Features/expand microarch for aarch64 (#13780)

* Add process to determine aarch64 microarchitecture

* add microarchitectures for thunderx2 and a64fx

* Add optimize flags for gcc on aarch64 family processors thunderx2 and a64fx.

* Add optimize flags for clang on aarch64 family processors thunderx2 and a64fx

* Add testing for thunderx2 and a64fx microarchitectures

lib/spack/llnl/util/cpu/detect.py

@@ -223,3 +223,15 @@ def compatibility_check_for_x86_64(info, target):
     return (target == arch_root or arch_root in target.ancestors) \
         and (target.vendor == vendor or target.vendor == 'generic') \
         and target.features.issubset(features)
+
+
+@compatibility_check(architecture_family='aarch64')
+def compatibility_check_for_aarch64(info, target):
+    basename = 'aarch64'
+    features = set(info.get('Features', '').split())
+    vendor = info.get('CPU implementer', 'generic')
+
+    arch_root = targets[basename]
+    return (target == arch_root or arch_root in target.ancestors) \
+        and (target.vendor == vendor or target.vendor == 'generic') \
+        and target.features.issubset(features)

lib/spack/llnl/util/cpu/microarchitectures.json

@@ -1161,6 +1161,97 @@
         }
       }
     },
+    "thunderx2": {
+      "from": "aarch64",
+      "vendor": "0x43",
+      "features": [
+        "fp",
+        "asimd",
+        "evtstrm",
+        "aes",
+        "pmull",
+        "sha1",
+        "sha2",
+        "crc32",
+        "atomics",
+        "cpuid",
+        "asimdrdm"
+      ],
+      "compilers": {
+        "gcc": [
+          {
+            "versions": "4.8:4.8.9",
+            "flags": "-march=armv8-a"
+          },
+          {
+            "versions": "4.9:5.9",
+            "flags": "-march=armv8-a+crc+crypto"
+          },
+          {
+            "versions": "6:6.9",
+            "flags": "-march=armv8.1-a+crc+crypto"
+          },
+          {
+            "versions": "7:",
+            "flags": "-mcpu=thunderx2t99"
+          }
+        ],
+        "clang": {
+          "versions": ":",
+          "flags": "-march=armv8-a -mcpu=generic"
+        }
+      }
+    },
+    "a64fx": {
+      "from": "aarch64",
+      "vendor": "0x46",
+      "features": [
+        "fp",
+        "asimd",
+        "evtstrm",
+        "aes",
+        "pmull",
+        "sha1",
+        "sha2",
+        "crc32",
+        "atomics",
+        "cpuid",
+        "asimdrdm",
+        "fphp",
+        "asimdhp",
+        "fcma",
+        "dcpop",
+        "sve"
+      ],
+      "compilers": {
+        "gcc": [
+          {
+            "versions": "4.8:4.8.9",
+            "flags": "-march=armv8-a"
+          },
+          {
+            "versions": "4.9:5.9",
+            "flags": "-march=armv8-a+crc+crypto"
+          },
+          {
+            "versions": "6:6.9",
+            "flags": "-march=armv8.1-a+crc+crypto"
+          },
+          {
+            "versions": "7:7.9",
+            "flags": "-arch=armv8.2a+crc+crypt+fp16"
+          },
+          {
+            "versions": "8:",
+            "flags": "-arch=armv8.2a+crc+aes+sh2+fp16+sve -msve-vector-bits=512"
+          }
+        ],
+        "clang": {
+          "versions": ":",
+          "flags": "-march=armv8-a -mcpu=generic"
+        }
+      }
+    },
     "arm": {
       "from": null,
       "vendor": "generic",

lib/spack/spack/test/data/targets/linux-centos7-thunderx2

@@ -0,0 +1,8 @@
+processor       : 0
+BogoMIPS        : 400.00
+Features        : fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics cpuid asimdrdm
+CPU implementer : 0x43
+CPU architecture: 8
+CPU variant     : 0x1
+CPU part        : 0x0af
+CPU revision    : 1

lib/spack/spack/test/llnl/util/cpu.py

@@ -32,6 +32,7 @@
     'linux-scientific7-piledriver',
     'linux-rhel6-piledriver',
     'linux-centos7-power8le',
+    'linux-centos7-thunderx2',
     'darwin-mojave-ivybridge',
     'darwin-mojave-haswell',
     'darwin-mojave-skylake',
@@ -121,6 +122,8 @@ def test_equality(supported_target):
     ('piledriver <= steamroller', True),
     ('zen2 >= zen', True),
     ('zen >= zen', True),
+    ('aarch64 <= thunderx2', True),
+    ('aarch64 <= a64fx', True),
     # Test unrelated microarchitectures
     ('power8 < skylake', False),
     ('power8 <= skylake', False),
@@ -205,12 +208,15 @@ def test_target_json_schema():
     ('nehalem', 'gcc', '4.9.3', '-march=nehalem -mtune=nehalem'),
     ('nehalem', 'gcc', '4.8.5', '-march=corei7 -mtune=corei7'),
     ('sandybridge', 'gcc', '4.8.5', '-march=corei7-avx -mtune=corei7-avx'),
+    ('thunderx2', 'gcc', '4.8.5', '-march=armv8-a'),
+    ('thunderx2', 'gcc', '4.9.3', '-march=armv8-a+crc+crypto'),
     # Test Clang / LLVM
     ('sandybridge', 'clang', '3.9.0', '-march=x86-64 -mcpu=sandybridge'),
     ('icelake', 'clang', '6.0.0', '-march=x86-64 -mcpu=icelake'),
     ('icelake', 'clang', '8.0.0', '-march=x86-64 -mcpu=icelake-client'),
     ('zen2', 'clang', '9.0.0', '-march=x86-64 -mcpu=znver2'),
     ('power9le', 'clang', '8.0.0', '-march=ppc64le -mcpu=pwr9'),
+    ('thunderx2', 'clang', '6.0.0', '-march=armv8-a -mcpu=generic'),
     # Test Intel on Intel CPUs
     ('sandybridge', 'intel', '17.0.2', '-march=corei7-avx -mtune=corei7-avx'),
     ('sandybridge', 'intel', '18.0.5',