AlexIndustrial/mesa
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

aco: support 8/16-bit loads in smem_combine()

Browse Source 
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/34162>
This commit is contained in:
Rhys Perry
2025-03-11 14:11:45 +00:00
committed by Marge Bot
parent 8abb787c6b
commit 75efc218f5
1 changed files with 22 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/amd/compiler/aco_optimizer.cpp
+22 -8
View File
@@ -791,7 +791,7 @@ parse_base_offset(opt_ctx& ctx, Instruction* instr, unsigned op_index, Temp* bas
}
void
skip_smem_offset_align(opt_ctx& ctx, SMEM_instruction* smem)
skip_smem_offset_align(opt_ctx& ctx, SMEM_instruction* smem, uint32_t align)
{
bool soe = smem->operands.size() >= (!smem->definitions.empty() ? 3 : 4);
if (soe && !smem->operands[1].isConstant())
@@ -808,10 +808,11 @@ skip_smem_offset_align(opt_ctx& ctx, SMEM_instruction* smem)
if (bitwise_instr->opcode != aco_opcode::s_and_b32)
return;
if (bitwise_instr->operands[0].constantEquals(-4) &&
uint32_t mask = ~(align - 1u);
if (bitwise_instr->operands[0].constantEquals(mask) &&
bitwise_instr->operands[1].isOfType(op.regClass().type()))
op.setTemp(bitwise_instr->operands[1].getTemp());
else if (bitwise_instr->operands[1].constantEquals(-4) &&
else if (bitwise_instr->operands[1].constantEquals(mask) &&
bitwise_instr->operands[0].isOfType(op.regClass().type()))
op.setTemp(bitwise_instr->operands[0].getTemp());
}
@@ -819,9 +820,22 @@ skip_smem_offset_align(opt_ctx& ctx, SMEM_instruction* smem)
void
smem_combine(opt_ctx& ctx, aco_ptr<Instruction>& instr)
{
uint32_t align = 4;
switch (instr->opcode) {
case aco_opcode::s_load_sbyte:
case aco_opcode::s_load_ubyte:
case aco_opcode::s_buffer_load_sbyte:
case aco_opcode::s_buffer_load_ubyte: align = 1; break;
case aco_opcode::s_load_sshort:
case aco_opcode::s_load_ushort:
case aco_opcode::s_buffer_load_sshort:
case aco_opcode::s_buffer_load_ushort: align = 2; break;
default: break;
}
/* skip &-4 before offset additions: load((a + 16) & -4, 0) */
if (!instr->operands.empty())
skip_smem_offset_align(ctx, &instr->smem());
if (!instr->operands.empty() && align > 1)
skip_smem_offset_align(ctx, &instr->smem(), align);
/* propagate constants and combine additions */
if (!instr->operands.empty() && instr->operands[1].isTemp()) {
@@ -834,7 +848,7 @@ smem_combine(opt_ctx& ctx, aco_ptr<Instruction>& instr)
instr->operands[1] = Operand::c32(info.val);
} else if (parse_base_offset(ctx, instr.get(), 1, &base, &offset, true) &&
base.regClass() == s1 && offset <= ctx.program->dev.smem_offset_max &&
ctx.program->gfx_level >= GFX9 && offset % 4u == 0) {
ctx.program->gfx_level >= GFX9 && offset % align == 0) {
bool soe = smem.operands.size() >= (!smem.definitions.empty() ? 3 : 4);
if (soe) {
if (ctx.info[smem.operands.back().tempId()].is_constant_or_literal(32) &&
@@ -860,8 +874,8 @@ smem_combine(opt_ctx& ctx, aco_ptr<Instruction>& instr)
}
/* skip &-4 after offset additions: load(a & -4, 16) */
if (!instr->operands.empty())
skip_smem_offset_align(ctx, &instr->smem());
if (!instr->operands.empty() && align > 1)
skip_smem_offset_align(ctx, &instr->smem(), align);
}
Operand