[clang] [llvm] [AArch64] Add support for Cortex-R82AE and improve Cortex-R82 (PR #90440)
via cfe-commits
cfe-commits at lists.llvm.org
Mon Apr 29 01:10:44 PDT 2024
llvmbot wrote:
<!--LLVM PR SUMMARY COMMENT-->
@llvm/pr-subscribers-backend-aarch64
Author: Jonathan Thackray (jthackray)
<details>
<summary>Changes</summary>
- Fix mismatches between function parameter definitions and declarations (#<!-- -->89512)
- Revert "[llvm][RISCV] Enable trailing fences for seq-cst stores by default (#<!-- -->87376)"
- Revert "[RISCV] Support RISCV Atomics ABI attributes (#<!-- -->84597)"
- [SelectionDAG] Treat CopyFromReg as freezing the value (#<!-- -->85932)
- [DAGCombiner] Do not always fold FREEZE over BUILD_VECTOR (#<!-- -->85932)
- [AArch64] Add support for Neoverse-N3, Neoverse-V3 and Neoverse-V3AE (#<!-- -->90143)
- [clang][X86] Fix -Wundef warning in cpuid.h (#<!-- -->89842)
- Add test cases for SELECT->AND miscompiles in DAGCombiner
- [M68k] Add support for MOVEQ instruction (#<!-- -->88542)
- [Transforms] Debug values are not remapped when cloning. (#<!-- -->87747)
- [RISCV][NFC] Future-proof reference to ISA manual in RISCVInstrInfoC.td
- [DAG] visitORCommutative - fold build_pair(not(x),not(y)) -> not(build_pair(x,y)) style patterns (#<!-- -->90050)
- [NFC][OpenACC] Remove stale FIXME comment in a test
- DAG: Simplify demanded bits for truncating atomic_store (#<!-- -->90113)
- [Offload] Remove remaining `__tgt_register_requires` references (#<!-- -->90198)
- Revert "[TableGen] Ignore inaccessible memory when checking pattern flags (#<!-- -->90061)"
- [SLP]Attempt to vectorize long stores, if short one failed.
- [mlir][MemRef] Add ExtractStridedMetadataOpCollapseShapeFolder (#<!-- -->89954)
- [mlir] Mark `isa/dyn_cast/cast/...` member functions deprecated. (#<!-- -->89998)
- [clang] Add test for CWG2149 "Brace elision and array length deduction" (#<!-- -->90079)
- [libc++][ranges] LWG3984: ranges::to's recursion branch may be ill-formed (#<!-- -->87964)
- [clang-tidy][NFC] Fix broken link in documentation of cert-env33-c (#<!-- -->90216)
- [mlir] Fix -Wdeprecated-declarations of cast in VCIXToLLVMIRTranslation.cpp (NFC)
- [mlir] Add sub-byte type emulation support for `memref.collapse_shape` (#<!-- -->89962)
- [MC] Rename temporary symbols of empty name to ".L0 " (#<!-- -->89693)
- [X86] Regenerate subreg-to-reg tests with update_llc_test_checks.py
- [C++17] Support __GCC_[CON|DE]STRUCTIVE_SIZE (#<!-- -->89446)
- [AArch64][SVE2] SVE2 NBSL instruction lowering. (#<!-- -->89732)
- [libc++][ranges] Exports operator|. (#<!-- -->90071)
- [NFC] update comments from an earlier version of SuffixTree (#<!-- -->89800)
- [scudo] Reflect the allowed values for M_DECAY_TIME on Android (#<!-- -->89114)
- [DXIL] Fix build warning (#<!-- -->90226)
- [OpenMP][AIX] Use syssmt() to get the number of SMTs per physical CPU (#<!-- -->89985)
- [RISCV] Flatten the ImpliedExts table in RISCVISAInfo.cpp (#<!-- -->89975)
- [LV] Add tests showing missed propgation of versiond stride values.
- [mlir][sparse] fold sparse convert into producer linalg op. (#<!-- -->89999)
- [Arm64EC] Improve alignment mangling in arm64ec thunks. (#<!-- -->90115)
- [RISCV] Add an instruction PrettyPrinter to llvm-objdump (#<!-- -->90093)
- [APINotes] Allow annotating a C++ type as non-copyable in Swift
- [lldb] Switch to llvm::DWARFUnitHeader (#<!-- -->89808)
- [SLP]Fix PR90224: check that users of gep are all vectorized.
- [lldb] Fix typo in CumulativeSystemTimeIsValid check (#<!-- -->89680)
- [Libomptarget] Rename `libomptarget.rtl.x86_64` to `libomptarget.rtl.host` (#<!-- -->86868)
- [RISCV] Consistently use uint32_t in Disassembler decode functions. NFC
- [Driver,test] Replace CHECK-NOT: warning with -### -Werror
- [HLSL][SPIR-V] Target `directx` is required
- Revert "[mlir] Mark `isa/dyn_cast/cast/...` member functions deprecated. (#<!-- -->89998)" (#<!-- -->90250)
- [RISCV] Fix off by 1 typo in decodeVMaskReg. NFC
- Implement the DWARF 6 language and version attributes. (#<!-- -->89980)
- [AMDGPU] Support byte_sel modifier on v_cvt_sr_fp8_f32 and v_cvt_sr_bf8_f32 (#<!-- -->90244)
- [ci] Add clang project dependency for bolt testing (#<!-- -->90262)
- [NFC] [HWASan] factor out debug record annotation (#<!-- -->90252)
- [lldb][sbapi] Fix API break in SBDebugger broadcast bits (#<!-- -->90261)
- [VPlan] Also propagate versioned strides to users via sext/zext.
- [flang][cuda] Avoid to issue data transfer in device context (#<!-- -->90247)
- [WebAssembly] Add half-precision feature (#<!-- -->90248)
- [BOLT][NFC] Use getEHFrameHdrSectionName() (#<!-- -->90257)
- [alpha.webkit.UncountedCallArgsChecker] Avoid emitting warnings for Ref, RefPtr, and their variants. (#<!-- -->90153)
- [ASan][Test] Remove hardcoded linker version from test (#<!-- -->90147)
- [AArch64] Add support for Cortex-R82AE and improve Cortex-R82
---
Full diff: https://github.com/llvm/llvm-project/pull/90440.diff
8 Files Affected:
- (modified) clang/docs/ReleaseNotes.rst (+1)
- (modified) clang/test/Misc/target-invalid-cpu-note.c (+2-2)
- (modified) llvm/docs/ReleaseNotes.rst (+1-1)
- (modified) llvm/include/llvm/TargetParser/AArch64TargetParser.h (+12-1)
- (modified) llvm/lib/Target/AArch64/AArch64Processors.td (+14-1)
- (modified) llvm/lib/Target/AArch64/AArch64Subtarget.cpp (+1)
- (modified) llvm/lib/TargetParser/Host.cpp (+1)
- (modified) llvm/unittests/TargetParser/TargetParserTest.cpp (+15-2)
``````````diff
diff --git a/clang/docs/ReleaseNotes.rst b/clang/docs/ReleaseNotes.rst
index 5d4d152b2eb540..c92d480023f4d4 100644
--- a/clang/docs/ReleaseNotes.rst
+++ b/clang/docs/ReleaseNotes.rst
@@ -645,6 +645,7 @@ Arm and AArch64 Support
* Arm Cortex-A78AE (cortex-a78ae).
* Arm Cortex-A520AE (cortex-a520ae).
* Arm Cortex-A720AE (cortex-a720ae).
+ * Arm Cortex-R82AE (cortex-r82ae).
* Arm Neoverse-N3 (neoverse-n3).
* Arm Neoverse-V3 (neoverse-v3).
* Arm Neoverse-V3AE (neoverse-v3ae).
diff --git a/clang/test/Misc/target-invalid-cpu-note.c b/clang/test/Misc/target-invalid-cpu-note.c
index 21d80b7134508f..768b243b04e3a3 100644
--- a/clang/test/Misc/target-invalid-cpu-note.c
+++ b/clang/test/Misc/target-invalid-cpu-note.c
@@ -5,11 +5,11 @@
// RUN: not %clang_cc1 -triple arm64--- -target-cpu not-a-cpu -fsyntax-only %s 2>&1 | FileCheck %s --check-prefix AARCH64
// AARCH64: error: unknown target CPU 'not-a-cpu'
-// AARCH64-NEXT: note: valid target CPU values are: cortex-a34, cortex-a35, cortex-a53, cortex-a55, cortex-a510, cortex-a520, cortex-a520ae, cortex-a57, cortex-a65, cortex-a65ae, cortex-a72, cortex-a73, cortex-a75, cortex-a76, cortex-a76ae, cortex-a77, cortex-a78, cortex-a78ae, cortex-a78c, cortex-a710, cortex-a715, cortex-a720, cortex-a720ae, cortex-r82, cortex-x1, cortex-x1c, cortex-x2, cortex-x3, cortex-x4, neoverse-e1, neoverse-n1, neoverse-n2, neoverse-n3, neoverse-512tvb, neoverse-v1, neoverse-v2, neoverse-v3, neoverse-v3ae, cyclone, apple-a7, apple-a8, apple-a9, apple-a10, apple-a11, apple-a12, apple-a13, apple-a14, apple-a15, apple-a16, apple-a17, apple-m1, apple-m2, apple-m3, apple-s4, apple-s5, exynos-m3, exynos-m4, exynos-m5, falkor, saphira, kryo, thunderx2t99, thunderx3t110, thunderx, thunderxt88, thunderxt81, thunderxt83, tsv110, a64fx, carmel, ampere1, ampere1a, ampere1b, cobalt-100, grace{{$}}
+// AARCH64-NEXT: note: valid target CPU values are: cortex-a34, cortex-a35, cortex-a53, cortex-a55, cortex-a510, cortex-a520, cortex-a520ae, cortex-a57, cortex-a65, cortex-a65ae, cortex-a72, cortex-a73, cortex-a75, cortex-a76, cortex-a76ae, cortex-a77, cortex-a78, cortex-a78ae, cortex-a78c, cortex-a710, cortex-a715, cortex-a720, cortex-a720ae, cortex-r82, cortex-r82ae, cortex-x1, cortex-x1c, cortex-x2, cortex-x3, cortex-x4, neoverse-e1, neoverse-n1, neoverse-n2, neoverse-n3, neoverse-512tvb, neoverse-v1, neoverse-v2, neoverse-v3, neoverse-v3ae, cyclone, apple-a7, apple-a8, apple-a9, apple-a10, apple-a11, apple-a12, apple-a13, apple-a14, apple-a15, apple-a16, apple-a17, apple-m1, apple-m2, apple-m3, apple-s4, apple-s5, exynos-m3, exynos-m4, exynos-m5, falkor, saphira, kryo, thunderx2t99, thunderx3t110, thunderx, thunderxt88, thunderxt81, thunderxt83, tsv110, a64fx, carmel, ampere1, ampere1a, ampere1b, cobalt-100, grace{{$}}
// RUN: not %clang_cc1 -triple arm64--- -tune-cpu not-a-cpu -fsyntax-only %s 2>&1 | FileCheck %s --check-prefix TUNE_AARCH64
// TUNE_AARCH64: error: unknown target CPU 'not-a-cpu'
-// TUNE_AARCH64-NEXT: note: valid target CPU values are: cortex-a34, cortex-a35, cortex-a53, cortex-a55, cortex-a510, cortex-a520, cortex-a520ae, cortex-a57, cortex-a65, cortex-a65ae, cortex-a72, cortex-a73, cortex-a75, cortex-a76, cortex-a76ae, cortex-a77, cortex-a78, cortex-a78ae, cortex-a78c, cortex-a710, cortex-a715, cortex-a720, cortex-a720ae, cortex-r82, cortex-x1, cortex-x1c, cortex-x2, cortex-x3, cortex-x4, neoverse-e1, neoverse-n1, neoverse-n2, neoverse-n3, neoverse-512tvb, neoverse-v1, neoverse-v2, neoverse-v3, neoverse-v3ae, cyclone, apple-a7, apple-a8, apple-a9, apple-a10, apple-a11, apple-a12, apple-a13, apple-a14, apple-a15, apple-a16, apple-a17, apple-m1, apple-m2, apple-m3, apple-s4, apple-s5, exynos-m3, exynos-m4, exynos-m5, falkor, saphira, kryo, thunderx2t99, thunderx3t110, thunderx, thunderxt88, thunderxt81, thunderxt83, tsv110, a64fx, carmel, ampere1, ampere1a, ampere1b, cobalt-100, grace{{$}}
+// TUNE_AARCH64-NEXT: note: valid target CPU values are: cortex-a34, cortex-a35, cortex-a53, cortex-a55, cortex-a510, cortex-a520, cortex-a520ae, cortex-a57, cortex-a65, cortex-a65ae, cortex-a72, cortex-a73, cortex-a75, cortex-a76, cortex-a76ae, cortex-a77, cortex-a78, cortex-a78ae, cortex-a78c, cortex-a710, cortex-a715, cortex-a720, cortex-a720ae, cortex-r82, cortex-r82ae, cortex-x1, cortex-x1c, cortex-x2, cortex-x3, cortex-x4, neoverse-e1, neoverse-n1, neoverse-n2, neoverse-n3, neoverse-512tvb, neoverse-v1, neoverse-v2, neoverse-v3, neoverse-v3ae, cyclone, apple-a7, apple-a8, apple-a9, apple-a10, apple-a11, apple-a12, apple-a13, apple-a14, apple-a15, apple-a16, apple-a17, apple-m1, apple-m2, apple-m3, apple-s4, apple-s5, exynos-m3, exynos-m4, exynos-m5, falkor, saphira, kryo, thunderx2t99, thunderx3t110, thunderx, thunderxt88, thunderxt81, thunderxt83, tsv110, a64fx, carmel, ampere1, ampere1a, ampere1b, cobalt-100, grace{{$}}
// RUN: not %clang_cc1 -triple i386--- -target-cpu not-a-cpu -fsyntax-only %s 2>&1 | FileCheck %s --check-prefix X86
// X86: error: unknown target CPU 'not-a-cpu'
diff --git a/llvm/docs/ReleaseNotes.rst b/llvm/docs/ReleaseNotes.rst
index 64a69832521290..4c07abb744a238 100644
--- a/llvm/docs/ReleaseNotes.rst
+++ b/llvm/docs/ReleaseNotes.rst
@@ -70,7 +70,7 @@ Changes to the AArch64 Backend
------------------------------
* Added support for Cortex-A78AE, Cortex-A520AE, Cortex-A720AE,
- Neoverse-N3, Neoverse-V3 and Neoverse-V3AE CPUs.
+ Cortex-R82AE, Neoverse-N3, Neoverse-V3 and Neoverse-V3AE CPUs.
Changes to the AMDGPU Backend
-----------------------------
diff --git a/llvm/include/llvm/TargetParser/AArch64TargetParser.h b/llvm/include/llvm/TargetParser/AArch64TargetParser.h
index 0d1cfd152151aa..c3d033e1659139 100644
--- a/llvm/include/llvm/TargetParser/AArch64TargetParser.h
+++ b/llvm/include/llvm/TargetParser/AArch64TargetParser.h
@@ -632,7 +632,18 @@ inline constexpr CpuInfo CpuInfos[] = {
AArch64::AEK_PAUTH, AArch64::AEK_SVE2BITPERM,
AArch64::AEK_FLAGM, AArch64::AEK_PERFMON,
AArch64::AEK_PREDRES, AArch64::AEK_PROFILE})},
- {"cortex-r82", ARMV8R, AArch64::ExtensionBitset({AArch64::AEK_LSE})},
+ {"cortex-r82", ARMV8R,
+ AArch64::ExtensionBitset({AArch64::AEK_CRC, AArch64::AEK_DOTPROD,
+ AArch64::AEK_FLAGM, AArch64::AEK_FP,
+ AArch64::AEK_RCPC, AArch64::AEK_PAUTH,
+ AArch64::AEK_PERFMON, AArch64::AEK_RAS,
+ AArch64::AEK_RDM, AArch64::AEK_PREDRES})},
+ {"cortex-r82ae", ARMV8R,
+ AArch64::ExtensionBitset({AArch64::AEK_CRC, AArch64::AEK_DOTPROD,
+ AArch64::AEK_FLAGM, AArch64::AEK_FP,
+ AArch64::AEK_RCPC, AArch64::AEK_PAUTH,
+ AArch64::AEK_PERFMON, AArch64::AEK_RAS,
+ AArch64::AEK_RDM, AArch64::AEK_PREDRES})},
{"cortex-x1", ARMV8_2A,
AArch64::ExtensionBitset({AArch64::AEK_AES, AArch64::AEK_SHA2,
AArch64::AEK_FP16, AArch64::AEK_DOTPROD,
diff --git a/llvm/lib/Target/AArch64/AArch64Processors.td b/llvm/lib/Target/AArch64/AArch64Processors.td
index 8772e51bf0ab42..f2286ae17dba56 100644
--- a/llvm/lib/Target/AArch64/AArch64Processors.td
+++ b/llvm/lib/Target/AArch64/AArch64Processors.td
@@ -194,6 +194,11 @@ def TuneR82 : SubtargetFeature<"cortex-r82", "ARMProcFamily",
"Cortex-R82 ARM processors", [
FeaturePostRAScheduler]>;
+def TuneR82AE : SubtargetFeature<"cortex-r82ae", "ARMProcFamily",
+ "CortexR82AE",
+ "Cortex-R82-AE ARM processors",
+ [FeaturePostRAScheduler]>;
+
def TuneX1 : SubtargetFeature<"cortex-x1", "ARMProcFamily", "CortexX1",
"Cortex-X1 ARM processors", [
FeatureCmpBccFusion,
@@ -667,7 +672,13 @@ def ProcessorFeatures {
list<SubtargetFeature> R82 = [HasV8_0rOps, FeaturePerfMon, FeatureFullFP16,
FeatureFP16FML, FeatureSSBS, FeaturePredRes,
FeatureSB, FeatureRDM, FeatureDotProd,
- FeatureComplxNum, FeatureJS];
+ FeatureComplxNum, FeatureJS,
+ FeatureCacheDeepPersist];
+ list<SubtargetFeature> R82AE = [HasV8_0rOps, FeaturePerfMon, FeatureFullFP16,
+ FeatureFP16FML, FeatureSSBS, FeaturePredRes,
+ FeatureSB, FeatureRDM, FeatureDotProd,
+ FeatureComplxNum, FeatureJS,
+ FeatureCacheDeepPersist];
list<SubtargetFeature> X1 = [HasV8_2aOps, FeatureCrypto, FeatureFPARMv8,
FeatureNEON, FeatureRCPC, FeaturePerfMon,
FeatureSPE, FeatureFullFP16, FeatureDotProd,
@@ -854,6 +865,8 @@ def : ProcessorModel<"cortex-a720ae", NeoverseN2Model, ProcessorFeatures.A720AE,
[TuneA720AE]>;
def : ProcessorModel<"cortex-r82", CortexA55Model, ProcessorFeatures.R82,
[TuneR82]>;
+def : ProcessorModel<"cortex-r82ae", CortexA55Model, ProcessorFeatures.R82AE,
+ [TuneR82AE]>;
def : ProcessorModel<"cortex-x1", CortexA57Model, ProcessorFeatures.X1,
[TuneX1]>;
def : ProcessorModel<"cortex-x1c", CortexA57Model, ProcessorFeatures.X1C,
diff --git a/llvm/lib/Target/AArch64/AArch64Subtarget.cpp b/llvm/lib/Target/AArch64/AArch64Subtarget.cpp
index 7d34dd1c776878..747d82639a9b4f 100644
--- a/llvm/lib/Target/AArch64/AArch64Subtarget.cpp
+++ b/llvm/lib/Target/AArch64/AArch64Subtarget.cpp
@@ -143,6 +143,7 @@ void AArch64Subtarget::initializeProperties(bool HasMinSize) {
case CortexA78AE:
case CortexA78C:
case CortexR82:
+ case CortexR82AE:
case CortexX1:
PrefFunctionAlignment = Align(16);
PrefLoopAlignment = Align(32);
diff --git a/llvm/lib/TargetParser/Host.cpp b/llvm/lib/TargetParser/Host.cpp
index 0a93b06f40c248..8823ae370ed2b0 100644
--- a/llvm/lib/TargetParser/Host.cpp
+++ b/llvm/lib/TargetParser/Host.cpp
@@ -214,6 +214,7 @@ StringRef sys::detail::getHostCPUNameForARM(StringRef ProcCpuinfoContent) {
.Case("0xc18", "cortex-r8")
.Case("0xd13", "cortex-r52")
.Case("0xd15", "cortex-r82")
+ .Case("0xd14", "cortex-r82ae")
.Case("0xd02", "cortex-a34")
.Case("0xd04", "cortex-a35")
.Case("0xd03", "cortex-a53")
diff --git a/llvm/unittests/TargetParser/TargetParserTest.cpp b/llvm/unittests/TargetParser/TargetParserTest.cpp
index 75e235008b4f25..816aea44a9bc51 100644
--- a/llvm/unittests/TargetParser/TargetParserTest.cpp
+++ b/llvm/unittests/TargetParser/TargetParserTest.cpp
@@ -1390,7 +1390,20 @@ INSTANTIATE_TEST_SUITE_P(
AArch64::AEK_DOTPROD, AArch64::AEK_FP, AArch64::AEK_SIMD,
AArch64::AEK_FP16, AArch64::AEK_FP16FML, AArch64::AEK_RAS,
AArch64::AEK_RCPC, AArch64::AEK_LSE, AArch64::AEK_SB,
- AArch64::AEK_JSCVT, AArch64::AEK_FCMA, AArch64::AEK_PAUTH}),
+ AArch64::AEK_JSCVT, AArch64::AEK_FCMA, AArch64::AEK_PAUTH,
+ AArch64::AEK_FLAGM, AArch64::AEK_PERFMON,
+ AArch64::AEK_PREDRES}),
+ "8-R"),
+ ARMCPUTestParams<AArch64::ExtensionBitset>(
+ "cortex-r82ae", "armv8-r", "crypto-neon-fp-armv8",
+ AArch64::ExtensionBitset(
+ {AArch64::AEK_CRC, AArch64::AEK_RDM, AArch64::AEK_SSBS,
+ AArch64::AEK_DOTPROD, AArch64::AEK_FP, AArch64::AEK_SIMD,
+ AArch64::AEK_FP16, AArch64::AEK_FP16FML, AArch64::AEK_RAS,
+ AArch64::AEK_RCPC, AArch64::AEK_LSE, AArch64::AEK_SB,
+ AArch64::AEK_JSCVT, AArch64::AEK_FCMA, AArch64::AEK_PAUTH,
+ AArch64::AEK_FLAGM, AArch64::AEK_PERFMON,
+ AArch64::AEK_PREDRES}),
"8-R"),
ARMCPUTestParams<AArch64::ExtensionBitset>(
"cortex-x1", "armv8.2-a", "crypto-neon-fp-armv8",
@@ -1806,7 +1819,7 @@ INSTANTIATE_TEST_SUITE_P(
ARMCPUTestParams<AArch64::ExtensionBitset>::PrintToStringParamName);
// Note: number of CPUs includes aliases.
-static constexpr unsigned NumAArch64CPUArchs = 75;
+static constexpr unsigned NumAArch64CPUArchs = 76;
TEST(TargetParserTest, testAArch64CPUArchList) {
SmallVector<StringRef, NumAArch64CPUArchs> List;
``````````
</details>
https://github.com/llvm/llvm-project/pull/90440
More information about the cfe-commits
mailing list