derek/gem5
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

arch-vega: Architected flat scratch and scratch insts

Browse Source 
Architected flat scratch is added in MI300 which store the scratch base
address in dedicated registers rather than in SGPRs. These registers are
used by scratch_ instructions. These are flat instruction which
explicitly target the private memory aperture. These instructions have a
different address calculation than global_ instructions.

This change implements architected flat scratch support, fixes the
address calculation of scratch_ instructions, and implements decodings
for some scratch_ instructions. Previous flat_ instructions which happen
to access the private memory aperture have no change in address
calculation. Since scratch_ instructions are identical to flat_
instruction except for address calculation, the decodings simply reuse
existing flat_ instruction definitions.

Change-Id: I1e1d15a2fbcc7a4a678157c35608f4f22b359e21
This commit is contained in:
Matthew Poremba
2024-05-15 12:00:47 -07:00
parent 8be5ce6fc9
commit c1803eafac
6 changed files with 160 additions and 47 deletions
Download Patch File Download Diff File Expand all files Collapse all files

24
src/arch/amdgpu/vega/gpu_decoder.cc
View File

@@ -9922,29 +9922,25 @@ namespace VegaISA
GPUStaticInst*
Decoder::decode_OP_SCRATCH__SCRATCH_LOAD_DWORD(MachInst iFmt)
{
fatal("Trying to decode instruction without a class\n");
return nullptr;
return new Inst_FLAT__FLAT_LOAD_DWORD(&iFmt->iFmt_FLAT);
}
GPUStaticInst*
Decoder::decode_OP_SCRATCH__SCRATCH_LOAD_DWORDX2(MachInst iFmt)
{
fatal("Trying to decode instruction without a class\n");
return nullptr;
return new Inst_FLAT__FLAT_LOAD_DWORDX2(&iFmt->iFmt_FLAT);
}
GPUStaticInst*
Decoder::decode_OP_SCRATCH__SCRATCH_LOAD_DWORDX3(MachInst iFmt)
{
fatal("Trying to decode instruction without a class\n");
return nullptr;
return new Inst_FLAT__FLAT_LOAD_DWORDX3(&iFmt->iFmt_FLAT);
}
GPUStaticInst*
Decoder::decode_OP_SCRATCH__SCRATCH_LOAD_DWORDX4(MachInst iFmt)
{
fatal("Trying to decode instruction without a class\n");
return nullptr;
return new Inst_FLAT__FLAT_LOAD_DWORDX4(&iFmt->iFmt_FLAT);
}
GPUStaticInst*
@@ -9977,29 +9973,25 @@ namespace VegaISA
GPUStaticInst*
Decoder::decode_OP_SCRATCH__SCRATCH_STORE_DWORD(MachInst iFmt)
{
fatal("Trying to decode instruction without a class\n");
return nullptr;
return new Inst_FLAT__FLAT_STORE_DWORD(&iFmt->iFmt_FLAT);
}
GPUStaticInst*
Decoder::decode_OP_SCRATCH__SCRATCH_STORE_DWORDX2(MachInst iFmt)
{
fatal("Trying to decode instruction without a class\n");
return nullptr;
return new Inst_FLAT__FLAT_STORE_DWORDX2(&iFmt->iFmt_FLAT);
}
GPUStaticInst*
Decoder::decode_OP_SCRATCH__SCRATCH_STORE_DWORDX3(MachInst iFmt)
{
fatal("Trying to decode instruction without a class\n");
return nullptr;
return new Inst_FLAT__FLAT_STORE_DWORDX3(&iFmt->iFmt_FLAT);
}
GPUStaticInst*
Decoder::decode_OP_SCRATCH__SCRATCH_STORE_DWORDX4(MachInst iFmt)
{
fatal("Trying to decode instruction without a class\n");
return nullptr;
return new Inst_FLAT__FLAT_STORE_DWORDX4(&iFmt->iFmt_FLAT);
}
GPUStaticInst*