From 022598d94f2cd718f62b2b9ed02edee86757dbd1 Mon Sep 17 00:00:00 2001
From: JanLJL <jan.laukemann@fau.de>
Date: Thu, 29 Oct 2020 13:00:02 +0100
Subject: [PATCH] autodetect ISA and default uarch for ISA

---
 osaca/osaca.py              | 21 ++++++++++++++++-----
 osaca/parser/base_parser.py | 20 +++++++++++++++++++-
 tests/test_base_parser.py   | 12 ++++++++++++
 tests/test_cli.py           | 14 ++++++++++++++
 4 files changed, 61 insertions(+), 6 deletions(-)

diff --git a/osaca/osaca.py b/osaca/osaca.py
index 656ad88..9c82f3c 100755
--- a/osaca/osaca.py
+++ b/osaca/osaca.py
@@ -31,6 +31,10 @@ SUPPORTED_ARCHS = [
     'N1',
     'A64FX',
 ]
+DEFAULT_ARCHS = {
+    'aarch64': 'A64FX',
+    'x86': 'ICL',
+}
 
 
 # Stolen from pip
@@ -85,7 +89,7 @@ def create_parser(parser=None):
         '--arch',
         type=str,
         help='Define architecture (SNB, IVB, HSW, BDW, SKX, CSX, ICL, ZEN1, ZEN2, TX2, N1, '
-        'A64FX).',
+        'A64FX). If no architecture is given, OSACA assumes a default uarch for x86/AArch64.',
     )
     parser.add_argument(
         '--fixed',
@@ -164,7 +168,12 @@ def check_arguments(args, parser):
     """
     supported_import_files = ['ibench', 'asmbench']
 
-    if 'arch' in args and (args.arch is None or args.arch.upper() not in SUPPORTED_ARCHS):
+    if args.arch is None and (args.check_db or 'import_data' in args):
+        parser.error(
+            'DB check and data import cannot work with a default microarchitecture. '
+            'Please see --help for all valid architecture codes.'
+        )
+    elif args.arch is not None and args.arch.upper() not in SUPPORTED_ARCHS:
         parser.error(
             'Microarchitecture not supported. Please see --help for all valid architecture codes.'
         )
@@ -241,13 +250,15 @@ def inspect(args, output_file=sys.stdout):
     :param output_file: Define the stream for output, defaults to :class:`sys.stdout`
     :type output_file: stream, optional
     """
-    arch = args.arch
+    # Read file
+    code = args.file.read()
+
+    # Detect ISA if necessary
+    arch = args.arch if args.arch is not None else DEFAULT_ARCHS[BaseParser.detect_ISA(code)]
     isa = MachineModel.get_isa_for_arch(arch)
     verbose = args.verbose
     ignore_unknown = args.ignore_unknown
 
-    # Read file
-    code = args.file.read()
     # Parse file
     parser = get_asm_parser(arch)
     parsed_code = parser.parse_file(code)
diff --git a/osaca/parser/base_parser.py b/osaca/parser/base_parser.py
index 2706e04..f2b8fe9 100755
--- a/osaca/parser/base_parser.py
+++ b/osaca/parser/base_parser.py
@@ -1,6 +1,7 @@
 #!/usr/bin/env python3
 """Parser superclass of specific parsers."""
-
+import operator
+import re
 
 class BaseParser(object):
     # Identifiers for operand types
@@ -18,6 +19,23 @@ class BaseParser(object):
     def __init__(self):
         self.construct_parser()
 
+    @staticmethod
+    def detect_ISA(file_content):
+        """Detect the ISA of the assembly based on the used registers and return the ISA code."""
+        # Check for the amount of registers in the code to determine the ISA
+        # 1) Check for xmm, ymm, zmm, rax, rbx, rcx, and rdx registers in x86
+        heuristics_x86ATT = [r'%[xyz]mm[0-9]', r'%r[abcd]x[0-9]']
+        # 2) check for v and z vector registers and x/w general-purpose registers
+        heuristics_aarch64 = [r'[vz][0-9][0-9]?\.[0-9][0-9]?[bhsd]', r'[wx][0-9]']
+        matches = {'x86': 0, 'aarch64': 0}
+
+        for h in heuristics_x86ATT:
+            matches['x86'] += len(re.findall(h, file_content))
+        for h in heuristics_aarch64:
+            matches['aarch64'] += len(re.findall(h, file_content))
+
+        return max(matches.items(), key=operator.itemgetter(1))[0]
+
     def parse_file(self, file_content, start_line=0):
         """
         Parse assembly file. This includes *not* extracting of the marked kernel and
diff --git a/tests/test_base_parser.py b/tests/test_base_parser.py
index 7465c05..4d76352 100755
--- a/tests/test_base_parser.py
+++ b/tests/test_base_parser.py
@@ -18,6 +18,12 @@ class TestBaseParser(unittest.TestCase):
             pass
         with open(self._find_file('triad_x86_iaca.s')) as f:
             self.triad_code = f.read()
+        with open(self._find_file('triad_arm_iaca.s')) as f:
+            self.triad_code_arm = f.read()
+        with open(self._find_file('kernel_x86.s')) as f:
+            self.x86_code = f.read()
+        with open(self._find_file('kernel_aarch64.s')) as f:
+            self.aarch64_code = f.read()
 
     ##################
     # Test
@@ -59,6 +65,12 @@ class TestBaseParser(unittest.TestCase):
         with self.assertRaises(NotImplementedError):
             self.parser.normalize_imd(imd_hex_1)
 
+    def test_detect_ISA(self):
+        self.assertEqual(BaseParser.detect_ISA(self.triad_code), 'x86')
+        self.assertEqual(BaseParser.detect_ISA(self.triad_code_arm), 'aarch64')
+        self.assertEqual(BaseParser.detect_ISA(self.x86_code), 'x86')
+        self.assertEqual(BaseParser.detect_ISA(self.aarch64_code), 'aarch64')
+
     ##################
     # Helper functions
     ##################
diff --git a/tests/test_cli.py b/tests/test_cli.py
index 8e44a2e..5030b36 100755
--- a/tests/test_cli.py
+++ b/tests/test_cli.py
@@ -153,6 +153,20 @@ class TestCLI(unittest.TestCase):
                 output = StringIO()
                 osaca.run(args, output_file=output)
 
+    def test_without_arch(self):
+        # Run test kernels without --arch flag
+        parser = osaca.create_parser()
+        # x86
+        kernel_x86 = 'kernel_x86.s'
+        args = parser.parse_args([self._find_test_file(kernel_x86)])
+        output = StringIO()
+        osaca.run(args, output_file=output)
+        # AArch64
+        kernel_aarch64 = 'kernel_aarch64.s'
+        args = parser.parse_args([self._find_test_file(kernel_aarch64)])
+        osaca.run(args, output_file=output)
+
+
     ##################
     # Helper functions
     ##################