diff --git a/tasks/task8.md b/tasks/task8.md
index 20a8a1b92..46eac39d6 100644
--- a/tasks/task8.md
+++ b/tasks/task8.md
@@ -6,6 +6,24 @@
 ## Задача
 
 - [ ] Реализовать **функцию**, основанную на тензорном алгоритме, решающую задачу достижимости между всеми парами вершин для заданного графа и заданной КС грамматики.
-  - Для преобразования грамматики в RSM использовать результаты предыдущих работ.
+  - Для преобразования грамматики в RSM использовать результаты предыдущих работ. Явно опишите **функции** преобразования CFG -> RSM и EBNF -> RSM
   - Для реализации матричных операций использовать [sciPy](https://docs.scipy.org/doc/scipy/reference/sparse.html).
+  - Необходимые функции:
+  ```python
+  def cfpq_with_tensor(
+    rsm: pyformlang.rsa.RecursiveAutomaton,
+    graph: nx.DiGraph,
+    final_nodes: set[int] = None,
+    start_nodes: set[int] = None,
+  ) -> set[tuple[int, int]]:
+    pass
+
+
+  def cfg_to_rsm(cfg: pyformlang.cfg.CFG) -> pyformlang.rsa.RecursiveAutomaton:
+    pass
+
+
+  def ebnf_to_rsm(ebnf: str) -> pyformlang.rsa.RecursiveAutomaton:
+    pass
+  ```
 - [ ] Добавить необходимые тесты.
diff --git a/tasks/task9.md b/tasks/task9.md
index 5c3c10ffa..190732a35 100644
--- a/tasks/task9.md
+++ b/tasks/task9.md
@@ -7,4 +7,14 @@
 
 - [ ] Реализовать **функцию**, основанную на алгоритме Generalized LL (работающего с RSM), решающую задачу достижимости между всеми парами вершин для заданного графа и заданной КС грамматики.
   - Для работы с графами и RSM использовать функции из предыдущих задач.
+  - Требуемая функция:
+  ```python
+  def cfpq_with_gll(
+        rsm: pyformlang.rsa.RecursiveAutomaton,
+        graph: nx.DiGraph,
+        start_nodes: set[int] = None,
+        final_nodes: set[int] = None,
+    ) -> set[tuple[int, int]]:
+    pass
+  ```
 - [ ] Добавить необходимые тесты.
diff --git a/tests/autotests/test_task8.py b/tests/autotests/test_task8.py
new file mode 100644
index 000000000..aebf0eab4
--- /dev/null
+++ b/tests/autotests/test_task8.py
@@ -0,0 +1,220 @@
+# This file contains test cases that you need to pass to get a grade
+# You MUST NOT touch anything here except ONE block below
+# You CAN modify this file IF AND ONLY IF you have found a bug and are willing to fix it
+# Otherwise, please report it
+import itertools
+import random
+from copy import deepcopy
+import cfpq_data as cd
+import networkx as nx
+import pytest
+from networkx import MultiDiGraph
+from pyformlang import cfg, rsa
+
+# Fix import statements in try block to run tests
+try:
+    from project.task2 import graph_to_nfa, regex_to_dfa
+    from project.task3 import FiniteAutomaton
+    from project.task4 import reachability_with_constraints
+    from project.task7 import cfpq_with_matrix
+    from project.task6 import cfpq_with_hellings
+    from project.task8 import cfpq_with_tensor, cfg_to_rsm, ebnf_to_rsm
+except ImportError:
+    pytestmark = pytest.mark.skip("Task 8 is not ready to test!")
+
+REGEXP_CFG: dict[str, list[cfg.CFG]] = {
+    "a": [cfg.CFG.from_text("S -> a"), cfg.CFG.from_text("S -> N B\nB -> $\nN -> a")],
+    "a*": [
+        cfg.CFG.from_text("S -> $ | a S"),
+        cfg.CFG.from_text("S -> $ | S S | a"),
+        cfg.CFG.from_text("S -> S a S | $"),
+    ],
+    "a b c": [cfg.CFG.from_text("S -> a b c"), cfg.CFG.from_text("S -> a B\nB -> b c")],
+    "a*b*": [
+        cfg.CFG.from_text("S -> S1 S2\nS2 -> $ | b S2\nS1 -> $ | a S1"),
+        cfg.CFG.from_text("S -> $ | S1 | a S\nS1 -> $ | b S1"),
+    ],
+    "(a b)*": [
+        cfg.CFG.from_text("S -> $ | a b S"),
+        cfg.CFG.from_text("S -> $ | S S1\nS1 -> a b"),
+    ],
+    "a b*c*": [
+        cfg.CFG.from_text("S -> S1 S2 S3\nS1 -> a\nS2 -> $ | S2 b\nS3 -> $ | c S3"),
+        cfg.CFG.from_text("S -> a S2 S3\nS2 -> S2 b | $\nS3 -> c | $ | S3 S3"),
+    ],
+    "(a|b|c|d|e)*": [
+        cfg.CFG.from_text("S -> $ | S1 S\nS1 -> a | b | c | d | e"),
+        cfg.CFG.from_text("S -> $ | a | b | c | d | e | S S"),
+        cfg.CFG.from_text("S -> $ | a S | b S | c S | e S | d S"),
+    ],
+    "((a | b) * c)*(d | e)": [
+        cfg.CFG.from_text(
+            "S -> S1 S2\nS1 -> S1 S1 | $ | S3 c\n S2 -> d | e\n S3 -> b S3 | $ | a S3"
+        ),
+        cfg.CFG.from_text("S -> S1 d | S1 e\nS1 -> S1 S3 c | $\nS3 -> b S3 | $ | a S3"),
+    ],
+}
+
+GRAMMARS = [
+    [
+        cfg.CFG.from_text("S -> $ | a S b | S S"),
+        cfg.CFG.from_text("S -> $ | a S b S"),
+        cfg.CFG.from_text("S -> $ | S a S b"),
+        cfg.CFG.from_text("S -> $ | a S b | S S S"),
+    ],
+    [
+        cfg.CFG.from_text("S -> $ | a S b | c S d | S S"),
+        cfg.CFG.from_text("S -> $ | a S b S | c S d S"),
+        cfg.CFG.from_text("S -> $ | S a S b | S c S d"),
+        cfg.CFG.from_text("S -> $ | a S b | c S d S | S S S"),
+    ],
+    [
+        cfg.CFG.from_text("S -> $ | S1 S S2\nS1 -> a | c\n S2 -> b | d\n S -> S S"),
+        cfg.CFG.from_text("S -> $ | S1 S S2 S\n S1 -> a | c\nS2 -> b | d"),
+        cfg.CFG.from_text("S -> $ | S a S b | S a S d | S c S d | S c S b"),
+        cfg.CFG.from_text("S -> $ | S1 S S2 | S S S\nS1 -> a | c\nS2-> b | d"),
+    ],
+    [
+        cfg.CFG.from_text("S -> S S | Se S1 Se\nSe -> $ | Se e\nS1 -> $ | a S1 b"),
+        cfg.CFG.from_text("S -> S1 | S S | e\nS1 -> $ | a S1 b"),
+        cfg.CFG.from_text("S -> S2 S | $\n S2 -> e | S1\n S1 -> $ | a S1 b"),
+        cfg.CFG.from_text("S -> $ | S1 S | e S\n S1 -> $ | a S1 b"),
+    ],
+    [
+        cfg.CFG.from_text("S -> a S | $"),
+        cfg.CFG.from_text("S -> S1 | a\nS1 -> a S1 | $"),
+    ],
+]
+
+GRAMMARS_DIFFERENT = [
+    cfg.CFG.from_text(
+        "S -> S1 | S2\nS1 -> Sab | S1 c\nSab -> $ | a Sab b\nS2 -> Sbc | a S2\nSbc -> b Sbc c"
+    ),
+    cfg.CFG.from_text("S -> a | b | S c S | S d S | e S f | g S"),
+    cfg.CFG.from_text("S -> $ | a S b | b S a | e S f | S S | c S d | f S c | f S e"),
+]
+
+EBNF_GRAMMARS = [
+    """S -> ( Sab c* ) | ( a* Sbc )
+    Sab -> a ( Sab | $ ) b
+    Sbc -> b ( Sbc | $ ) c""",
+    "S -> a | b | (S ( c | d ) S ) | ( e S f ) | ( g S )",
+    "S -> ( ( a S b ) | ( b S a ) | ( c S d ) | ( d S c ) | ( e S f ) | (f S e) )*",
+]
+
+LABELS = ["a", "b", "c", "d", "e", "f", "g", "h"]
+
+LABEL = "label"
+IS_FINAL = "is_final"
+IS_START = "is_start"
+
+
+def generate_rnd_start_and_final(graph):
+    start_nodes = set(
+        random.choices(
+            list(graph.nodes().keys()), k=random.randint(1, len(graph.nodes))
+        )
+    )
+    final_nodes = set(
+        random.choices(
+            list(graph.nodes().keys()), k=random.randint(1, len(graph.nodes))
+        )
+    )
+
+    for node, data in graph.nodes(data=True):
+        if node in start_nodes:
+            data[IS_START] = True
+        if node in final_nodes:
+            data[IS_FINAL] = True
+    return start_nodes, final_nodes
+
+
+@pytest.fixture(scope="function", params=range(5))
+def graph(request) -> MultiDiGraph:
+    n_of_nodes = random.randint(20, 40)
+    return cd.graphs.labeled_scale_free_graph(n_of_nodes, labels=LABELS)
+
+
+class TestReachabilityTensorAlgorithm:
+    @pytest.mark.parametrize(
+        "regex_str, cfg_list", REGEXP_CFG.items(), ids=lambda regexp_cfgs: regexp_cfgs
+    )
+    def test_rpq_cfpq_tensor(self, graph, regex_str, cfg_list) -> None:
+        start_nodes, final_nodes = generate_rnd_start_and_final(graph)
+
+        for cf_gram in cfg_list:
+            cfpq: set[tuple[int, int]] = cfpq_with_tensor(
+                cfg_to_rsm(cf_gram), deepcopy(graph), start_nodes, final_nodes
+            )
+            rpq: dict[int, set[int]] = reachability_with_constraints(
+                FiniteAutomaton(graph_to_nfa(graph, start_nodes, final_nodes)),
+                FiniteAutomaton(regex_to_dfa(regex_str)),
+            )
+            rpq_set = set()
+            for node_from, nodes_to in rpq.items():
+                for node_to in nodes_to:
+                    rpq_set.add((node_from, node_to))
+            assert cfpq == rpq_set
+
+    @pytest.mark.parametrize("eq_grammars", GRAMMARS, ids=lambda grammars: grammars)
+    def test_different_grammars(self, graph, eq_grammars):
+        start_nodes, final_nodes = generate_rnd_start_and_final(graph)
+        eq_cfpqs = [
+            cfpq_with_tensor(
+                cfg_to_rsm(deepcopy(cf_gram)), deepcopy(graph), start_nodes, final_nodes
+            )
+            for cf_gram in eq_grammars
+        ]
+        for a, b in itertools.combinations(eq_cfpqs, 2):
+            assert a == b
+
+    @pytest.mark.parametrize("grammar", GRAMMARS_DIFFERENT, ids=lambda g: g)
+    def test_hellings_matrix_tensor(self, graph, grammar):
+        start_nodes, final_nodes = generate_rnd_start_and_final(graph)
+        hellings = cfpq_with_hellings(
+            deepcopy(grammar), deepcopy(graph), start_nodes, final_nodes
+        )
+        matrix = cfpq_with_matrix(
+            deepcopy(grammar), deepcopy(graph), start_nodes, final_nodes
+        )
+        tensor = cfpq_with_tensor(
+            cfg_to_rsm(deepcopy(grammar)), deepcopy(graph), start_nodes, final_nodes
+        )
+        assert (hellings == matrix) and (matrix == tensor)
+
+    @pytest.mark.parametrize(
+        "cfg_grammar, ebnf_grammar",
+        (zip(GRAMMARS_DIFFERENT, EBNF_GRAMMARS)),
+        ids=lambda t: t,
+    )
+    def test_ebnf_cfg(self, graph, cfg_grammar, ebnf_grammar):
+        start_nodes, final_nodes = generate_rnd_start_and_final(graph)
+        cfg_cfpq = cfpq_with_tensor(
+            cfg_to_rsm(cfg_grammar), deepcopy(graph), start_nodes, final_nodes
+        )
+        ebnf_cfpq = cfpq_with_tensor(
+            ebnf_to_rsm(ebnf_grammar), deepcopy(graph), start_nodes, final_nodes
+        )
+        assert ebnf_cfpq == cfg_cfpq
+
+    @pytest.mark.parametrize(
+        "regex_str, cfg_list", REGEXP_CFG.items(), ids=lambda regexp_cfgs: regexp_cfgs
+    )
+    def test_cfpq_tensor(self, graph, regex_str, cfg_list):
+        start_nodes, final_nodes = generate_rnd_start_and_final(graph)
+        eq_cfpqs = [
+            cfpq_with_tensor(
+                cfg_to_rsm(deepcopy(cf_gram)), deepcopy(graph), start_nodes, final_nodes
+            )
+            for cf_gram in cfg_list
+        ]
+        eq_cfpqs.append(
+            cfpq_with_tensor(
+                ebnf_to_rsm(f"S -> {regex_str}"),
+                deepcopy(graph),
+                start_nodes,
+                final_nodes,
+            )
+        )
+        for a, b in itertools.combinations(eq_cfpqs, 2):
+            assert a == b
diff --git a/tests/autotests/test_task9.py b/tests/autotests/test_task9.py
new file mode 100644
index 000000000..e9411a2e9
--- /dev/null
+++ b/tests/autotests/test_task9.py
@@ -0,0 +1,224 @@
+# This file contains test cases that you need to pass to get a grade
+# You MUST NOT touch anything here except ONE block below
+# You CAN modify this file IF AND ONLY IF you have found a bug and are willing to fix it
+# Otherwise, please report it
+import itertools
+import random
+from copy import deepcopy
+import cfpq_data as cd
+import networkx as nx
+import pytest
+from networkx import MultiDiGraph
+from pyformlang import cfg, rsa
+
+# Fix import statements in try block to run tests
+try:
+    from project.task2 import graph_to_nfa, regex_to_dfa
+    from project.task3 import FiniteAutomaton
+    from project.task4 import reachability_with_constraints
+    from project.task7 import cfpq_with_matrix
+    from project.task6 import cfpq_with_hellings
+    from project.task8 import cfpq_with_tensor, cfg_to_rsm, ebnf_to_rsm
+    from project.task9 import cfpq_with_gll
+except ImportError:
+    pytestmark = pytest.mark.skip("Task 9 is not ready to test!")
+
+REGEXP_CFG: dict[str, list[cfg.CFG]] = {
+    "a": [cfg.CFG.from_text("S -> a"), cfg.CFG.from_text("S -> N B\nB -> $\nN -> a")],
+    "a*": [
+        cfg.CFG.from_text("S -> $ | a S"),
+        cfg.CFG.from_text("S -> $ | S S | a"),
+        cfg.CFG.from_text("S -> S a S | $"),
+    ],
+    "a b c": [cfg.CFG.from_text("S -> a b c"), cfg.CFG.from_text("S -> a B\nB -> b c")],
+    "a*b*": [
+        cfg.CFG.from_text("S -> S1 S2\nS2 -> $ | b S2\nS1 -> $ | a S1"),
+        cfg.CFG.from_text("S -> $ | S1 | a S\nS1 -> $ | b S1"),
+    ],
+    "(a b)*": [
+        cfg.CFG.from_text("S -> $ | a b S"),
+        cfg.CFG.from_text("S -> $ | S S1\nS1 -> a b"),
+    ],
+    "a b*c*": [
+        cfg.CFG.from_text("S -> S1 S2 S3\nS1 -> a\nS2 -> $ | S2 b\nS3 -> $ | c S3"),
+        cfg.CFG.from_text("S -> a S2 S3\nS2 -> S2 b | $\nS3 -> c | $ | S3 S3"),
+    ],
+    "(a|b|c|d|e)*": [
+        cfg.CFG.from_text("S -> $ | S1 S\nS1 -> a | b | c | d | e"),
+        cfg.CFG.from_text("S -> $ | a | b | c | d | e | S S"),
+        cfg.CFG.from_text("S -> $ | a S | b S | c S | e S | d S"),
+    ],
+    "((a | b) * c)*(d | e)": [
+        cfg.CFG.from_text(
+            "S -> S1 S2\nS1 -> S1 S1 | $ | S3 c\n S2 -> d | e\n S3 -> b S3 | $ | a S3"
+        ),
+        cfg.CFG.from_text("S -> S1 d | S1 e\nS1 -> S1 S3 c | $\nS3 -> b S3 | $ | a S3"),
+    ],
+}
+
+GRAMMARS = [
+    [
+        cfg.CFG.from_text("S -> $ | a S b | S S"),
+        cfg.CFG.from_text("S -> $ | a S b S"),
+        cfg.CFG.from_text("S -> $ | S a S b"),
+        cfg.CFG.from_text("S -> $ | a S b | S S S"),
+    ],
+    [
+        cfg.CFG.from_text("S -> $ | a S b | c S d | S S"),
+        cfg.CFG.from_text("S -> $ | a S b S | c S d S"),
+        cfg.CFG.from_text("S -> $ | S a S b | S c S d"),
+        cfg.CFG.from_text("S -> $ | a S b | c S d S | S S S"),
+    ],
+    [
+        cfg.CFG.from_text("S -> $ | S1 S S2\nS1 -> a | c\n S2 -> b | d\n S -> S S"),
+        cfg.CFG.from_text("S -> $ | S1 S S2 S\n S1 -> a | c\nS2 -> b | d"),
+        cfg.CFG.from_text("S -> $ | S a S b | S a S d | S c S d | S c S b"),
+        cfg.CFG.from_text("S -> $ | S1 S S2 | S S S\nS1 -> a | c\nS2-> b | d"),
+    ],
+    [
+        cfg.CFG.from_text("S -> S S | Se S1 Se\nSe -> $ | Se e\nS1 -> $ | a S1 b"),
+        cfg.CFG.from_text("S -> S1 | S S | e\nS1 -> $ | a S1 b"),
+        cfg.CFG.from_text("S -> S2 S | $\n S2 -> e | S1\n S1 -> $ | a S1 b"),
+        cfg.CFG.from_text("S -> $ | S1 S | e S\n S1 -> $ | a S1 b"),
+    ],
+    [
+        cfg.CFG.from_text("S -> a S | $"),
+        cfg.CFG.from_text("S -> S1 | a\nS1 -> a S1 | $"),
+    ],
+]
+
+GRAMMARS_DIFFERENT = [
+    cfg.CFG.from_text(
+        "S -> S1 | S2\nS1 -> Sab | S1 c\nSab -> $ | a Sab b\nS2 -> Sbc | a S2\nSbc -> b Sbc c"
+    ),
+    cfg.CFG.from_text("S -> a | b | S c S | S d S | e S f | g S"),
+    cfg.CFG.from_text("S -> $ | a S b | b S a | e S f | S S | c S d | f S c | f S e"),
+]
+
+EBNF_GRAMMARS = [
+    """S -> ( Sab c* ) | ( a* Sbc )
+    Sab -> a ( Sab | $ ) b
+    Sbc -> b ( Sbc | $ ) c""",
+    "S -> a | b | (S ( c | d ) S ) | ( e S f ) | ( g S )",
+    "S -> ( ( a S b ) | ( b S a ) | ( c S d ) | ( d S c ) | ( e S f ) | (f S e) )*",
+]
+
+LABELS = ["a", "b", "c", "d", "e", "f", "g", "h"]
+
+LABEL = "label"
+IS_FINAL = "is_final"
+IS_START = "is_start"
+
+
+def generate_rnd_start_and_final(graph):
+    start_nodes = set(
+        random.choices(
+            list(graph.nodes().keys()), k=random.randint(1, len(graph.nodes))
+        )
+    )
+    final_nodes = set(
+        random.choices(
+            list(graph.nodes().keys()), k=random.randint(1, len(graph.nodes))
+        )
+    )
+
+    for node, data in graph.nodes(data=True):
+        if node in start_nodes:
+            data[IS_START] = True
+        if node in final_nodes:
+            data[IS_FINAL] = True
+    return start_nodes, final_nodes
+
+
+@pytest.fixture(scope="function", params=range(5))
+def graph(request) -> MultiDiGraph:
+    n_of_nodes = random.randint(20, 40)
+    return cd.graphs.labeled_scale_free_graph(n_of_nodes, labels=LABELS)
+
+
+class TestReachabilityGllAlgorithm:
+    @pytest.mark.parametrize(
+        "regex_str, cfg_list", REGEXP_CFG.items(), ids=lambda regexp_cfgs: regexp_cfgs
+    )
+    def test_rpq_cfpq_gll(self, graph, regex_str, cfg_list) -> None:
+        start_nodes, final_nodes = generate_rnd_start_and_final(graph)
+
+        for cf_gram in cfg_list:
+            cfpq: set[tuple[int, int]] = cfpq_with_gll(
+                cfg_to_rsm(cf_gram), deepcopy(graph), start_nodes, final_nodes
+            )
+            rpq: dict[int, set[int]] = reachability_with_constraints(
+                FiniteAutomaton(graph_to_nfa(graph, start_nodes, final_nodes)),
+                FiniteAutomaton(regex_to_dfa(regex_str)),
+            )
+            rpq_set = set()
+            for node_from, nodes_to in rpq.items():
+                for node_to in nodes_to:
+                    rpq_set.add((node_from, node_to))
+            assert cfpq == rpq_set
+
+    @pytest.mark.parametrize("eq_grammars", GRAMMARS, ids=lambda grammars: grammars)
+    def test_different_grammars(self, graph, eq_grammars):
+        start_nodes, final_nodes = generate_rnd_start_and_final(graph)
+        eq_cfpqs = [
+            cfpq_with_gll(
+                cfg_to_rsm(deepcopy(cf_gram)), deepcopy(graph), start_nodes, final_nodes
+            )
+            for cf_gram in eq_grammars
+        ]
+        for a, b in itertools.combinations(eq_cfpqs, 2):
+            assert a == b
+
+    @pytest.mark.parametrize("grammar", GRAMMARS_DIFFERENT, ids=lambda g: g)
+    def test_hellings_matrix_tensor(self, graph, grammar):
+        start_nodes, final_nodes = generate_rnd_start_and_final(graph)
+        hellings = cfpq_with_hellings(
+            deepcopy(grammar), deepcopy(graph), start_nodes, final_nodes
+        )
+        matrix = cfpq_with_matrix(
+            deepcopy(grammar), deepcopy(graph), start_nodes, final_nodes
+        )
+        tensor = cfpq_with_tensor(
+            cfg_to_rsm(deepcopy(grammar)), deepcopy(graph), start_nodes, final_nodes
+        )
+        gll = cfpq_with_gll(
+            cfg_to_rsm(deepcopy(grammar)), deepcopy(graph), start_nodes, final_nodes
+        )
+        assert (hellings == matrix) and (matrix == tensor) and (tensor == gll)
+
+    @pytest.mark.parametrize(
+        "cfg_grammar, ebnf_grammar",
+        (zip(GRAMMARS_DIFFERENT, EBNF_GRAMMARS)),
+        ids=lambda t: t,
+    )
+    def test_ebnf_cfg(self, graph, cfg_grammar, ebnf_grammar):
+        start_nodes, final_nodes = generate_rnd_start_and_final(graph)
+        cfg_cfpq = cfpq_with_gll(
+            cfg_to_rsm(cfg_grammar), deepcopy(graph), start_nodes, final_nodes
+        )
+        ebnf_cfpq = cfpq_with_gll(
+            ebnf_to_rsm(ebnf_grammar), deepcopy(graph), start_nodes, final_nodes
+        )
+        assert ebnf_cfpq == cfg_cfpq
+
+    @pytest.mark.parametrize(
+        "regex_str, cfg_list", REGEXP_CFG.items(), ids=lambda regexp_cfgs: regexp_cfgs
+    )
+    def test_cfpq_gll(self, graph, regex_str, cfg_list):
+        start_nodes, final_nodes = generate_rnd_start_and_final(graph)
+        eq_cfpqs = [
+            cfpq_with_gll(
+                cfg_to_rsm(deepcopy(cf_gram)), deepcopy(graph), start_nodes, final_nodes
+            )
+            for cf_gram in cfg_list
+        ]
+        eq_cfpqs.append(
+            cfpq_with_gll(
+                ebnf_to_rsm(f"S -> {regex_str}"),
+                deepcopy(graph),
+                start_nodes,
+                final_nodes,
+            )
+        )
+        for a, b in itertools.combinations(eq_cfpqs, 2):
+            assert a == b