stiles <johnstiles@google.com> > > Commit-Queue: Brian Osman <brianosman@google.com> > > Auto-Submit: John Stiles <johnstiles@google.com> > > Reviewed-by: Brian Osman <brianosman@google.com> > > TBR=brianosman@google.com,johnstiles@google.com > > Change-Id: Ica506467b0a4e03d0cbe482034acfa2d9f8d2c16 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337560 > Reviewed-by: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,johnstiles@google.com Change-Id: Ia93263f3269c057e7eaa69ca2b05e783d18c0199 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337944 Auto-Submit: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-11-24 03:19:04 +00:00
OpDecorate %242 RelaxedPrecision
OpDecorate %243 RelaxedPrecision
OpDecorate %250 RelaxedPrecision
OpDecorate %251 RelaxedPrecision
Revert "Simplify _blend_set_color_saturation, removing an instruction." This reverts commit ed289e777cfa41c059b94e64e0c749fe47db563b. Reason for revert: causing strange artifacts, only on Adreno Original change's description: > Simplify _blend_set_color_saturation, removing an instruction. > > This tightens up our intrinsics slightly; after inlining, it eliminates > one scratch variable. (We no longer need to copy `sda` into `hueColor` > as hueColor is now unchanged.) > > Change-Id: Iece5ba2fe11cde54481704a1787114a2c2a66d9b > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/336599 > Commit-Queue: John Stiles <johnstiles@google.com> > Commit-Queue: Brian Osman <brianosman@google.com> > Auto-Submit: John Stiles <johnstiles@google.com> > Reviewed-by: Brian Osman <brianosman@google.com> TBR=brianosman@google.com,johnstiles@google.com Change-Id: Ica506467b0a4e03d0cbe482034acfa2d9f8d2c16 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337560 Reviewed-by: John Stiles <johnstiles@google.com>
2020-11-23 15:55:00 +00:00
OpDecorate %252 RelaxedPrecision
Revert "Eliminate inliner temporary variables for functions with a single exit." This reverts commit 345d72124d35b7c434be7bd9b6b16a66fbd980e6. Reason for revert: can break ES2 for-loop rules Original change's description: > Eliminate inliner temporary variables for functions with a single exit. > > When we determine that a function only contains a single return > statement, there is no need to create a temporary variable and store the > result expression into a variable. Instead, we can directly replace the > function-call expression with the return-statement's expression. > > This dramatically simplifies the final optimized output from chains of > very simple inlined functions, which is a very common pattern for trees > of Skia fragment processors. > > Change-Id: I6789064a321daf43db2e1cef4915f25ed74d6131 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/344665 > Commit-Queue: John Stiles <johnstiles@google.com> > Reviewed-by: Brian Osman <brianosman@google.com> > Auto-Submit: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,ethannicholas@google.com,johnstiles@google.com Change-Id: I60845f22159605a06047b030e2686a769121a35a No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://skia-review.googlesource.com/c/skia/+/345120 Reviewed-by: John Stiles <johnstiles@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-12-16 18:24:48 +00:00
OpDecorate %255 RelaxedPrecision
OpDecorate %256 RelaxedPrecision
Reland "Simplify _blend_set_color_saturation, removing an instruction." This reverts commit e81fb87bb444c7f1dd087479b0a868f365a465ae. Reason for revert: checking results with less-aggressive inliner Original change's description: > Revert "Simplify _blend_set_color_saturation, removing an instruction." > > This reverts commit ed289e777cfa41c059b94e64e0c749fe47db563b. > > Reason for revert: causing strange artifacts, only on Adreno > > Original change's description: > > Simplify _blend_set_color_saturation, removing an instruction. > > > > This tightens up our intrinsics slightly; after inlining, it eliminates > > one scratch variable. (We no longer need to copy `sda` into `hueColor` > > as hueColor is now unchanged.) > > > > Change-Id: Iece5ba2fe11cde54481704a1787114a2c2a66d9b > > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/336599 > > Commit-Queue: John Stiles <johnstiles@google.com> > > Commit-Queue: Brian Osman <brianosman@google.com> > > Auto-Submit: John Stiles <johnstiles@google.com> > > Reviewed-by: Brian Osman <brianosman@google.com> > > TBR=brianosman@google.com,johnstiles@google.com > > Change-Id: Ica506467b0a4e03d0cbe482034acfa2d9f8d2c16 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337560 > Reviewed-by: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,johnstiles@google.com Change-Id: Ia93263f3269c057e7eaa69ca2b05e783d18c0199 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337944 Auto-Submit: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-11-24 03:19:04 +00:00
OpDecorate %257 RelaxedPrecision
OpDecorate %259 RelaxedPrecision
Revert "Eliminate inliner temporary variables for functions with a single exit." This reverts commit 345d72124d35b7c434be7bd9b6b16a66fbd980e6. Reason for revert: can break ES2 for-loop rules Original change's description: > Eliminate inliner temporary variables for functions with a single exit. > > When we determine that a function only contains a single return > statement, there is no need to create a temporary variable and store the > result expression into a variable. Instead, we can directly replace the > function-call expression with the return-statement's expression. > > This dramatically simplifies the final optimized output from chains of > very simple inlined functions, which is a very common pattern for trees > of Skia fragment processors. > > Change-Id: I6789064a321daf43db2e1cef4915f25ed74d6131 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/344665 > Commit-Queue: John Stiles <johnstiles@google.com> > Reviewed-by: Brian Osman <brianosman@google.com> > Auto-Submit: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,ethannicholas@google.com,johnstiles@google.com Change-Id: I60845f22159605a06047b030e2686a769121a35a No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://skia-review.googlesource.com/c/skia/+/345120 Reviewed-by: John Stiles <johnstiles@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-12-16 18:24:48 +00:00
OpDecorate %260 RelaxedPrecision
Avoid inlining functions that are called repeatedly. Previously, we'd gauge suitability for inlining by counting the nodes in a function; past a certain limit, the function was considered "too big." Now, we also incorporate the number of times that function is called. So if a function is called three times, and its size is 20 nodes, it would be considered to have an inlining cost of 60 (3 * 20) instead of 20. This should tamp down the aggressive nature of the inliner in cases like gaussian convolution or complicated blends, and will hopefully satisfy Pinpoint. No change visible in Nanobench (which doesn't test any of these sorts of patterns, but certainly inlines things): http://screen/AwD5hkgkEfjVx4g Change-Id: Ie5e32898245ac854adb9ddd52d87001df6a67125 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337676 Commit-Queue: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com> Auto-Submit: John Stiles <johnstiles@google.com>
2020-11-23 19:48:06 +00:00
OpDecorate %261 RelaxedPrecision
Enable SPIR-V disassembly output from skslc. We now have SPIR-V golden outputs for `blend` and `shared` tests. This exposes a handful of SPIR-V limitations for us to address. Change-Id: Ie5278889b8a61432403d06231b17765885bee0ac Bug: skia:10694 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337182 Commit-Queue: John Stiles <johnstiles@google.com> Commit-Queue: Brian Osman <brianosman@google.com> Auto-Submit: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com>
2020-11-20 21:28:50 +00:00
OpDecorate %266 RelaxedPrecision
Avoid inlining functions that are called repeatedly. Previously, we'd gauge suitability for inlining by counting the nodes in a function; past a certain limit, the function was considered "too big." Now, we also incorporate the number of times that function is called. So if a function is called three times, and its size is 20 nodes, it would be considered to have an inlining cost of 60 (3 * 20) instead of 20. This should tamp down the aggressive nature of the inliner in cases like gaussian convolution or complicated blends, and will hopefully satisfy Pinpoint. No change visible in Nanobench (which doesn't test any of these sorts of patterns, but certainly inlines things): http://screen/AwD5hkgkEfjVx4g Change-Id: Ie5e32898245ac854adb9ddd52d87001df6a67125 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337676 Commit-Queue: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com> Auto-Submit: John Stiles <johnstiles@google.com>
2020-11-23 19:48:06 +00:00
OpDecorate %267 RelaxedPrecision
Reland "Simplify _blend_set_color_saturation, removing an instruction." This reverts commit e81fb87bb444c7f1dd087479b0a868f365a465ae. Reason for revert: checking results with less-aggressive inliner Original change's description: > Revert "Simplify _blend_set_color_saturation, removing an instruction." > > This reverts commit ed289e777cfa41c059b94e64e0c749fe47db563b. > > Reason for revert: causing strange artifacts, only on Adreno > > Original change's description: > > Simplify _blend_set_color_saturation, removing an instruction. > > > > This tightens up our intrinsics slightly; after inlining, it eliminates > > one scratch variable. (We no longer need to copy `sda` into `hueColor` > > as hueColor is now unchanged.) > > > > Change-Id: Iece5ba2fe11cde54481704a1787114a2c2a66d9b > > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/336599 > > Commit-Queue: John Stiles <johnstiles@google.com> > > Commit-Queue: Brian Osman <brianosman@google.com> > > Auto-Submit: John Stiles <johnstiles@google.com> > > Reviewed-by: Brian Osman <brianosman@google.com> > > TBR=brianosman@google.com,johnstiles@google.com > > Change-Id: Ica506467b0a4e03d0cbe482034acfa2d9f8d2c16 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337560 > Reviewed-by: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,johnstiles@google.com Change-Id: Ia93263f3269c057e7eaa69ca2b05e783d18c0199 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337944 Auto-Submit: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-11-24 03:19:04 +00:00
OpDecorate %268 RelaxedPrecision
Revert "Eliminate inliner temporary variables for functions with a single exit." This reverts commit 345d72124d35b7c434be7bd9b6b16a66fbd980e6. Reason for revert: can break ES2 for-loop rules Original change's description: > Eliminate inliner temporary variables for functions with a single exit. > > When we determine that a function only contains a single return > statement, there is no need to create a temporary variable and store the > result expression into a variable. Instead, we can directly replace the > function-call expression with the return-statement's expression. > > This dramatically simplifies the final optimized output from chains of > very simple inlined functions, which is a very common pattern for trees > of Skia fragment processors. > > Change-Id: I6789064a321daf43db2e1cef4915f25ed74d6131 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/344665 > Commit-Queue: John Stiles <johnstiles@google.com> > Reviewed-by: Brian Osman <brianosman@google.com> > Auto-Submit: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,ethannicholas@google.com,johnstiles@google.com Change-Id: I60845f22159605a06047b030e2686a769121a35a No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://skia-review.googlesource.com/c/skia/+/345120 Reviewed-by: John Stiles <johnstiles@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-12-16 18:24:48 +00:00
OpDecorate %269 RelaxedPrecision
OpDecorate %271 RelaxedPrecision
Reland "Simplify _blend_set_color_saturation, removing an instruction." This reverts commit e81fb87bb444c7f1dd087479b0a868f365a465ae. Reason for revert: checking results with less-aggressive inliner Original change's description: > Revert "Simplify _blend_set_color_saturation, removing an instruction." > > This reverts commit ed289e777cfa41c059b94e64e0c749fe47db563b. > > Reason for revert: causing strange artifacts, only on Adreno > > Original change's description: > > Simplify _blend_set_color_saturation, removing an instruction. > > > > This tightens up our intrinsics slightly; after inlining, it eliminates > > one scratch variable. (We no longer need to copy `sda` into `hueColor` > > as hueColor is now unchanged.) > > > > Change-Id: Iece5ba2fe11cde54481704a1787114a2c2a66d9b > > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/336599 > > Commit-Queue: John Stiles <johnstiles@google.com> > > Commit-Queue: Brian Osman <brianosman@google.com> > > Auto-Submit: John Stiles <johnstiles@google.com> > > Reviewed-by: Brian Osman <brianosman@google.com> > > TBR=brianosman@google.com,johnstiles@google.com > > Change-Id: Ica506467b0a4e03d0cbe482034acfa2d9f8d2c16 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337560 > Reviewed-by: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,johnstiles@google.com Change-Id: Ia93263f3269c057e7eaa69ca2b05e783d18c0199 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337944 Auto-Submit: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-11-24 03:19:04 +00:00
OpDecorate %272 RelaxedPrecision
Revert "Eliminate inliner temporary variables for functions with a single exit." This reverts commit 345d72124d35b7c434be7bd9b6b16a66fbd980e6. Reason for revert: can break ES2 for-loop rules Original change's description: > Eliminate inliner temporary variables for functions with a single exit. > > When we determine that a function only contains a single return > statement, there is no need to create a temporary variable and store the > result expression into a variable. Instead, we can directly replace the > function-call expression with the return-statement's expression. > > This dramatically simplifies the final optimized output from chains of > very simple inlined functions, which is a very common pattern for trees > of Skia fragment processors. > > Change-Id: I6789064a321daf43db2e1cef4915f25ed74d6131 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/344665 > Commit-Queue: John Stiles <johnstiles@google.com> > Reviewed-by: Brian Osman <brianosman@google.com> > Auto-Submit: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,ethannicholas@google.com,johnstiles@google.com Change-Id: I60845f22159605a06047b030e2686a769121a35a No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://skia-review.googlesource.com/c/skia/+/345120 Reviewed-by: John Stiles <johnstiles@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-12-16 18:24:48 +00:00
OpDecorate %273 RelaxedPrecision
Reland "Simplify _blend_set_color_saturation, removing an instruction." This reverts commit e81fb87bb444c7f1dd087479b0a868f365a465ae. Reason for revert: checking results with less-aggressive inliner Original change's description: > Revert "Simplify _blend_set_color_saturation, removing an instruction." > > This reverts commit ed289e777cfa41c059b94e64e0c749fe47db563b. > > Reason for revert: causing strange artifacts, only on Adreno > > Original change's description: > > Simplify _blend_set_color_saturation, removing an instruction. > > > > This tightens up our intrinsics slightly; after inlining, it eliminates > > one scratch variable. (We no longer need to copy `sda` into `hueColor` > > as hueColor is now unchanged.) > > > > Change-Id: Iece5ba2fe11cde54481704a1787114a2c2a66d9b > > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/336599 > > Commit-Queue: John Stiles <johnstiles@google.com> > > Commit-Queue: Brian Osman <brianosman@google.com> > > Auto-Submit: John Stiles <johnstiles@google.com> > > Reviewed-by: Brian Osman <brianosman@google.com> > > TBR=brianosman@google.com,johnstiles@google.com > > Change-Id: Ica506467b0a4e03d0cbe482034acfa2d9f8d2c16 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337560 > Reviewed-by: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,johnstiles@google.com Change-Id: Ia93263f3269c057e7eaa69ca2b05e783d18c0199 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337944 Auto-Submit: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-11-24 03:19:04 +00:00
OpDecorate %274 RelaxedPrecision
Revert "Simplify _blend_set_color_saturation, removing an instruction." This reverts commit ed289e777cfa41c059b94e64e0c749fe47db563b. Reason for revert: causing strange artifacts, only on Adreno Original change's description: > Simplify _blend_set_color_saturation, removing an instruction. > > This tightens up our intrinsics slightly; after inlining, it eliminates > one scratch variable. (We no longer need to copy `sda` into `hueColor` > as hueColor is now unchanged.) > > Change-Id: Iece5ba2fe11cde54481704a1787114a2c2a66d9b > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/336599 > Commit-Queue: John Stiles <johnstiles@google.com> > Commit-Queue: Brian Osman <brianosman@google.com> > Auto-Submit: John Stiles <johnstiles@google.com> > Reviewed-by: Brian Osman <brianosman@google.com> TBR=brianosman@google.com,johnstiles@google.com Change-Id: Ica506467b0a4e03d0cbe482034acfa2d9f8d2c16 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337560 Reviewed-by: John Stiles <johnstiles@google.com>
2020-11-23 15:55:00 +00:00
OpDecorate %276 RelaxedPrecision
Avoid inlining functions that are called repeatedly. Previously, we'd gauge suitability for inlining by counting the nodes in a function; past a certain limit, the function was considered "too big." Now, we also incorporate the number of times that function is called. So if a function is called three times, and its size is 20 nodes, it would be considered to have an inlining cost of 60 (3 * 20) instead of 20. This should tamp down the aggressive nature of the inliner in cases like gaussian convolution or complicated blends, and will hopefully satisfy Pinpoint. No change visible in Nanobench (which doesn't test any of these sorts of patterns, but certainly inlines things): http://screen/AwD5hkgkEfjVx4g Change-Id: Ie5e32898245ac854adb9ddd52d87001df6a67125 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337676 Commit-Queue: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com> Auto-Submit: John Stiles <johnstiles@google.com>
2020-11-23 19:48:06 +00:00
OpDecorate %277 RelaxedPrecision
OpDecorate %278 RelaxedPrecision
Revert "Eliminate inliner temporary variables for functions with a single exit." This reverts commit 345d72124d35b7c434be7bd9b6b16a66fbd980e6. Reason for revert: can break ES2 for-loop rules Original change's description: > Eliminate inliner temporary variables for functions with a single exit. > > When we determine that a function only contains a single return > statement, there is no need to create a temporary variable and store the > result expression into a variable. Instead, we can directly replace the > function-call expression with the return-statement's expression. > > This dramatically simplifies the final optimized output from chains of > very simple inlined functions, which is a very common pattern for trees > of Skia fragment processors. > > Change-Id: I6789064a321daf43db2e1cef4915f25ed74d6131 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/344665 > Commit-Queue: John Stiles <johnstiles@google.com> > Reviewed-by: Brian Osman <brianosman@google.com> > Auto-Submit: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,ethannicholas@google.com,johnstiles@google.com Change-Id: I60845f22159605a06047b030e2686a769121a35a No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://skia-review.googlesource.com/c/skia/+/345120 Reviewed-by: John Stiles <johnstiles@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-12-16 18:24:48 +00:00
OpDecorate %282 RelaxedPrecision
OpDecorate %284 RelaxedPrecision
OpDecorate %286 RelaxedPrecision
OpDecorate %287 RelaxedPrecision
OpDecorate %288 RelaxedPrecision
OpDecorate %290 RelaxedPrecision
Enable SPIR-V disassembly output from skslc. We now have SPIR-V golden outputs for `blend` and `shared` tests. This exposes a handful of SPIR-V limitations for us to address. Change-Id: Ie5278889b8a61432403d06231b17765885bee0ac Bug: skia:10694 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337182 Commit-Queue: John Stiles <johnstiles@google.com> Commit-Queue: Brian Osman <brianosman@google.com> Auto-Submit: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com>
2020-11-20 21:28:50 +00:00
%float = OpTypeFloat 32
%v4float = OpTypeVector %float 4
%_ptr_Output_v4float = OpTypePointer Output %v4float
%sk_FragColor = OpVariable %_ptr_Output_v4float Output
%bool = OpTypeBool
%_ptr_Input_bool = OpTypePointer Input %bool
%sk_Clockwise = OpVariable %_ptr_Input_bool Input
%_ptr_Input_v4float = OpTypePointer Input %v4float
%src = OpVariable %_ptr_Input_v4float Input
%dst = OpVariable %_ptr_Input_v4float Input
%v3float = OpTypeVector %float 3
%_ptr_Function_v3float = OpTypePointer Function %v3float
Avoid inlining functions that are called repeatedly. Previously, we'd gauge suitability for inlining by counting the nodes in a function; past a certain limit, the function was considered "too big." Now, we also incorporate the number of times that function is called. So if a function is called three times, and its size is 20 nodes, it would be considered to have an inlining cost of 60 (3 * 20) instead of 20. This should tamp down the aggressive nature of the inliner in cases like gaussian convolution or complicated blends, and will hopefully satisfy Pinpoint. No change visible in Nanobench (which doesn't test any of these sorts of patterns, but certainly inlines things): http://screen/AwD5hkgkEfjVx4g Change-Id: Ie5e32898245ac854adb9ddd52d87001df6a67125 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337676 Commit-Queue: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com> Auto-Submit: John Stiles <johnstiles@google.com>
2020-11-23 19:48:06 +00:00
%_ptr_Function_float = OpTypePointer Function %float
%15 = OpTypeFunction %v3float %_ptr_Function_v3float %_ptr_Function_float
Enable SPIR-V disassembly output from skslc. We now have SPIR-V golden outputs for `blend` and `shared` tests. This exposes a handful of SPIR-V limitations for us to address. Change-Id: Ie5278889b8a61432403d06231b17765885bee0ac Bug: skia:10694 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337182 Commit-Queue: John Stiles <johnstiles@google.com> Commit-Queue: Brian Osman <brianosman@google.com> Auto-Submit: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com>
2020-11-20 21:28:50 +00:00
%float_0 = OpConstant %float 0
Avoid inlining functions that are called repeatedly. Previously, we'd gauge suitability for inlining by counting the nodes in a function; past a certain limit, the function was considered "too big." Now, we also incorporate the number of times that function is called. So if a function is called three times, and its size is 20 nodes, it would be considered to have an inlining cost of 60 (3 * 20) instead of 20. This should tamp down the aggressive nature of the inliner in cases like gaussian convolution or complicated blends, and will hopefully satisfy Pinpoint. No change visible in Nanobench (which doesn't test any of these sorts of patterns, but certainly inlines things): http://screen/AwD5hkgkEfjVx4g Change-Id: Ie5e32898245ac854adb9ddd52d87001df6a67125 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337676 Commit-Queue: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com> Auto-Submit: John Stiles <johnstiles@google.com>
2020-11-23 19:48:06 +00:00
%46 = OpConstantComposite %v3float %float_0 %float_0 %float_0
%void = OpTypeVoid
%49 = OpTypeFunction %void
Revert "Eliminate inliner temporary variables for functions with a single exit." This reverts commit 345d72124d35b7c434be7bd9b6b16a66fbd980e6. Reason for revert: can break ES2 for-loop rules Original change's description: > Eliminate inliner temporary variables for functions with a single exit. > > When we determine that a function only contains a single return > statement, there is no need to create a temporary variable and store the > result expression into a variable. Instead, we can directly replace the > function-call expression with the return-statement's expression. > > This dramatically simplifies the final optimized output from chains of > very simple inlined functions, which is a very common pattern for trees > of Skia fragment processors. > > Change-Id: I6789064a321daf43db2e1cef4915f25ed74d6131 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/344665 > Commit-Queue: John Stiles <johnstiles@google.com> > Reviewed-by: Brian Osman <brianosman@google.com> > Auto-Submit: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,ethannicholas@google.com,johnstiles@google.com Change-Id: I60845f22159605a06047b030e2686a769121a35a No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://skia-review.googlesource.com/c/skia/+/345120 Reviewed-by: John Stiles <johnstiles@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-12-16 18:24:48 +00:00
%_ptr_Function_v4float = OpTypePointer Function %v4float
Enable SPIR-V disassembly output from skslc. We now have SPIR-V golden outputs for `blend` and `shared` tests. This exposes a handful of SPIR-V limitations for us to address. Change-Id: Ie5278889b8a61432403d06231b17765885bee0ac Bug: skia:10694 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337182 Commit-Queue: John Stiles <johnstiles@google.com> Commit-Queue: Brian Osman <brianosman@google.com> Auto-Submit: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com>
2020-11-20 21:28:50 +00:00
%float_0_300000012 = OpConstant %float 0.300000012
%float_0_589999974 = OpConstant %float 0.589999974
%float_0_109999999 = OpConstant %float 0.109999999
Declare all inlined variables at the topmost scope possible. By itself, this is uninteresting and even perhaps slightly counterproductive (as it separates vardecl from its initializer, increasing LOC). However, this enables a followup CL (http://review.skia.org/344665) which allows single-return functions to be inlined without the creation of a temporary variable at all. This applies to the majority of fragment processors in a typical Ganesh hierarchy. This change will greatly reduce the number of inliner-created temporary copies when compiling a typical tree of FPs. Change-Id: I03423a13cf35050637dabace4a32973a08a4ed0a Reviewed-on: https://skia-review.googlesource.com/c/skia/+/344764 Reviewed-by: Brian Osman <brianosman@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-12-15 22:50:21 +00:00
%180 = OpConstantComposite %v3float %float_0_300000012 %float_0_589999974 %float_0_109999999
Revert "Eliminate inliner temporary variables for functions with a single exit." This reverts commit 345d72124d35b7c434be7bd9b6b16a66fbd980e6. Reason for revert: can break ES2 for-loop rules Original change's description: > Eliminate inliner temporary variables for functions with a single exit. > > When we determine that a function only contains a single return > statement, there is no need to create a temporary variable and store the > result expression into a variable. Instead, we can directly replace the > function-call expression with the return-statement's expression. > > This dramatically simplifies the final optimized output from chains of > very simple inlined functions, which is a very common pattern for trees > of Skia fragment processors. > > Change-Id: I6789064a321daf43db2e1cef4915f25ed74d6131 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/344665 > Commit-Queue: John Stiles <johnstiles@google.com> > Reviewed-by: Brian Osman <brianosman@google.com> > Auto-Submit: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,ethannicholas@google.com,johnstiles@google.com Change-Id: I60845f22159605a06047b030e2686a769121a35a No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://skia-review.googlesource.com/c/skia/+/345120 Reviewed-by: John Stiles <johnstiles@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-12-16 18:24:48 +00:00
%187 = OpConstantComposite %v3float %float_0_300000012 %float_0_589999974 %float_0_109999999
Enable SPIR-V disassembly output from skslc. We now have SPIR-V golden outputs for `blend` and `shared` tests. This exposes a handful of SPIR-V limitations for us to address. Change-Id: Ie5278889b8a61432403d06231b17765885bee0ac Bug: skia:10694 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337182 Commit-Queue: John Stiles <johnstiles@google.com> Commit-Queue: Brian Osman <brianosman@google.com> Auto-Submit: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com>
2020-11-20 21:28:50 +00:00
%false = OpConstantFalse %bool
%float_1 = OpConstant %float 1
Avoid inlining functions that are called repeatedly. Previously, we'd gauge suitability for inlining by counting the nodes in a function; past a certain limit, the function was considered "too big." Now, we also incorporate the number of times that function is called. So if a function is called three times, and its size is 20 nodes, it would be considered to have an inlining cost of 60 (3 * 20) instead of 20. This should tamp down the aggressive nature of the inliner in cases like gaussian convolution or complicated blends, and will hopefully satisfy Pinpoint. No change visible in Nanobench (which doesn't test any of these sorts of patterns, but certainly inlines things): http://screen/AwD5hkgkEfjVx4g Change-Id: Ie5e32898245ac854adb9ddd52d87001df6a67125 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337676 Commit-Queue: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com> Auto-Submit: John Stiles <johnstiles@google.com>
2020-11-23 19:48:06 +00:00
%_blend_set_color_saturation_helper = OpFunction %v3float None %15
%18 = OpFunctionParameter %_ptr_Function_v3float
%19 = OpFunctionParameter %_ptr_Function_float
%20 = OpLabel
%26 = OpVariable %_ptr_Function_v3float Function
%21 = OpLoad %v3float %18
%22 = OpCompositeExtract %float %21 0
%23 = OpLoad %v3float %18
%24 = OpCompositeExtract %float %23 2
%25 = OpFOrdLessThan %bool %22 %24
OpSelectionMerge %29 None
OpBranchConditional %25 %27 %28
%27 = OpLabel
%31 = OpLoad %float %19
%32 = OpLoad %v3float %18
%33 = OpCompositeExtract %float %32 1
%34 = OpLoad %v3float %18
%35 = OpCompositeExtract %float %34 0
%36 = OpFSub %float %33 %35
%37 = OpFMul %float %31 %36
%38 = OpLoad %v3float %18
%39 = OpCompositeExtract %float %38 2
%40 = OpLoad %v3float %18
%41 = OpCompositeExtract %float %40 0
%42 = OpFSub %float %39 %41
%43 = OpFDiv %float %37 %42
%44 = OpLoad %float %19
%45 = OpCompositeConstruct %v3float %float_0 %43 %44
OpStore %26 %45
OpBranch %29
%28 = OpLabel
OpStore %26 %46
OpBranch %29
%29 = OpLabel
%47 = OpLoad %v3float %26
OpReturnValue %47
OpFunctionEnd
%main = OpFunction %void None %49
%50 = OpLabel
Revert "Eliminate inliner temporary variables for functions with a single exit." This reverts commit 345d72124d35b7c434be7bd9b6b16a66fbd980e6. Reason for revert: can break ES2 for-loop rules Original change's description: > Eliminate inliner temporary variables for functions with a single exit. > > When we determine that a function only contains a single return > statement, there is no need to create a temporary variable and store the > result expression into a variable. Instead, we can directly replace the > function-call expression with the return-statement's expression. > > This dramatically simplifies the final optimized output from chains of > very simple inlined functions, which is a very common pattern for trees > of Skia fragment processors. > > Change-Id: I6789064a321daf43db2e1cef4915f25ed74d6131 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/344665 > Commit-Queue: John Stiles <johnstiles@google.com> > Reviewed-by: Brian Osman <brianosman@google.com> > Auto-Submit: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,ethannicholas@google.com,johnstiles@google.com Change-Id: I60845f22159605a06047b030e2686a769121a35a No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://skia-review.googlesource.com/c/skia/+/345120 Reviewed-by: John Stiles <johnstiles@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-12-16 18:24:48 +00:00
%_0_blend_saturation = OpVariable %_ptr_Function_v4float Function
Enable SPIR-V disassembly output from skslc. We now have SPIR-V golden outputs for `blend` and `shared` tests. This exposes a handful of SPIR-V limitations for us to address. Change-Id: Ie5278889b8a61432403d06231b17765885bee0ac Bug: skia:10694 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337182 Commit-Queue: John Stiles <johnstiles@google.com> Commit-Queue: Brian Osman <brianosman@google.com> Auto-Submit: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com>
2020-11-20 21:28:50 +00:00
%_1_alpha = OpVariable %_ptr_Function_float Function
%_2_sda = OpVariable %_ptr_Function_v3float Function
%_3_dsa = OpVariable %_ptr_Function_v3float Function
%_4_blend_set_color_saturation = OpVariable %_ptr_Function_v3float Function
Revert "Eliminate inliner temporary variables for functions with a single exit." This reverts commit 345d72124d35b7c434be7bd9b6b16a66fbd980e6. Reason for revert: can break ES2 for-loop rules Original change's description: > Eliminate inliner temporary variables for functions with a single exit. > > When we determine that a function only contains a single return > statement, there is no need to create a temporary variable and store the > result expression into a variable. Instead, we can directly replace the > function-call expression with the return-statement's expression. > > This dramatically simplifies the final optimized output from chains of > very simple inlined functions, which is a very common pattern for trees > of Skia fragment processors. > > Change-Id: I6789064a321daf43db2e1cef4915f25ed74d6131 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/344665 > Commit-Queue: John Stiles <johnstiles@google.com> > Reviewed-by: Brian Osman <brianosman@google.com> > Auto-Submit: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,ethannicholas@google.com,johnstiles@google.com Change-Id: I60845f22159605a06047b030e2686a769121a35a No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://skia-review.googlesource.com/c/skia/+/345120 Reviewed-by: John Stiles <johnstiles@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-12-16 18:24:48 +00:00
%_5_13_blend_color_saturation = OpVariable %_ptr_Function_float Function
%_6_sat = OpVariable %_ptr_Function_float Function
%109 = OpVariable %_ptr_Function_v3float Function
%111 = OpVariable %_ptr_Function_float Function
%123 = OpVariable %_ptr_Function_v3float Function
%125 = OpVariable %_ptr_Function_float Function
%130 = OpVariable %_ptr_Function_v3float Function
%132 = OpVariable %_ptr_Function_float Function
%145 = OpVariable %_ptr_Function_v3float Function
%147 = OpVariable %_ptr_Function_float Function
%160 = OpVariable %_ptr_Function_v3float Function
%162 = OpVariable %_ptr_Function_float Function
%167 = OpVariable %_ptr_Function_v3float Function
%169 = OpVariable %_ptr_Function_float Function
%_7_blend_set_color_luminance = OpVariable %_ptr_Function_v3float Function
%_8_11_blend_color_luminance = OpVariable %_ptr_Function_float Function
%_9_lum = OpVariable %_ptr_Function_float Function
%_10_12_blend_color_luminance = OpVariable %_ptr_Function_float Function
%_11_result = OpVariable %_ptr_Function_v3float Function
%_12_minComp = OpVariable %_ptr_Function_float Function
%_13_maxComp = OpVariable %_ptr_Function_float Function
%246 = OpVariable %_ptr_Function_v3float Function
%56 = OpLoad %v4float %dst
Avoid inlining functions that are called repeatedly. Previously, we'd gauge suitability for inlining by counting the nodes in a function; past a certain limit, the function was considered "too big." Now, we also incorporate the number of times that function is called. So if a function is called three times, and its size is 20 nodes, it would be considered to have an inlining cost of 60 (3 * 20) instead of 20. This should tamp down the aggressive nature of the inliner in cases like gaussian convolution or complicated blends, and will hopefully satisfy Pinpoint. No change visible in Nanobench (which doesn't test any of these sorts of patterns, but certainly inlines things): http://screen/AwD5hkgkEfjVx4g Change-Id: Ie5e32898245ac854adb9ddd52d87001df6a67125 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337676 Commit-Queue: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com> Auto-Submit: John Stiles <johnstiles@google.com>
2020-11-23 19:48:06 +00:00
%57 = OpCompositeExtract %float %56 3
Revert "Eliminate inliner temporary variables for functions with a single exit." This reverts commit 345d72124d35b7c434be7bd9b6b16a66fbd980e6. Reason for revert: can break ES2 for-loop rules Original change's description: > Eliminate inliner temporary variables for functions with a single exit. > > When we determine that a function only contains a single return > statement, there is no need to create a temporary variable and store the > result expression into a variable. Instead, we can directly replace the > function-call expression with the return-statement's expression. > > This dramatically simplifies the final optimized output from chains of > very simple inlined functions, which is a very common pattern for trees > of Skia fragment processors. > > Change-Id: I6789064a321daf43db2e1cef4915f25ed74d6131 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/344665 > Commit-Queue: John Stiles <johnstiles@google.com> > Reviewed-by: Brian Osman <brianosman@google.com> > Auto-Submit: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,ethannicholas@google.com,johnstiles@google.com Change-Id: I60845f22159605a06047b030e2686a769121a35a No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://skia-review.googlesource.com/c/skia/+/345120 Reviewed-by: John Stiles <johnstiles@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-12-16 18:24:48 +00:00
%58 = OpLoad %v4float %src
%59 = OpCompositeExtract %float %58 3
%60 = OpFMul %float %57 %59
OpStore %_1_alpha %60
%61 = OpLoad %v4float %src
%62 = OpVectorShuffle %v3float %61 %61 0 1 2
%63 = OpLoad %v4float %dst
%64 = OpCompositeExtract %float %63 3
%65 = OpVectorTimesScalar %v3float %62 %64
OpStore %_2_sda %65
%66 = OpLoad %v4float %dst
%67 = OpVectorShuffle %v3float %66 %66 0 1 2
%68 = OpLoad %v4float %src
%69 = OpCompositeExtract %float %68 3
%70 = OpVectorTimesScalar %v3float %67 %69
OpStore %_3_dsa %70
%76 = OpLoad %v3float %_2_sda
%77 = OpCompositeExtract %float %76 0
%78 = OpLoad %v3float %_2_sda
%79 = OpCompositeExtract %float %78 1
%75 = OpExtInst %float %1 FMax %77 %79
%80 = OpLoad %v3float %_2_sda
%81 = OpCompositeExtract %float %80 2
%74 = OpExtInst %float %1 FMax %75 %81
%84 = OpLoad %v3float %_2_sda
%85 = OpCompositeExtract %float %84 0
%86 = OpLoad %v3float %_2_sda
%87 = OpCompositeExtract %float %86 1
%83 = OpExtInst %float %1 FMin %85 %87
%88 = OpLoad %v3float %_2_sda
%89 = OpCompositeExtract %float %88 2
%82 = OpExtInst %float %1 FMin %83 %89
%90 = OpFSub %float %74 %82
OpStore %_5_13_blend_color_saturation %90
%91 = OpLoad %float %_5_13_blend_color_saturation
OpStore %_6_sat %91
%92 = OpLoad %v3float %_3_dsa
%93 = OpCompositeExtract %float %92 0
%94 = OpLoad %v3float %_3_dsa
%95 = OpCompositeExtract %float %94 1
%96 = OpFOrdLessThanEqual %bool %93 %95
OpSelectionMerge %99 None
OpBranchConditional %96 %97 %98
%97 = OpLabel
%100 = OpLoad %v3float %_3_dsa
%101 = OpCompositeExtract %float %100 1
%102 = OpLoad %v3float %_3_dsa
%103 = OpCompositeExtract %float %102 2
%104 = OpFOrdLessThanEqual %bool %101 %103
OpSelectionMerge %107 None
OpBranchConditional %104 %105 %106
%105 = OpLabel
%108 = OpLoad %v3float %_3_dsa
OpStore %109 %108
%110 = OpLoad %float %_6_sat
OpStore %111 %110
%112 = OpFunctionCall %v3float %_blend_set_color_saturation_helper %109 %111
OpStore %_4_blend_set_color_saturation %112
OpBranch %107
%106 = OpLabel
%113 = OpLoad %v3float %_3_dsa
%114 = OpCompositeExtract %float %113 0
%115 = OpLoad %v3float %_3_dsa
%116 = OpCompositeExtract %float %115 2
%117 = OpFOrdLessThanEqual %bool %114 %116
OpSelectionMerge %120 None
OpBranchConditional %117 %118 %119
%118 = OpLabel
%121 = OpLoad %v3float %_3_dsa
%122 = OpVectorShuffle %v3float %121 %121 0 2 1
OpStore %123 %122
%124 = OpLoad %float %_6_sat
OpStore %125 %124
%126 = OpFunctionCall %v3float %_blend_set_color_saturation_helper %123 %125
%127 = OpVectorShuffle %v3float %126 %126 0 2 1
OpStore %_4_blend_set_color_saturation %127
OpBranch %120
%119 = OpLabel
%128 = OpLoad %v3float %_3_dsa
%129 = OpVectorShuffle %v3float %128 %128 2 0 1
OpStore %130 %129
%131 = OpLoad %float %_6_sat
OpStore %132 %131
%133 = OpFunctionCall %v3float %_blend_set_color_saturation_helper %130 %132
%134 = OpVectorShuffle %v3float %133 %133 1 2 0
OpStore %_4_blend_set_color_saturation %134
OpBranch %120
%120 = OpLabel
OpBranch %107
%107 = OpLabel
OpBranch %99
%98 = OpLabel
%135 = OpLoad %v3float %_3_dsa
%136 = OpCompositeExtract %float %135 0
%137 = OpLoad %v3float %_3_dsa
%138 = OpCompositeExtract %float %137 2
%139 = OpFOrdLessThanEqual %bool %136 %138
OpSelectionMerge %142 None
OpBranchConditional %139 %140 %141
%140 = OpLabel
%143 = OpLoad %v3float %_3_dsa
%144 = OpVectorShuffle %v3float %143 %143 1 0 2
OpStore %145 %144
%146 = OpLoad %float %_6_sat
OpStore %147 %146
%148 = OpFunctionCall %v3float %_blend_set_color_saturation_helper %145 %147
%149 = OpVectorShuffle %v3float %148 %148 1 0 2
OpStore %_4_blend_set_color_saturation %149
OpBranch %142
%141 = OpLabel
%150 = OpLoad %v3float %_3_dsa
%151 = OpCompositeExtract %float %150 1
%152 = OpLoad %v3float %_3_dsa
%153 = OpCompositeExtract %float %152 2
%154 = OpFOrdLessThanEqual %bool %151 %153
OpSelectionMerge %157 None
OpBranchConditional %154 %155 %156
%155 = OpLabel
%158 = OpLoad %v3float %_3_dsa
%159 = OpVectorShuffle %v3float %158 %158 1 2 0
OpStore %160 %159
%161 = OpLoad %float %_6_sat
OpStore %162 %161
%163 = OpFunctionCall %v3float %_blend_set_color_saturation_helper %160 %162
%164 = OpVectorShuffle %v3float %163 %163 2 0 1
OpStore %_4_blend_set_color_saturation %164
OpBranch %157
%156 = OpLabel
%165 = OpLoad %v3float %_3_dsa
%166 = OpVectorShuffle %v3float %165 %165 2 1 0
OpStore %167 %166
%168 = OpLoad %float %_6_sat
OpStore %169 %168
%170 = OpFunctionCall %v3float %_blend_set_color_saturation_helper %167 %169
%171 = OpVectorShuffle %v3float %170 %170 2 1 0
OpStore %_4_blend_set_color_saturation %171
OpBranch %157
%157 = OpLabel
OpBranch %142
%142 = OpLabel
OpBranch %99
%99 = OpLabel
%184 = OpLoad %v3float %_3_dsa
%179 = OpDot %float %180 %184
OpStore %_8_11_blend_color_luminance %179
%185 = OpLoad %float %_8_11_blend_color_luminance
OpStore %_9_lum %185
%188 = OpLoad %v3float %_4_blend_set_color_saturation
%186 = OpDot %float %187 %188
OpStore %_10_12_blend_color_luminance %186
%189 = OpLoad %float %_9_lum
%190 = OpLoad %float %_10_12_blend_color_luminance
%191 = OpFSub %float %189 %190
%192 = OpLoad %v3float %_4_blend_set_color_saturation
%193 = OpCompositeConstruct %v3float %191 %191 %191
%194 = OpFAdd %v3float %193 %192
OpStore %_11_result %194
%197 = OpLoad %v3float %_11_result
%198 = OpCompositeExtract %float %197 0
%199 = OpLoad %v3float %_11_result
%200 = OpCompositeExtract %float %199 1
%196 = OpExtInst %float %1 FMin %198 %200
%201 = OpLoad %v3float %_11_result
%202 = OpCompositeExtract %float %201 2
%195 = OpExtInst %float %1 FMin %196 %202
OpStore %_12_minComp %195
%205 = OpLoad %v3float %_11_result
%206 = OpCompositeExtract %float %205 0
%207 = OpLoad %v3float %_11_result
%208 = OpCompositeExtract %float %207 1
%204 = OpExtInst %float %1 FMax %206 %208
%209 = OpLoad %v3float %_11_result
%210 = OpCompositeExtract %float %209 2
%203 = OpExtInst %float %1 FMax %204 %210
OpStore %_13_maxComp %203
%212 = OpLoad %float %_12_minComp
%213 = OpFOrdLessThan %bool %212 %float_0
OpSelectionMerge %215 None
OpBranchConditional %213 %214 %215
%214 = OpLabel
%216 = OpLoad %float %_9_lum
%217 = OpLoad %float %_12_minComp
%218 = OpFOrdNotEqual %bool %216 %217
OpBranch %215
%215 = OpLabel
%219 = OpPhi %bool %false %99 %218 %214
OpSelectionMerge %221 None
OpBranchConditional %219 %220 %221
%220 = OpLabel
%222 = OpLoad %float %_9_lum
%223 = OpLoad %v3float %_11_result
%224 = OpLoad %float %_9_lum
%225 = OpCompositeConstruct %v3float %224 %224 %224
%226 = OpFSub %v3float %223 %225
%227 = OpLoad %float %_9_lum
%228 = OpVectorTimesScalar %v3float %226 %227
%229 = OpLoad %float %_9_lum
%230 = OpLoad %float %_12_minComp
%231 = OpFSub %float %229 %230
%233 = OpFDiv %float %float_1 %231
%234 = OpVectorTimesScalar %v3float %228 %233
%235 = OpCompositeConstruct %v3float %222 %222 %222
%236 = OpFAdd %v3float %235 %234
OpStore %_11_result %236
OpBranch %221
%221 = OpLabel
%237 = OpLoad %float %_13_maxComp
%238 = OpLoad %float %_1_alpha
%239 = OpFOrdGreaterThan %bool %237 %238
OpSelectionMerge %241 None
OpBranchConditional %239 %240 %241
Reland "Simplify _blend_set_color_saturation, removing an instruction." This reverts commit e81fb87bb444c7f1dd087479b0a868f365a465ae. Reason for revert: checking results with less-aggressive inliner Original change's description: > Revert "Simplify _blend_set_color_saturation, removing an instruction." > > This reverts commit ed289e777cfa41c059b94e64e0c749fe47db563b. > > Reason for revert: causing strange artifacts, only on Adreno > > Original change's description: > > Simplify _blend_set_color_saturation, removing an instruction. > > > > This tightens up our intrinsics slightly; after inlining, it eliminates > > one scratch variable. (We no longer need to copy `sda` into `hueColor` > > as hueColor is now unchanged.) > > > > Change-Id: Iece5ba2fe11cde54481704a1787114a2c2a66d9b > > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/336599 > > Commit-Queue: John Stiles <johnstiles@google.com> > > Commit-Queue: Brian Osman <brianosman@google.com> > > Auto-Submit: John Stiles <johnstiles@google.com> > > Reviewed-by: Brian Osman <brianosman@google.com> > > TBR=brianosman@google.com,johnstiles@google.com > > Change-Id: Ica506467b0a4e03d0cbe482034acfa2d9f8d2c16 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337560 > Reviewed-by: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,johnstiles@google.com Change-Id: Ia93263f3269c057e7eaa69ca2b05e783d18c0199 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337944 Auto-Submit: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-11-24 03:19:04 +00:00
%240 = OpLabel
Revert "Eliminate inliner temporary variables for functions with a single exit." This reverts commit 345d72124d35b7c434be7bd9b6b16a66fbd980e6. Reason for revert: can break ES2 for-loop rules Original change's description: > Eliminate inliner temporary variables for functions with a single exit. > > When we determine that a function only contains a single return > statement, there is no need to create a temporary variable and store the > result expression into a variable. Instead, we can directly replace the > function-call expression with the return-statement's expression. > > This dramatically simplifies the final optimized output from chains of > very simple inlined functions, which is a very common pattern for trees > of Skia fragment processors. > > Change-Id: I6789064a321daf43db2e1cef4915f25ed74d6131 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/344665 > Commit-Queue: John Stiles <johnstiles@google.com> > Reviewed-by: Brian Osman <brianosman@google.com> > Auto-Submit: John Stiles <johnstiles@google.com> TBR=brianosman@google.com,ethannicholas@google.com,johnstiles@google.com Change-Id: I60845f22159605a06047b030e2686a769121a35a No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://skia-review.googlesource.com/c/skia/+/345120 Reviewed-by: John Stiles <johnstiles@google.com> Commit-Queue: John Stiles <johnstiles@google.com>
2020-12-16 18:24:48 +00:00
%242 = OpLoad %float %_13_maxComp
%243 = OpLoad %float %_9_lum
%244 = OpFOrdNotEqual %bool %242 %243
OpBranch %241
%241 = OpLabel
%245 = OpPhi %bool %false %221 %244 %240
OpSelectionMerge %249 None
OpBranchConditional %245 %247 %248
%247 = OpLabel
%250 = OpLoad %float %_9_lum
%251 = OpLoad %v3float %_11_result
%252 = OpLoad %float %_9_lum
%253 = OpCompositeConstruct %v3float %252 %252 %252
%254 = OpFSub %v3float %251 %253
%255 = OpLoad %float %_1_alpha
%256 = OpLoad %float %_9_lum
%257 = OpFSub %float %255 %256
%258 = OpVectorTimesScalar %v3float %254 %257
%259 = OpLoad %float %_13_maxComp
%260 = OpLoad %float %_9_lum
%261 = OpFSub %float %259 %260
%262 = OpFDiv %float %float_1 %261
%263 = OpVectorTimesScalar %v3float %258 %262
%264 = OpCompositeConstruct %v3float %250 %250 %250
%265 = OpFAdd %v3float %264 %263
OpStore %246 %265
OpBranch %249
%248 = OpLabel
%266 = OpLoad %v3float %_11_result
OpStore %246 %266
OpBranch %249
%249 = OpLabel
%267 = OpLoad %v3float %246
OpStore %_7_blend_set_color_luminance %267
%268 = OpLoad %v3float %_7_blend_set_color_luminance
%269 = OpLoad %v4float %dst
%270 = OpVectorShuffle %v3float %269 %269 0 1 2
%271 = OpFAdd %v3float %268 %270
%272 = OpLoad %v3float %_3_dsa
%273 = OpFSub %v3float %271 %272
%274 = OpLoad %v4float %src
%275 = OpVectorShuffle %v3float %274 %274 0 1 2
%276 = OpFAdd %v3float %273 %275
%277 = OpLoad %v3float %_2_sda
%278 = OpFSub %v3float %276 %277
%279 = OpCompositeExtract %float %278 0
%280 = OpCompositeExtract %float %278 1
%281 = OpCompositeExtract %float %278 2
%282 = OpLoad %v4float %src
%283 = OpCompositeExtract %float %282 3
%284 = OpLoad %v4float %dst
%285 = OpCompositeExtract %float %284 3
%286 = OpFAdd %float %283 %285
%287 = OpLoad %float %_1_alpha
%288 = OpFSub %float %286 %287
%289 = OpCompositeConstruct %v4float %279 %280 %281 %288
OpStore %_0_blend_saturation %289
%290 = OpLoad %v4float %_0_blend_saturation
OpStore %sk_FragColor %290
Enable SPIR-V disassembly output from skslc. We now have SPIR-V golden outputs for `blend` and `shared` tests. This exposes a handful of SPIR-V limitations for us to address. Change-Id: Ie5278889b8a61432403d06231b17765885bee0ac Bug: skia:10694 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/337182 Commit-Queue: John Stiles <johnstiles@google.com> Commit-Queue: Brian Osman <brianosman@google.com> Auto-Submit: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com>
2020-11-20 21:28:50 +00:00
OpReturn
OpFunctionEnd
Reference in New Issue Copy Permalink