diff --git a/slothy/targets/aarch64/aarch64_neon.py b/slothy/targets/aarch64/aarch64_neon.py
index d84e83f3..79c24b28 100644
--- a/slothy/targets/aarch64/aarch64_neon.py
+++ b/slothy/targets/aarch64/aarch64_neon.py
@@ -1665,12 +1665,32 @@ class neg(AArch64BasicArithmetic): # pylint: disable=missing-docstring,invalid-n
     inputs = ["Xa"]
     outputs = ["Xd"]
 
+class negs(AArch64BasicArithmetic): # pylint: disable=missing-docstring,invalid-name
+    pattern = "negs <Xd>, <Xa>"
+    inputs = ["Xa"]
+    outputs = ["Xd"]
+    modifiesFlags=True
+
 class ngc_zero(AArch64BasicArithmetic): # pylint: disable=missing-docstring,invalid-name
     pattern = "ngc <Xd>, xzr"
     inputs = []
     outputs = ["Xd"]
     dependsOnFlags=True
 
+class ngcs(AArch64BasicArithmetic): # pylint: disable=missing-docstring,invalid-name
+    pattern = "ngcs <Xd>, <Xa>"
+    inputs = ["Xa"]
+    outputs = ["Xd"]
+    modifiesFlags=True
+    dependsOnFlags=True
+
+class ngcs_zero(AArch64BasicArithmetic): # pylint: disable=missing-docstring,invalid-name
+    pattern = "ngcs <Xd>, xzr"
+    inputs = []
+    outputs = ["Xd"]
+    modifiesFlags=True
+    dependsOnFlags=True
+
 class adds(AArch64BasicArithmetic): # pylint: disable=missing-docstring,invalid-name
     pattern = "adds <Xd>, <Xa>, <imm>"
     inputs = ["Xa"]
@@ -1720,6 +1740,13 @@ class sbcs_zero(AArch64BasicArithmetic): # pylint: disable=missing-docstring,inv
     modifiesFlags=True
     dependsOnFlags=True
 
+class sbcs_to_zero(AArch64BasicArithmetic): # pylint: disable=missing-docstring,invalid-name
+    pattern = "sbcs xzr, <Xa>, <Xb>"
+    inputs = ["Xa", "Xb"]
+    outputs = []
+    modifiesFlags=True
+    dependsOnFlags=True
+
 class sbcs_zero_to_zero(AArch64BasicArithmetic): # pylint: disable=missing-docstring,invalid-name
     pattern = "sbcs xzr, <Xa>, xzr"
     inputs = ["Xa"]
@@ -1980,6 +2007,12 @@ class csel_xzr_ne(AArch64ConditionalSelect): # pylint: disable=missing-docstring
     outputs = ["Xd"]
     dependsOnFlags=True
 
+class csel_xzr2_ne(AArch64ConditionalSelect): # pylint: disable=missing-docstring,invalid-name
+    pattern = "csel <Xd>, xzr, <Xe>, <flag>"
+    inputs = ["Xe"]
+    outputs = ["Xd"]
+    dependsOnFlags=True
+
 class csel_ne(AArch64ConditionalSelect): # pylint: disable=missing-docstring,invalid-name
     pattern = "csel <Xd>, <Xe>, <Xf>, <flag>"
     inputs = ["Xe", "Xf"]
@@ -2125,6 +2158,11 @@ class tst_xform(Tst): # pylint: disable=missing-docstring,invalid-name
     inputs = ["Xa", "Xb"]
     modifiesFlags=True
 
+class cmp(Tst): # pylint: disable=missing-docstring,invalid-name
+    pattern = "cmp <Xa>, <Xb>"
+    inputs = ["Xa","Xb"]
+    modifiesFlags=True
+
 class cmp_xzr(Tst): # pylint: disable=missing-docstring,invalid-name
     pattern = "cmp <Xa>, xzr"
     inputs = ["Xa"]
diff --git a/slothy/targets/aarch64/neoverse_n1_experimental.py b/slothy/targets/aarch64/neoverse_n1_experimental.py
index 76dd0b21..ef7c62a6 100644
--- a/slothy/targets/aarch64/neoverse_n1_experimental.py
+++ b/slothy/targets/aarch64/neoverse_n1_experimental.py
@@ -121,6 +121,7 @@ def get_min_max_objective(slothy):
     Tst                       : ExecutionUnit.I(),
     AArch64ShiftedArithmetic  : ExecutionUnit.M(),
     Fmov                      : ExecutionUnit.M(),
+    umull_wform               : ExecutionUnit.M(),
     (AArch64HighMultiply,
      AArch64Multiply) : ExecutionUnit.M(),
     vdup                     : ExecutionUnit.M(),
@@ -158,6 +159,7 @@ def get_min_max_objective(slothy):
     (AArch64HighMultiply)      : 4,
     (AArch64Multiply)          : 3,
     (vdup)                     : 1,
+    umull_wform                : 1,
 }
 
 default_latencies = {
@@ -193,6 +195,7 @@ def get_min_max_objective(slothy):
     AArch64HighMultiply       : 5,
     AArch64Multiply           : 4,
     (vdup)                    : 3,
+    umull_wform               : 2,
 }
 
 def get_latency(src, out_idx, dst):