This is an archive of the discontinued LLVM Phabricator instance.

Differential D29179

[Doc][LangRef] Fix typo-ish error in description of Masked Gather
ClosedPublic

Authored by zvi on Jan 26 2017, 7:29 AM.

Details

Reviewers
delena
Commits
rGb26530cd69d7: [Doc][LangRef] Fix typo-ish error in description of Masked Gather
rL293206: [Doc][LangRef] Fix typo-ish error in description of Masked Gather
Summary

Fix the example of equivalent expansion for when mask is all ones.

Diff Detail

Repository
rL LLVM

Event Timeline

zvi created this revision.Jan 26 2017, 7:29 AM
zvi updated this revision to Diff 85908.Jan 26 2017, 7:45 AM
delena accepted this revision.Jan 26 2017, 12:21 PM
This revision is now accepted and ready to land.Jan 26 2017, 12:21 PM
Closed by commit rL293206: [Doc][LangRef] Fix typo-ish error in description of Masked Gather (authored by zvi). · Explain WhyJan 26 2017, 12:40 PM
This revision was automatically updated to reflect the committed changes.

Revision Contents

PathSize
llvm/
trunk/
docs/
2 lines

Diff 85948

llvm/trunk/docs/LangRef.rst

  • This file is larger than 256 KB, so syntax highlighting is disabled by default.
Show First 20 Lines • Show All 11,825 Lines • ▼ Show 20 Lines
"""""""""" """"""""""
The '``llvm.masked.gather``' intrinsic is designed for conditional reading of multiple scalar values from arbitrary memory locations in a single IR operation. It is useful for targets that support vector masked gathers and allows vectorizing basic blocks with data and control divergence. Other targets may support this intrinsic differently, for example by lowering it into a sequence of scalar load operations. The '``llvm.masked.gather``' intrinsic is designed for conditional reading of multiple scalar values from arbitrary memory locations in a single IR operation. It is useful for targets that support vector masked gathers and allows vectorizing basic blocks with data and control divergence. Other targets may support this intrinsic differently, for example by lowering it into a sequence of scalar load operations.
The semantics of this operation are equivalent to a sequence of conditional scalar loads with subsequent gathering all loaded values into a single vector. The mask restricts memory access to certain lanes and facilitates vectorization of predicated basic blocks. The semantics of this operation are equivalent to a sequence of conditional scalar loads with subsequent gathering all loaded values into a single vector. The mask restricts memory access to certain lanes and facilitates vectorization of predicated basic blocks.
:: ::
%res = call <4 x double> @llvm.masked.gather.v4f64 (<4 x double*> %ptrs, i32 8, <4 x i1>%mask, <4 x double> <true, true, true, true>) %res = call <4 x double> @llvm.masked.gather.v4f64 (<4 x double*> %ptrs, i32 8, <4 x i1> <i1 true, i1 true, i1 true, i1 true>, <4 x double> undef)
;; The gather with all-true mask is equivalent to the following instruction sequence ;; The gather with all-true mask is equivalent to the following instruction sequence
%ptr0 = extractelement <4 x double*> %ptrs, i32 0 %ptr0 = extractelement <4 x double*> %ptrs, i32 0
%ptr1 = extractelement <4 x double*> %ptrs, i32 1 %ptr1 = extractelement <4 x double*> %ptrs, i32 1
%ptr2 = extractelement <4 x double*> %ptrs, i32 2 %ptr2 = extractelement <4 x double*> %ptrs, i32 2
%ptr3 = extractelement <4 x double*> %ptrs, i32 3 %ptr3 = extractelement <4 x double*> %ptrs, i32 3
%val0 = load double, double* %ptr0, align 8 %val0 = load double, double* %ptr0, align 8
▲ Show 20 Lines • Show All 915 Lines • Show Last 20 Lines