raphiaura/KytyPS5
1
0
Fork 0
mirror of https://github.com/KytyPS5/KytyPS5.git synced 2026-10-20 10:56:58 +02:00
Code Issues Packages Projects Releases Wiki Activity

shader: took fixes from PR 427

Browse Source 
- apply c8f34f9's ResourceTracking portability fix
- complete V_CMP_LT_U64 and S_FLBIT_I32_B32 support
- add the sceAmprCommandBufferWaitOnAddress NID from f625a47
- add decoder and compute tests for both shader operations

Co-authored-by: Mac <me@itsmemac.com>
This commit is contained in:
nmzik
2026-08-31 23:41:24 +04:00
parent 6453bafd1d
commit 1342c6f1e1
10 changed files with 111 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tests/ShaderRecompilerComputeTests.cpp
+62
View File
@@ -12960,6 +12960,36 @@ TestCase ScalarFlbitI32B64Gpu() {
O::BUFFER_STORE_DWORD, O::S_ENDPGM}};
}
TestCase ScalarFlbitI32B32Gpu() {
using O = ShaderOpcode;
std::vector<u32> code;
AppendSMovLiteral(&code, 0, 0x00000000u);
code.push_back(EncodeSop1(0x15, 20, 0));
AppendSMovLiteral(&code, 0, 0x0000ccccu);
code.push_back(EncodeSop1(0x15, 21, 0));
AppendSMovLiteral(&code, 0, 0x80000000u);
code.push_back(EncodeSop1(0x15, 22, 0));
AppendSMovLiteral(&code, 0, 0x7fffffffu);
code.push_back(EncodeSop1(0x15, 23, 0));
AppendStoreSgpr(&code, 20, 0);
AppendStoreSgpr(&code, 21, 1);
AppendStoreSgpr(&code, 22, 2);
AppendStoreSgpr(&code, 23, 3);
AppendEnd(&code);
return {"ScalarFlbitI32B32Gpu",
code,
{},
{0xffffffffu, 16, 0, 1},
{O::S_MOV_B32, O::S_FLBIT_I32_B32, O::V_MOV_B32,
O::BUFFER_STORE_DWORD, O::S_ENDPGM}};
}
TestCase ScalarSaveExecOps() {
using O = ShaderOpcode;
@@ -16091,6 +16121,36 @@ TestCase VectorVop3CompareGtU64OnGpu() {
return test;
}
TestCase VectorVopcCompareLtU64OnGpu() {
using O = ShaderOpcode;
std::vector<u32> code;
AppendVMovU32(&code, 0, 5);
AppendVMovU32(&code, 1, 0);
AppendVMovU32(&code, 2, 1);
AppendSMovLiteral(&code, 4, 4);
AppendSMovLiteral(&code, 5, 0);
code.push_back(EncodeVopc(0xe1, 4, 0));
code.push_back(EncodeVop2(0x01, 3, InlineU32(0), 2));
AppendStoreVgpr(&code, 3, 0);
AppendSMovLiteral(&code, 4, 6);
code.push_back(EncodeVopc(0xe1, 4, 0));
code.push_back(EncodeVop2(0x01, 4, InlineU32(0), 2));
AppendStoreVgpr(&code, 4, 1);
AppendEnd(&code);
TestCase test{"VectorVopcCompareLtU64OnGpu",
code,
{},
{1, 0},
{O::V_MOV_B32, O::S_MOV_B32, O::V_CMP_LT_U64,
O::V_CNDMASK_B32, O::BUFFER_STORE_DWORD, O::S_ENDPGM}};
test.decoded_counts = {{"V_CMP_LT_U64 vcc_lo, s4, v0", 2}};
return test;
}
TestCase VectorVop3CompareEqI64OnGpu() {
using O = ShaderOpcode;
@@ -21148,6 +21208,7 @@ std::vector<TestCase> MakeCases() {
AddCase(ScalarMovB32SpecialAliases);
AddCase(ScalarQuadmaskB64);
AddCase(ScalarFlbitI32B64Gpu);
AddCase(ScalarFlbitI32B32Gpu);
AddCase(ScalarSaveExecOps);
AddCase(ScalarOrn2SaveexecUsesSourceOrNotExec);
AddCase(ScalarGetpcWritesNextInstructionPc);
@@ -21238,6 +21299,7 @@ std::vector<TestCase> MakeCases() {
AddCase(VectorVop3CompareEqI64OnGpu);
AddCase(VectorVop3CompareEqU64OnGpu);
AddCase(VectorVop3CompareGtU64OnGpu);
AddCase(VectorVopcCompareLtU64OnGpu);
AddCase(VectorVop3CompareNeU64OnGpu);
AddCase(VectorVopcCmpxNeU64CapturedExecMask);
AddCase(VectorVop3CmpxNeI64CapturedExecMask);