diff --git a/.github/scripts/ci-test-assertions.sh b/.github/scripts/ci-test-assertions.sh
index 0a6a7d2f..64e75996 100755
--- a/.github/scripts/ci-test-assertions.sh
+++ b/.github/scripts/ci-test-assertions.sh
@@ -19,6 +19,12 @@ export MMTK_PLAN=SemiSpace
 build/linux-x86_64-normal-server-$DEBUG_LEVEL/jdk/bin/java -XX:+UseThirdPartyHeap -server -XX:MetaspaceSize=100M -Xms500M -Xmx500M -jar benchmarks/dacapo-2006-10-MR2.jar fop
 build/linux-x86_64-normal-server-$DEBUG_LEVEL/jdk/bin/java -XX:+UseThirdPartyHeap -server -XX:MetaspaceSize=100M -Xms500M -Xmx500M -jar benchmarks/dacapo-2006-10-MR2.jar luindex
 
+# --- Immix ---
+export MMTK_PLAN=Immix
+
+build/linux-x86_64-normal-server-$DEBUG_LEVEL/jdk/bin/java -XX:+UseThirdPartyHeap -server -XX:MetaspaceSize=100M -Xms500M -Xmx500M -jar benchmarks/dacapo-2006-10-MR2.jar fop
+build/linux-x86_64-normal-server-$DEBUG_LEVEL/jdk/bin/java -XX:+UseThirdPartyHeap -server -XX:MetaspaceSize=100M -Xms500M -Xmx500M -jar benchmarks/dacapo-2006-10-MR2.jar luindex
+
 # -- GenCopy --
 export MMTK_PLAN=GenCopy
 
diff --git a/.github/scripts/ci-test-normal.sh b/.github/scripts/ci-test-normal.sh
index 3ea6bead..e6ee752a 100755
--- a/.github/scripts/ci-test-normal.sh
+++ b/.github/scripts/ci-test-normal.sh
@@ -32,6 +32,24 @@ build/linux-x86_64-normal-server-$DEBUG_LEVEL/jdk/bin/java -XX:+UseThirdPartyHea
 build/linux-x86_64-normal-server-$DEBUG_LEVEL/jdk/bin/java -XX:+UseThirdPartyHeap -server -XX:MetaspaceSize=100M -Xms500M -Xmx500M -jar benchmarks/dacapo-2006-10-MR2.jar hsqldb
 build/linux-x86_64-normal-server-$DEBUG_LEVEL/jdk/bin/java -XX:+UseThirdPartyHeap -server -XX:MetaspaceSize=100M -Xms500M -Xmx500M -jar benchmarks/dacapo-2006-10-MR2.jar eclipse
 
+# --- Immix ---
+export MMTK_PLAN=Immix
+
+# Test - the benchmarks that are commented out do not work yet
+# Note: the command line options are necessary for now to ensure the benchmarks work. We may later change the options if we do not have these many constraints.
+build/linux-x86_64-normal-server-$DEBUG_LEVEL/jdk/bin/java -XX:+UseThirdPartyHeap -server -XX:MetaspaceSize=100M -Xms500M -Xmx500M -jar benchmarks/dacapo-2006-10-MR2.jar antlr
+#build/linux-x86_64-normal-server-$DEBUG_LEVEL/jdk/bin/java -XX:+UseThirdPartyHeap -server -XX:MetaspaceSize=100M -Xms500M -Xmx500M -jar benchmarks/dacapo-2006-10-MR2.jar bloat - does not work for stock build
+build/linux-x86_64-normal-server-$DEBUG_LEVEL/jdk/bin/java -XX:+UseThirdPartyHeap -server -XX:MetaspaceSize=100M -Xms500M -Xmx500M -jar benchmarks/dacapo-2006-10-MR2.jar fop
+#build/linux-x86_64-normal-server-$DEBUG_LEVEL/jdk/bin/java -XX:+UseThirdPartyHeap -server -XX:MetaspaceSize=100M -Xms500M -Xmx500M -jar benchmarks/dacapo-2006-10-MR2.jar jython - does not work for stock build
+build/linux-x86_64-normal-server-$DEBUG_LEVEL/jdk/bin/java -XX:+UseThirdPartyHeap -server -XX:MetaspaceSize=100M -Xms500M -Xmx500M -jar benchmarks/dacapo-2006-10-MR2.jar luindex
+#build/linux-x86_64-normal-server-$DEBUG_LEVEL/jdk/bin/java -XX:+UseThirdPartyHeap -server -XX:MetaspaceSize=100M -Xms500M -Xmx500M -jar benchmarks/dacapo-2006-10-MR2.jar lusearch - validation failed
+build/linux-x86_64-normal-server-$DEBUG_LEVEL/jdk/bin/java -XX:+UseThirdPartyHeap -server -XX:MetaspaceSize=100M -Xms500M -Xmx500M -jar benchmarks/dacapo-2006-10-MR2.jar pmd
+#build/linux-x86_64-normal-server-$DEBUG_LEVEL/jdk/bin/java -XX:+UseThirdPartyHeap -server -XX:MetaspaceSize=100M -Xms500M -Xmx500M -jar benchmarks/dacapo-2006-10-MR2.jar xalan - mmtk-core gets stuck in slowdebug build
+
+# These benchmarks take 40s+ for slowdebug build, we may consider removing them from the CI
+build/linux-x86_64-normal-server-$DEBUG_LEVEL/jdk/bin/java -XX:+UseThirdPartyHeap -server -XX:MetaspaceSize=100M -Xms500M -Xmx500M -jar benchmarks/dacapo-2006-10-MR2.jar hsqldb
+build/linux-x86_64-normal-server-$DEBUG_LEVEL/jdk/bin/java -XX:+UseThirdPartyHeap -server -XX:MetaspaceSize=100M -Xms500M -Xmx500M -jar benchmarks/dacapo-2006-10-MR2.jar eclipse
+
 # --- GenCopy ---
 export MMTK_PLAN=GenCopy
 
diff --git a/mmtk/Cargo.toml b/mmtk/Cargo.toml
index bad132ee..3fd768bd 100644
--- a/mmtk/Cargo.toml
+++ b/mmtk/Cargo.toml
@@ -20,7 +20,7 @@ lazy_static = "1.1"
 # - change branch
 # - change repo name
 # But other changes including adding/removing whitespaces in commented lines may break the CI.
-mmtk = { git = "https://github.com/mmtk/mmtk-core.git", rev = "f1655d8644cc260b4c23d3ceb74ca31599447581" }
+mmtk = { git = "https://github.com/mmtk/mmtk-core.git", rev = "093da769a71067dcd4f37db6f453213e7dace660" }
 # Uncomment the following to build locally
 # mmtk = { path = "../repos/mmtk-core" }
 
@@ -42,3 +42,4 @@ semispace = []
 gencopy = []
 marksweep = []
 pageprotect = []
+immix = []
diff --git a/mmtk/src/lib.rs b/mmtk/src/lib.rs
index bc328de1..a335bb6b 100644
--- a/mmtk/src/lib.rs
+++ b/mmtk/src/lib.rs
@@ -112,6 +112,8 @@ lazy_static! {
         std::env::set_var("MMTK_PLAN", "MarkSweep");
         #[cfg(feature = "pageprotect")]
         std::env::set_var("MMTK_PLAN", "PageProtect");
+        #[cfg(feature = "immix")]
+        std::env::set_var("MMTK_PLAN", "Immix");
         MMTK::new()
     };
 }
diff --git a/openjdk/mmtk.h b/openjdk/mmtk.h
index 2481e331..e8ad38d5 100644
--- a/openjdk/mmtk.h
+++ b/openjdk/mmtk.h
@@ -51,6 +51,7 @@ struct AllocatorSelector {
 #define TAG_BUMP_POINTER    0
 #define TAG_LARGE_OBJECT    1
 #define TAG_MALLOC          2
+#define TAG_IMMIX           3
 
 extern AllocatorSelector get_allocator_mapping(int allocator);
 extern size_t get_max_non_los_default_alloc_bytes();
diff --git a/openjdk/mmtkBarrierSetAssembler_x86.cpp b/openjdk/mmtkBarrierSetAssembler_x86.cpp
index f6625073..2089cf83 100644
--- a/openjdk/mmtkBarrierSetAssembler_x86.cpp
+++ b/openjdk/mmtkBarrierSetAssembler_x86.cpp
@@ -70,18 +70,29 @@ void MMTkBarrierSetAssembler::eden_allocate(MacroAssembler* masm, Register threa
     }
 
     // Only bump pointer allocator is implemented.
-    if (selector.tag != TAG_BUMP_POINTER) {
+    if (selector.tag != TAG_BUMP_POINTER && selector.tag != TAG_IMMIX) {
       fatal("unimplemented allocator fastpath\n");
     }
 
-    // Calculate offsets of top and end. We now assume we are using bump pointer.
-    int allocator_base_offset = in_bytes(JavaThread::third_party_heap_mutator_offset())
-      + in_bytes(byte_offset_of(MMTkMutatorContext, allocators))
-      + in_bytes(byte_offset_of(Allocators, bump_pointer))
-      + selector.index * sizeof(BumpAllocator);
+    // Calculat offsets of top and end. We now assume we are using bump pointer.
+    int allocator_base_offset;
+    Address cursor, limit;
 
-    Address cursor = Address(r15_thread, allocator_base_offset + in_bytes(byte_offset_of(BumpAllocator, cursor)));
-    Address limit = Address(r15_thread, allocator_base_offset + in_bytes(byte_offset_of(BumpAllocator, limit)));
+    if (selector.tag == TAG_IMMIX) {
+      allocator_base_offset = in_bytes(JavaThread::third_party_heap_mutator_offset())
+        + in_bytes(byte_offset_of(MMTkMutatorContext, allocators))
+        + in_bytes(byte_offset_of(Allocators, immix))
+        + selector.index * sizeof(ImmixAllocator);
+      cursor = Address(r15_thread, allocator_base_offset + in_bytes(byte_offset_of(ImmixAllocator, cursor)));
+      limit = Address(r15_thread, allocator_base_offset + in_bytes(byte_offset_of(ImmixAllocator, limit)));
+    } else {
+      allocator_base_offset = in_bytes(JavaThread::third_party_heap_mutator_offset())
+        + in_bytes(byte_offset_of(MMTkMutatorContext, allocators))
+        + in_bytes(byte_offset_of(Allocators, bump_pointer))
+        + selector.index * sizeof(BumpAllocator);
+      cursor = Address(r15_thread, allocator_base_offset + in_bytes(byte_offset_of(BumpAllocator, cursor)));
+      limit = Address(r15_thread, allocator_base_offset + in_bytes(byte_offset_of(BumpAllocator, limit)));
+    }
     // obj = load lab.cursor
     __ movptr(obj, cursor);
     // end = obj + size
diff --git a/openjdk/mmtkBarrierSetC2.cpp b/openjdk/mmtkBarrierSetC2.cpp
index aea2c0db..d5418864 100644
--- a/openjdk/mmtkBarrierSetC2.cpp
+++ b/openjdk/mmtkBarrierSetC2.cpp
@@ -177,19 +177,28 @@ void MMTkBarrierSetC2::expand_allocate(
 
     {
       // Only bump pointer allocator fastpath is implemented.
-      if (selector.tag != TAG_BUMP_POINTER) {
+      if (selector.tag != TAG_BUMP_POINTER && selector.tag != TAG_IMMIX) {
         fatal("unimplemented allocator fastpath\n");
       }
 
       // Calculat offsets of top and end. We now assume we are using bump pointer.
-      int allocator_base_offset = in_bytes(JavaThread::third_party_heap_mutator_offset())
-        + in_bytes(byte_offset_of(MMTkMutatorContext, allocators))
-        + in_bytes(byte_offset_of(Allocators, bump_pointer))
-        + selector.index * sizeof(BumpAllocator);
-
+      int allocators_base_offset = in_bytes(JavaThread::third_party_heap_mutator_offset())
+        + in_bytes(byte_offset_of(MMTkMutatorContext, allocators));
+      int tlab_top_offset, tlab_end_offset;
+      if (selector.tag == TAG_IMMIX) {
+        int allocator_base_offset = allocators_base_offset
+          + in_bytes(byte_offset_of(Allocators, immix))
+          + selector.index * sizeof(ImmixAllocator);
+        tlab_top_offset = allocator_base_offset + in_bytes(byte_offset_of(ImmixAllocator, cursor));
+        tlab_end_offset = allocator_base_offset + in_bytes(byte_offset_of(ImmixAllocator, limit));
+      } else {
+        int allocator_base_offset = allocators_base_offset
+          + in_bytes(byte_offset_of(Allocators, bump_pointer))
+          + selector.index * sizeof(BumpAllocator);
+        tlab_top_offset = allocator_base_offset + in_bytes(byte_offset_of(BumpAllocator, cursor));
+        tlab_end_offset = allocator_base_offset + in_bytes(byte_offset_of(BumpAllocator, limit));
+      }
       Node* thread = x->transform_later(new ThreadLocalNode());
-      int tlab_top_offset = allocator_base_offset + in_bytes(byte_offset_of(BumpAllocator, cursor));
-      int tlab_end_offset = allocator_base_offset + in_bytes(byte_offset_of(BumpAllocator, limit));
       eden_top_adr = x->basic_plus_adr(x->top()/*not oop*/, thread, tlab_top_offset);
       eden_end_adr = x->basic_plus_adr(x->top()/*not oop*/, thread, tlab_end_offset);
     }
diff --git a/openjdk/mmtkMutator.hpp b/openjdk/mmtkMutator.hpp
index 12cff9d6..410499e0 100644
--- a/openjdk/mmtkMutator.hpp
+++ b/openjdk/mmtkMutator.hpp
@@ -22,6 +22,7 @@ struct RustDynPtr {
 const int MAX_BUMP_ALLOCATORS = 5;
 const int MAX_LARGE_OBJECT_ALLOCATORS = 1;
 const int MAX_MALLOC_ALLOCATORS = 1;
+const int MAX_IMMIX_ALLOCATORS = 1;
 
 // The following types should have the same layout as the types with the same name in MMTk core (Rust)
 
@@ -39,6 +40,22 @@ struct LargeObjectAllocator {
     RustDynPtr plan;
 };
 
+struct ImmixAllocator {
+    void* tls;
+    void* cursor;
+    void* limit;
+    void* immix_space;
+    RustDynPtr plan;
+    uint8_t hot;
+    uint8_t copy;
+    void* large_cursor;
+    void* large_limit;
+    uint8_t request_for_large;
+    uint8_t _align[7];
+    uint8_t line_opt_tag;
+    uintptr_t line_opt;
+};
+
 struct MallocAllocator {
     void* tls;
     void* space;
@@ -49,6 +66,7 @@ struct Allocators {
     BumpAllocator bump_pointer[MAX_BUMP_ALLOCATORS];
     LargeObjectAllocator large_object[MAX_LARGE_OBJECT_ALLOCATORS];
     MallocAllocator malloc[MAX_MALLOC_ALLOCATORS];
+    ImmixAllocator immix[MAX_IMMIX_ALLOCATORS];
 };
 
 struct MutatorConfig {