Improved Scan #855

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

Closed

devshgraphicsprogramming wants to merge 24 commits into master from improve_scan

Member

devshgraphicsprogramming commented Mar 19, 2025

Description

Testing

TODO list:


          intial changes

d4e3738

devshgraphicsprogramming commented

View reviewed changes

include/nbl/builtin/hlsl/subgroup/arithmetic_portability.hlsl Outdated Show resolved Hide resolved

keptsecret added 4 commits

March 27, 2025 15:04


          subgroup2 implementations

10d9c39


          some fixes, example

f2a281c


          changed template parameters

4622f1f


          working subgroup2 template and funcs

abfaf67

devshgraphicsprogramming commented

View reviewed changes

include/nbl/builtin/hlsl/subgroup2/arithmetic_portability.hlsl Outdated Show resolved Hide resolved

keptsecret added 6 commits

March 31, 2025 16:40


          fix reduction bug

f2d6d8a


          minor fix

eeec20a


          latest example

53ffc60


          merge master, fix conflicts

0efeb8d


          new example number


          partial spec for items per invoc =1

e88f51a

devshgraphicsprogramming commented

View reviewed changes

include/nbl/builtin/hlsl/subgroup/ballot.hlsl Outdated Show resolved Hide resolved

devshgraphicsprogramming commented

View reviewed changes

include/nbl/builtin/hlsl/subgroup2/arithmetic_portability.hlsl Outdated Show resolved Hide resolved

devshgraphicsprogramming commented

View reviewed changes

include/nbl/builtin/hlsl/subgroup2/arithmetic_portability_impl.hlsl Outdated Show resolved Hide resolved

devshgraphicsprogramming commented

View reviewed changes

include/nbl/builtin/hlsl/subgroup2/arithmetic_portability.hlsl Outdated Show resolved Hide resolved

devshgraphicsprogramming commented

View reviewed changes

include/nbl/builtin/hlsl/subgroup2/arithmetic_portability_impl.hlsl Outdated Show resolved Hide resolved

devshgraphicsprogramming commented

View reviewed changes

include/nbl/builtin/hlsl/subgroup2/arithmetic_portability_impl.hlsl Outdated Show resolved Hide resolved

devshgraphicsprogramming commented

View reviewed changes

include/nbl/builtin/hlsl/subgroup2/arithmetic_portability_impl.hlsl Outdated Show resolved Hide resolved

devshgraphicsprogramming commented

View reviewed changes

include/nbl/builtin/hlsl/subgroup2/arithmetic_portability_impl.hlsl Show resolved Hide resolved


          changes to Params, Config handling types

a8e02a3

devshgraphicsprogramming commented

View reviewed changes

include/nbl/builtin/hlsl/subgroup2/arithmetic_portability.hlsl Outdated Show resolved Hide resolved


          rework specializations for native, emulated funcs

237ac09

devshgraphicsprogramming commented

View reviewed changes

include/nbl/builtin/hlsl/subgroup2/arithmetic_portability.hlsl Outdated Show resolved Hide resolved

devshgraphicsprogramming commented

View reviewed changes

include/nbl/builtin/hlsl/subgroup2/arithmetic_portability_impl.hlsl Outdated Show resolved Hide resolved

keptsecret added 4 commits

April 11, 2025 10:25


          added OpSelect intrinsic for mix, fix mix behavior with bool

859c313


          use mix instead of ternary op

c5a3223


          fixes to subgroup2 funcs

87bca2b


          changes to handle coalesced data loads

49fd605

keptsecret added 7 commits

April 21, 2025 14:53


          merge master, fix example conflicts

4ae51a1


          fixes to inclusive_scan for coalesced

609ad85


          removed redundant code

6b692f4


          enabled handling vectors in spirv group ops with templates and enable_if

d0acb31


          added impl component wise inclusive scan for inclusive scan

fc92538


          revert to scans using consecutive data loads

8ad4843


          merge master, fix conflicts

0c4b937

devshgraphicsprogramming commented

View reviewed changes

include/nbl/builtin/hlsl/spirv_intrinsics/subgroup_arithmetic.hlsl Show resolved Hide resolved

devshgraphicsprogramming commented

View reviewed changes

include/nbl/builtin/hlsl/subgroup2/arithmetic_portability.hlsl Show resolved Hide resolved

devshgraphicsprogramming commented

View reviewed changes

include/nbl/builtin/hlsl/subgroup2/arithmetic_portability.hlsl

+                  using type_t = vector<scalar_t, _ItemsPerInvocation>;
+                  NBL_CONSTEXPR_STATIC_INLINE int32_t ItemsPerInvocation = _ItemsPerInvocation;
+                  NBL_CONSTEXPR_STATIC_INLINE bool UseNativeIntrinsics = device_capabilities_traits<device_capabilities>::shaderSubgroupArithmetic /*&& /*some heuristic for when its faster*/;

Member Author

devshgraphicsprogramming May 12, 2025

write a TODO to add a IHV enum to device_capabilities_traits so we can && !is_nvidia :P

Member Author

devshgraphicsprogramming May 12, 2025

also do a using alias for device_traits = device_capabilities_traits<device_capabilities>

devshgraphicsprogramming commented

View reviewed changes

include/nbl/builtin/hlsl/subgroup2/arithmetic_portability_impl.hlsl Show resolved Hide resolved

devshgraphicsprogramming commented

View reviewed changes

include/nbl/builtin/hlsl/subgroup2/arithmetic_portability_impl.hlsl Show resolved Hide resolved

devshgraphicsprogramming commented

View reviewed changes

include/nbl/builtin/hlsl/subgroup2/arithmetic_portability_impl.hlsl Show resolved Hide resolved

Member Author

devshgraphicsprogramming commented May 12, 2025

continued in #873

devshgraphicsprogramming closed this

devshgraphicsprogramming deleted the improve_scan branch

May 13, 2025 05:29

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet