diff --git a/src/compiler/backend/mips/code-generator-mips.cc b/src/compiler/backend/mips/code-generator-mips.cc
index d626507b3f..b67f403f0e 100644
--- a/src/compiler/backend/mips/code-generator-mips.cc
+++ b/src/compiler/backend/mips/code-generator-mips.cc
@@ -2436,6 +2436,12 @@ CodeGenerator::CodeGenResult CodeGenerator::AssembleArchInstruction(
                  i.InputSimd128Register(0));
       break;
     }
+    case kMipsI32x4Abs: {
+      CpuFeatureScope msa_scope(tasm(), MIPS_SIMD);
+      __ asub_s_w(i.OutputSimd128Register(), i.InputSimd128Register(0),
+                  kSimd128RegZero);
+      break;
+    }
     case kMipsI16x8Splat: {
       CpuFeatureScope msa_scope(tasm(), MIPS_SIMD);
       __ fill_h(i.OutputSimd128Register(), i.InputRegister(0));
@@ -2597,6 +2603,12 @@ CodeGenerator::CodeGenResult CodeGenerator::AssembleArchInstruction(
                   i.InputSimd128Register(0));
       break;
     }
+    case kMipsI16x8Abs: {
+      CpuFeatureScope msa_scope(tasm(), MIPS_SIMD);
+      __ asub_s_h(i.OutputSimd128Register(), i.InputSimd128Register(0),
+                  kSimd128RegZero);
+      break;
+    }
     case kMipsI8x16Splat: {
       CpuFeatureScope msa_scope(tasm(), MIPS_SIMD);
       __ fill_b(i.OutputSimd128Register(), i.InputRegister(0));
@@ -2758,6 +2770,12 @@ CodeGenerator::CodeGenResult CodeGenerator::AssembleArchInstruction(
                   i.InputSimd128Register(0));
       break;
     }
+    case kMipsI8x16Abs: {
+      CpuFeatureScope msa_scope(tasm(), MIPS_SIMD);
+      __ asub_s_b(i.OutputSimd128Register(), i.InputSimd128Register(0),
+                  kSimd128RegZero);
+      break;
+    }
     case kMipsS128And: {
       CpuFeatureScope msa_scope(tasm(), MIPS_SIMD);
       __ and_v(i.OutputSimd128Register(), i.InputSimd128Register(0),
diff --git a/src/compiler/backend/mips/instruction-codes-mips.h b/src/compiler/backend/mips/instruction-codes-mips.h
index c3458e7b91..0a37dd7068 100644
--- a/src/compiler/backend/mips/instruction-codes-mips.h
+++ b/src/compiler/backend/mips/instruction-codes-mips.h
@@ -203,6 +203,7 @@ namespace compiler {
   V(MipsI32x4GeS)                  \
   V(MipsI32x4GtU)                  \
   V(MipsI32x4GeU)                  \
+  V(MipsI32x4Abs)                  \
   V(MipsI16x8Splat)                \
   V(MipsI16x8ExtractLaneU)         \
   V(MipsI16x8ExtractLaneS)         \
@@ -230,6 +231,7 @@ namespace compiler {
   V(MipsI16x8GtU)                  \
   V(MipsI16x8GeU)                  \
   V(MipsI16x8RoundingAverageU)     \
+  V(MipsI16x8Abs)                  \
   V(MipsI8x16Splat)                \
   V(MipsI8x16ExtractLaneU)         \
   V(MipsI8x16ExtractLaneS)         \
@@ -256,6 +258,7 @@ namespace compiler {
   V(MipsI8x16GtU)                  \
   V(MipsI8x16GeU)                  \
   V(MipsI8x16RoundingAverageU)     \
+  V(MipsI8x16Abs)                  \
   V(MipsS128And)                   \
   V(MipsS128Or)                    \
   V(MipsS128Xor)                   \
diff --git a/src/compiler/backend/mips/instruction-scheduler-mips.cc b/src/compiler/backend/mips/instruction-scheduler-mips.cc
index 64d7d27857..81bbfbbfb9 100644
--- a/src/compiler/backend/mips/instruction-scheduler-mips.cc
+++ b/src/compiler/backend/mips/instruction-scheduler-mips.cc
@@ -137,6 +137,7 @@ int InstructionScheduler::GetTargetInstructionFlags(
     case kMipsI16x8UConvertI32x4:
     case kMipsI16x8UConvertI8x16High:
     case kMipsI16x8UConvertI8x16Low:
+    case kMipsI16x8Abs:
     case kMipsI32x4Add:
     case kMipsI32x4AddHoriz:
     case kMipsI32x4Eq:
@@ -164,6 +165,7 @@ int InstructionScheduler::GetTargetInstructionFlags(
     case kMipsI32x4UConvertF32x4:
     case kMipsI32x4UConvertI16x8High:
     case kMipsI32x4UConvertI16x8Low:
+    case kMipsI32x4Abs:
     case kMipsI8x16Add:
     case kMipsI8x16AddSaturateS:
     case kMipsI8x16AddSaturateU:
@@ -192,6 +194,7 @@ int InstructionScheduler::GetTargetInstructionFlags(
     case kMipsI8x16SubSaturateS:
     case kMipsI8x16SubSaturateU:
     case kMipsI8x16UConvertI16x8:
+    case kMipsI8x16Abs:
     case kMipsIns:
     case kMipsLsa:
     case kMipsMaddD:
diff --git a/src/compiler/backend/mips/instruction-selector-mips.cc b/src/compiler/backend/mips/instruction-selector-mips.cc
index 2f15078874..cd87f36913 100644
--- a/src/compiler/backend/mips/instruction-selector-mips.cc
+++ b/src/compiler/backend/mips/instruction-selector-mips.cc
@@ -2169,6 +2169,7 @@ void InstructionSelector::VisitInt64AbsWithOverflow(Node* node) {
   V(I32x4GeS, kMipsI32x4GeS)                           \
   V(I32x4GtU, kMipsI32x4GtU)                           \
   V(I32x4GeU, kMipsI32x4GeU)                           \
+  V(I32x4Abs, kMipsI32x4Abs)                           \
   V(I16x8Add, kMipsI16x8Add)                           \
   V(I16x8AddSaturateS, kMipsI16x8AddSaturateS)         \
   V(I16x8AddSaturateU, kMipsI16x8AddSaturateU)         \
@@ -2190,6 +2191,7 @@ void InstructionSelector::VisitInt64AbsWithOverflow(Node* node) {
   V(I16x8SConvertI32x4, kMipsI16x8SConvertI32x4)       \
   V(I16x8UConvertI32x4, kMipsI16x8UConvertI32x4)       \
   V(I16x8RoundingAverageU, kMipsI16x8RoundingAverageU) \
+  V(I16x8Abs, kMipsI16x8Abs)                           \
   V(I8x16Add, kMipsI8x16Add)                           \
   V(I8x16AddSaturateS, kMipsI8x16AddSaturateS)         \
   V(I8x16AddSaturateU, kMipsI8x16AddSaturateU)         \
@@ -2210,6 +2212,7 @@ void InstructionSelector::VisitInt64AbsWithOverflow(Node* node) {
   V(I8x16RoundingAverageU, kMipsI8x16RoundingAverageU) \
   V(I8x16SConvertI16x8, kMipsI8x16SConvertI16x8)       \
   V(I8x16UConvertI16x8, kMipsI8x16UConvertI16x8)       \
+  V(I8x16Abs, kMipsI8x16Abs)                           \
   V(S128And, kMipsS128And)                             \
   V(S128Or, kMipsS128Or)                               \
   V(S128Xor, kMipsS128Xor)                             \
diff --git a/src/compiler/backend/mips64/code-generator-mips64.cc b/src/compiler/backend/mips64/code-generator-mips64.cc
index 7491cbddf6..05e737556a 100644
--- a/src/compiler/backend/mips64/code-generator-mips64.cc
+++ b/src/compiler/backend/mips64/code-generator-mips64.cc
@@ -2490,6 +2490,12 @@ CodeGenerator::CodeGenResult CodeGenerator::AssembleArchInstruction(
                  i.InputSimd128Register(0));
       break;
     }
+    case kMips64I32x4Abs: {
+      CpuFeatureScope msa_scope(tasm(), MIPS_SIMD);
+      __ asub_s_w(i.OutputSimd128Register(), i.InputSimd128Register(0),
+                  kSimd128RegZero);
+      break;
+    }
     case kMips64I16x8Splat: {
       CpuFeatureScope msa_scope(tasm(), MIPS_SIMD);
       __ fill_h(i.OutputSimd128Register(), i.InputRegister(0));
@@ -2651,6 +2657,12 @@ CodeGenerator::CodeGenResult CodeGenerator::AssembleArchInstruction(
                   i.InputSimd128Register(0));
       break;
     }
+    case kMips64I16x8Abs: {
+      CpuFeatureScope msa_scope(tasm(), MIPS_SIMD);
+      __ asub_s_h(i.OutputSimd128Register(), i.InputSimd128Register(0),
+                  kSimd128RegZero);
+      break;
+    }
     case kMips64I8x16Splat: {
       CpuFeatureScope msa_scope(tasm(), MIPS_SIMD);
       __ fill_b(i.OutputSimd128Register(), i.InputRegister(0));
@@ -2812,6 +2824,12 @@ CodeGenerator::CodeGenResult CodeGenerator::AssembleArchInstruction(
                   i.InputSimd128Register(0));
       break;
     }
+    case kMips64I8x16Abs: {
+      CpuFeatureScope msa_scope(tasm(), MIPS_SIMD);
+      __ asub_s_b(i.OutputSimd128Register(), i.InputSimd128Register(0),
+                  kSimd128RegZero);
+      break;
+    }
     case kMips64S128And: {
       CpuFeatureScope msa_scope(tasm(), MIPS_SIMD);
       __ and_v(i.OutputSimd128Register(), i.InputSimd128Register(0),
diff --git a/src/compiler/backend/mips64/instruction-codes-mips64.h b/src/compiler/backend/mips64/instruction-codes-mips64.h
index bed1080cbe..c752381c8c 100644
--- a/src/compiler/backend/mips64/instruction-codes-mips64.h
+++ b/src/compiler/backend/mips64/instruction-codes-mips64.h
@@ -233,6 +233,7 @@ namespace compiler {
   V(Mips64I32x4GeS)                          \
   V(Mips64I32x4GtU)                          \
   V(Mips64I32x4GeU)                          \
+  V(Mips64I32x4Abs)                          \
   V(Mips64I16x8Splat)                        \
   V(Mips64I16x8ExtractLaneU)                 \
   V(Mips64I16x8ExtractLaneS)                 \
@@ -260,6 +261,7 @@ namespace compiler {
   V(Mips64I16x8GtU)                          \
   V(Mips64I16x8GeU)                          \
   V(Mips64I16x8RoundingAverageU)             \
+  V(Mips64I16x8Abs)                          \
   V(Mips64I8x16Splat)                        \
   V(Mips64I8x16ExtractLaneU)                 \
   V(Mips64I8x16ExtractLaneS)                 \
@@ -286,6 +288,7 @@ namespace compiler {
   V(Mips64I8x16GtU)                          \
   V(Mips64I8x16GeU)                          \
   V(Mips64I8x16RoundingAverageU)             \
+  V(Mips64I8x16Abs)                          \
   V(Mips64S128And)                           \
   V(Mips64S128Or)                            \
   V(Mips64S128Xor)                           \
diff --git a/src/compiler/backend/mips64/instruction-scheduler-mips64.cc b/src/compiler/backend/mips64/instruction-scheduler-mips64.cc
index d6d0bfa475..0261d915fb 100644
--- a/src/compiler/backend/mips64/instruction-scheduler-mips64.cc
+++ b/src/compiler/backend/mips64/instruction-scheduler-mips64.cc
@@ -167,6 +167,7 @@ int InstructionScheduler::GetTargetInstructionFlags(
     case kMips64I16x8UConvertI8x16High:
     case kMips64I16x8UConvertI8x16Low:
     case kMips64I16x8RoundingAverageU:
+    case kMips64I16x8Abs:
     case kMips64I32x4Add:
     case kMips64I32x4AddHoriz:
     case kMips64I32x4Eq:
@@ -194,6 +195,7 @@ int InstructionScheduler::GetTargetInstructionFlags(
     case kMips64I32x4UConvertF32x4:
     case kMips64I32x4UConvertI16x8High:
     case kMips64I32x4UConvertI16x8Low:
+    case kMips64I32x4Abs:
     case kMips64I8x16Add:
     case kMips64I8x16AddSaturateS:
     case kMips64I8x16AddSaturateU:
@@ -220,6 +222,7 @@ int InstructionScheduler::GetTargetInstructionFlags(
     case kMips64I8x16SubSaturateS:
     case kMips64I8x16SubSaturateU:
     case kMips64I8x16RoundingAverageU:
+    case kMips64I8x16Abs:
     case kMips64Ins:
     case kMips64Lsa:
     case kMips64MaxD:
diff --git a/src/compiler/backend/mips64/instruction-selector-mips64.cc b/src/compiler/backend/mips64/instruction-selector-mips64.cc
index cab89aa112..5a0e41ccbe 100644
--- a/src/compiler/backend/mips64/instruction-selector-mips64.cc
+++ b/src/compiler/backend/mips64/instruction-selector-mips64.cc
@@ -2757,12 +2757,15 @@ void InstructionSelector::VisitInt64AbsWithOverflow(Node* node) {
   V(I32x4SConvertI16x8High, kMips64I32x4SConvertI16x8High) \
   V(I32x4UConvertI16x8Low, kMips64I32x4UConvertI16x8Low)   \
   V(I32x4UConvertI16x8High, kMips64I32x4UConvertI16x8High) \
+  V(I32x4Abs, kMips64I32x4Abs)                             \
   V(I16x8Neg, kMips64I16x8Neg)                             \
   V(I16x8SConvertI8x16Low, kMips64I16x8SConvertI8x16Low)   \
   V(I16x8SConvertI8x16High, kMips64I16x8SConvertI8x16High) \
   V(I16x8UConvertI8x16Low, kMips64I16x8UConvertI8x16Low)   \
   V(I16x8UConvertI8x16High, kMips64I16x8UConvertI8x16High) \
+  V(I16x8Abs, kMips64I16x8Abs)                             \
   V(I8x16Neg, kMips64I8x16Neg)                             \
+  V(I8x16Abs, kMips64I8x16Abs)                             \
   V(S128Not, kMips64S128Not)                               \
   V(S1x4AnyTrue, kMips64S1x4AnyTrue)                       \
   V(S1x4AllTrue, kMips64S1x4AllTrue)                       \