Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Add --relax-float-ops and --convert-relaxed-to-half #2808

Merged
merged 1 commit into from
Sep 3, 2019
Merged

Add --relax-float-ops and --convert-relaxed-to-half #2808

merged 1 commit into from
Sep 3, 2019

Conversation

greg-lunarg
Copy link
Contributor

The first pass applies the RelaxedPrecision decoration to all executable
instructions with float32 based type results. The second pass converts
all executable instructions with RelaxedPrecision result to the equivalent
float16 type, inserting converts where necessary.

@greg-lunarg greg-lunarg force-pushed the half4 branch 2 times, most recently from 11e20fb to e0b803f Compare August 21, 2019 23:47
@s-perron s-perron self-requested a review August 22, 2019 13:33
s-perron
s-perron requested changes Aug 22, 2019
View reviewed changes
Copy link
Collaborator

@s-perron s-perron left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Here is the first part of my review. Something came up and will not finish it today.

include/spirv-tools/optimizer.hpp Outdated Show resolved Hide resolved
source/opt/convert_to_half_pass.cpp Outdated Show resolved Hide resolved
source/opt/convert_to_half_pass.cpp Outdated Show resolved Hide resolved
source/opt/convert_to_half_pass.cpp Outdated Show resolved Hide resolved
source/opt/convert_to_half_pass.cpp Show resolved Hide resolved
source/opt/convert_to_half_pass.cpp Outdated Show resolved Hide resolved
source/opt/convert_to_half_pass.cpp Outdated Show resolved Hide resolved
@greg-lunarg greg-lunarg force-pushed the half4 branch 2 times, most recently from 7dcf46b to c79f657 Compare August 28, 2019 23:54
s-perron
s-perron reviewed Aug 29, 2019
View reviewed changes
tools/opt/opt.cpp Outdated Show resolved Hide resolved
s-perron
s-perron reviewed Aug 29, 2019
View reviewed changes
tools/opt/opt.cpp Outdated Show resolved Hide resolved
s-perron
s-perron requested changes Aug 29, 2019
View reviewed changes
Copy link
Collaborator

@s-perron s-perron left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I have still not looked at the convert to half pass, but I have finished with the rest.

source/opt/relax_float_ops_pass.h Outdated Show resolved Hide resolved
source/opt/ir_builder.h Outdated Show resolved Hide resolved
source/opt/convert_to_half_pass.h Outdated Show resolved Hide resolved
source/opt/convert_to_half_pass.h Outdated Show resolved Hide resolved
source/opt/convert_to_half_pass.cpp Outdated Show resolved Hide resolved
source/opt/convert_to_half_pass.cpp Outdated Show resolved Hide resolved
source/opt/relax_float_ops_pass.cpp Outdated Show resolved Hide resolved
source/opt/relax_float_ops_pass.cpp Outdated Show resolved Hide resolved
@greg-lunarg
Copy link
Contributor Author

I think I have addressed all concerns to this point.

@greg-lunarg greg-lunarg force-pushed the half4 branch 5 times, most recently from 90c1d57 to 1c79ee3 Compare September 2, 2019 22:08
@greg-lunarg
Add --relax-float-ops and --convert-relaxed-to-half
7b254d2 
The first pass applies the RelaxedPrecision decoration to all executable
instructions with float32 based type results. The second pass converts
all executable instructions with RelaxedPrecision result to the equivalent
float16 type, inserting converts where necessary.
@s-perron s-perron merged commit d11725b into KhronosGroup:master Sep 3, 2019
dneto0 pushed a commit to dneto0/SPIRV-Tools that referenced this pull request Sep 14, 2024
@zoddicus
Rolling 4 dependencies and update known_failures (KhronosGroup#803)
4baa46a 
Roll third_party/glslang/ 56f61cc..664ad41 (2 commits)

KhronosGroup/glslang@56f61cc...664ad41

$ git log 56f61cc..664ad41 --date=short --no-merges --format='%ad %ae %s'
2019-09-05 cepheus Fix KhronosGroup#1879: Check for valid variable before checking for unsized arrays.
2019-09-04 greg Update spirv-tools and spriv-headers known good.

Roll third_party/googletest/ 565f1b848..3f05f651a (3 commits)

google/googletest@565f1b8...3f05f65

$ git log 565f1b848..3f05f651a --date=short --no-merges --format='%ad %ae %s'
2019-09-05 absl-team Googletest export
2019-09-05 absl-team Googletest export
2019-08-29 krystian.kuzniarek Googletest export

Roll third_party/spirv-cross/ a06997a6a..b32a1b415 (16 commits)

KhronosGroup/SPIRV-Cross@a06997a...b32a1b4

$ git log a06997a6a..b32a1b415 --date=short --no-merges --format='%ad %ae %s'
2019-09-06 wadetb MSL: Fix array copies to/from interpolators
2019-09-06 post Run format_all.sh.
2019-09-05 post Refactor into stronger types in public API.
2019-09-06 post Add dynamic offsets to C API.
2019-09-06 post Fix some issues on certain compilers.
2019-09-05 cdavis MSL: Support dynamic offsets for buffers in argument buffers.
2019-09-04 cdavis MSL: Force storage images on iOS to use discrete descriptors.
2019-09-05 lifeng.pan Fix ParsedIR::mark_used_as_array_length(uint32_t id)
2019-09-04 post Add test case for interlocks in control flow.
2019-09-04 post Make sure not to propagate loads outside interlock region.
2019-09-04 post Add interlock test for split functions doing begin/end.
2019-09-04 post Deal with complex interlock cases in GLSL.
2019-09-04 post Add test shader for simple case of interlocked callstack.
2019-09-04 post Deal with call stacks when analyzing access.
2019-09-04 post Analyze complex cases for fragment interlocks.
2019-08-04 cdavis Support the SPV_EXT_fragment_shader_interlock extension.

Roll third_party/spirv-tools/ b54d950..76261e2 (10 commits)

KhronosGroup/SPIRV-Tools@b54d950...76261e2

$ git log b54d950..76261e2 --date=short --no-merges --format='%ad %ae %s'
2019-09-06 stevenperron Replace CubeFaceCoord and CubeFaceIndexAMD (KhronosGroup#2840)
2019-09-05 stevenperron Fold Min, Max, and Clamp instructions. (KhronosGroup#2836)
2019-09-05 stevenperron Replace uses of SPV_AMD_shader_trinary_minmax extension (KhronosGroup#2835)
2019-09-04 zoddicus For WebGPU<->Vulkan optimization, set correct execution environment (KhronosGroup#2834)
2019-09-04 40687079+rumblehhh Export SPIRV-Tools targets on installation (KhronosGroup#2785)
2019-09-04 jmadill GN: Add Chromium GoogleTest deps. (KhronosGroup#2832)
2019-09-03 stevenperron Upadate CHANGES
2019-09-03 greg Instrument: Be sure Float16 capability on when generating float16 null (KhronosGroup#2831)
2019-09-03 greg Add --relax-float-ops and --convert-relaxed-to-half (KhronosGroup#2808)
2019-09-03 jmadill GN: Make SPIRV-Tools target use public_deps. (KhronosGroup#2828)

Created with:
  roll-dep third_party/effcee third_party/glslang third_party/googletest third_party/re2 third_party/spirv-cross third_party/spirv-headers third_party/spirv-tools
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@s-perron s-perron s-perron approved these changes

Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
2 participants
@greg-lunarg @s-perron