From ffbee7acdcaafad1cddd4cdac060905922f93cfd Mon Sep 17 00:00:00 2001 From: Dmitry Preobrazhensky Date: Wed, 13 Jun 2018 15:32:46 +0000 Subject: [AMDGPU][MC][GFX8][GFX9] Allow LDS direct reads for BUFFER_LOAD_DWORDX2/X3/X4 See bug 37653: https://bugs.llvm.org/show_bug.cgi?id=37653 Reviewers: artem.tamazov, arsenm Differential Revision: https://reviews.llvm.org/D47885 llvm-svn: 334609 --- llvm/lib/Target/AMDGPU/BUFInstructions.td | 22 +++++++++++++++++++--- 1 file changed, 19 insertions(+), 3 deletions(-) (limited to 'llvm/lib') diff --git a/llvm/lib/Target/AMDGPU/BUFInstructions.td b/llvm/lib/Target/AMDGPU/BUFInstructions.td index 18a885c71b2..7c0b5e9d8f6 100644 --- a/llvm/lib/Target/AMDGPU/BUFInstructions.td +++ b/llvm/lib/Target/AMDGPU/BUFInstructions.td @@ -798,6 +798,22 @@ defm BUFFER_LOAD_DWORDX3 : MUBUF_Pseudo_Loads < defm BUFFER_LOAD_DWORDX4 : MUBUF_Pseudo_Loads < "buffer_load_dwordx4", VReg_128, v4i32, mubuf_load >; + +// This is not described in AMD documentation, +// but 'lds' versions of these opcodes are available +// in at least GFX8+ chips. See Bug 37653. +let SubtargetPredicate = isVI in { +defm BUFFER_LOAD_DWORDX2_LDS : MUBUF_Pseudo_Loads < + "buffer_load_dwordx2", VReg_64, v2i32, null_frag, 0, 1 +>; +defm BUFFER_LOAD_DWORDX3_LDS : MUBUF_Pseudo_Loads < + "buffer_load_dwordx3", VReg_96, untyped, null_frag, 0, 1 +>; +defm BUFFER_LOAD_DWORDX4_LDS : MUBUF_Pseudo_Loads < + "buffer_load_dwordx4", VReg_128, v4i32, null_frag, 0, 1 +>; +} + defm BUFFER_STORE_BYTE : MUBUF_Pseudo_Stores < "buffer_store_byte", VGPR_32, i32, truncstorei8_global >; @@ -1934,9 +1950,9 @@ defm BUFFER_LOAD_SBYTE : MUBUF_Real_AllAddr_Lds_vi <0x11>; defm BUFFER_LOAD_USHORT : MUBUF_Real_AllAddr_Lds_vi <0x12>; defm BUFFER_LOAD_SSHORT : MUBUF_Real_AllAddr_Lds_vi <0x13>; defm BUFFER_LOAD_DWORD : MUBUF_Real_AllAddr_Lds_vi <0x14>; -defm BUFFER_LOAD_DWORDX2 : MUBUF_Real_AllAddr_vi <0x15>; -defm BUFFER_LOAD_DWORDX3 : MUBUF_Real_AllAddr_vi <0x16>; -defm BUFFER_LOAD_DWORDX4 : MUBUF_Real_AllAddr_vi <0x17>; +defm BUFFER_LOAD_DWORDX2 : MUBUF_Real_AllAddr_Lds_vi <0x15>; +defm BUFFER_LOAD_DWORDX3 : MUBUF_Real_AllAddr_Lds_vi <0x16>; +defm BUFFER_LOAD_DWORDX4 : MUBUF_Real_AllAddr_Lds_vi <0x17>; defm BUFFER_STORE_BYTE : MUBUF_Real_AllAddr_vi <0x18>; defm BUFFER_STORE_BYTE_D16_HI : MUBUF_Real_AllAddr_vi <0x19>; defm BUFFER_STORE_SHORT : MUBUF_Real_AllAddr_vi <0x1a>; -- cgit v1.2.3