diff --git a/include/gpu/GrRecordingContext.h b/include/gpu/GrRecordingContext.h
index b11d85d..b20bf34 100644
--- a/include/gpu/GrRecordingContext.h
+++ b/include/gpu/GrRecordingContext.h
@@ -96,10 +96,7 @@
     // GrRecordingContext. Arenas does not maintain ownership of the pools it groups together.
     class Arenas {
     public:
-        Arenas(GrMemoryPool*, SkArenaAlloc*, GrSubRunAllocator*);
-
-        // For storing GrOp-derived classes recorded by a GrRecordingContext
-        GrMemoryPool* opMemoryPool() { return fOpMemoryPool; }
+        Arenas(SkArenaAlloc*, GrSubRunAllocator*);
 
         // For storing pipelines and other complex data as-needed by ops
         SkArenaAlloc* recordTimeAllocator() { return fRecordTimeAllocator; }
@@ -108,7 +105,6 @@
         GrSubRunAllocator* recordTimeSubRunAllocator() { return fRecordTimeSubRunAllocator; }
 
     private:
-        GrMemoryPool* fOpMemoryPool;
         SkArenaAlloc* fRecordTimeAllocator;
         GrSubRunAllocator* fRecordTimeSubRunAllocator;
     };
@@ -129,7 +125,6 @@
         OwnedArenas& operator=(OwnedArenas&&);
 
     private:
-        std::unique_ptr<GrMemoryPool> fOpMemoryPool;
         std::unique_ptr<SkArenaAlloc> fRecordTimeAllocator;
         std::unique_ptr<GrSubRunAllocator> fRecordTimeSubRunAllocator;
     };
diff --git a/src/gpu/GrDirectContextPriv.h b/src/gpu/GrDirectContextPriv.h
index 6299549..a6a8c02 100644
--- a/src/gpu/GrDirectContextPriv.h
+++ b/src/gpu/GrDirectContextPriv.h
@@ -53,7 +53,6 @@
     // from GrRecordingContext
     GrDrawingManager* drawingManager() { return fContext->drawingManager(); }
 
-    GrMemoryPool* opMemoryPool() { return fContext->arenas().opMemoryPool(); }
     SkArenaAlloc* recordTimeAllocator() { return fContext->arenas().recordTimeAllocator(); }
     GrRecordingContext::Arenas arenas() { return fContext->arenas(); }
 
diff --git a/src/gpu/GrDrawingManager.cpp b/src/gpu/GrDrawingManager.cpp
index 9077e2c..8ff9ef1 100644
--- a/src/gpu/GrDrawingManager.cpp
+++ b/src/gpu/GrDrawingManager.cpp
@@ -215,15 +215,6 @@
     }
     this->removeRenderTasks();
 
-#ifdef SK_DEBUG
-    // In non-DDL mode this checks that all the flushed ops have been freed from the memory pool.
-    // When we move to partial flushes this assert will no longer be valid.
-    // In DDL mode this check is somewhat superfluous since the memory for most of the ops/opsTasks
-    // will be stored in the DDL's GrOpMemoryPools.
-    GrMemoryPool* opMemoryPool = fContext->priv().opMemoryPool();
-    opMemoryPool->isEmpty();
-#endif
-
     gpu->executeFlushInfo(proxies, access, info, newState);
 
     // Give the cache a chance to purge resources that become purgeable due to flushing.
diff --git a/src/gpu/GrRecordingContext.cpp b/src/gpu/GrRecordingContext.cpp
index d5a5ee2..8b9f622 100644
--- a/src/gpu/GrRecordingContext.cpp
+++ b/src/gpu/GrRecordingContext.cpp
@@ -96,14 +96,11 @@
     fDrawingManager.reset();
 }
 
-GrRecordingContext::Arenas::Arenas(GrMemoryPool* opMemoryPool,
-                                   SkArenaAlloc* recordTimeAllocator,
+GrRecordingContext::Arenas::Arenas(SkArenaAlloc* recordTimeAllocator,
                                    GrSubRunAllocator* subRunAllocator)
-        : fOpMemoryPool(opMemoryPool)
-        , fRecordTimeAllocator(recordTimeAllocator)
+        : fRecordTimeAllocator(recordTimeAllocator)
         , fRecordTimeSubRunAllocator(subRunAllocator) {
     // OwnedArenas should instantiate these before passing the bare pointer off to this struct.
-    SkASSERT(opMemoryPool);
     SkASSERT(recordTimeAllocator);
     SkASSERT(subRunAllocator);
 }
@@ -114,20 +111,12 @@
 GrRecordingContext::OwnedArenas::~OwnedArenas() {}
 
 GrRecordingContext::OwnedArenas& GrRecordingContext::OwnedArenas::operator=(OwnedArenas&& a) {
-    fOpMemoryPool = std::move(a.fOpMemoryPool);
     fRecordTimeAllocator = std::move(a.fRecordTimeAllocator);
     fRecordTimeSubRunAllocator = std::move(a.fRecordTimeSubRunAllocator);
     return *this;
 }
 
 GrRecordingContext::Arenas GrRecordingContext::OwnedArenas::get() {
-    if (!fOpMemoryPool) {
-        // DDL TODO: should the size of the memory pool be decreased in DDL mode? CPU-side memory
-        // consumed in DDL mode vs. normal mode for a single skp might be a good metric of wasted
-        // memory.
-        fOpMemoryPool = GrMemoryPool::Make(16384, 16384);
-    }
-
     if (!fRecordTimeAllocator) {
         // TODO: empirically determine a better number for SkArenaAlloc's firstHeapAllocation param
         fRecordTimeAllocator = std::make_unique<SkArenaAlloc>(sizeof(GrPipeline) * 100);
@@ -137,7 +126,7 @@
         fRecordTimeSubRunAllocator = std::make_unique<GrSubRunAllocator>();
     }
 
-    return {fOpMemoryPool.get(), fRecordTimeAllocator.get(), fRecordTimeSubRunAllocator.get()};
+    return {fRecordTimeAllocator.get(), fRecordTimeSubRunAllocator.get()};
 }
 
 GrRecordingContext::OwnedArenas&& GrRecordingContext::detachArenas() {
diff --git a/src/gpu/GrRecordingContextPriv.h b/src/gpu/GrRecordingContextPriv.h
index 3c1e5f2..08e1c6e 100644
--- a/src/gpu/GrRecordingContextPriv.h
+++ b/src/gpu/GrRecordingContextPriv.h
@@ -41,7 +41,6 @@
     // from GrRecordingContext
     GrDrawingManager* drawingManager() { return fContext->drawingManager(); }
 
-    GrMemoryPool* opMemoryPool() { return fContext->arenas().opMemoryPool(); }
     SkArenaAlloc* recordTimeAllocator() { return fContext->arenas().recordTimeAllocator(); }
     GrSubRunAllocator* recordTimeSubRunAllocator() {
         return fContext->arenas().recordTimeSubRunAllocator();
diff --git a/src/gpu/ops/GrAtlasTextOp.cpp b/src/gpu/ops/GrAtlasTextOp.cpp
index 3b037ba..b2b9b25 100644
--- a/src/gpu/ops/GrAtlasTextOp.cpp
+++ b/src/gpu/ops/GrAtlasTextOp.cpp
@@ -35,7 +35,7 @@
 #include <utility>
 
 // If we have thread local, then cache memory for a single GrAtlasTextOp.
-#if !defined(GR_OP_ALLOCATE_USE_POOL) && defined(GR_HAS_THREAD_LOCAL)
+#if defined(GR_HAS_THREAD_LOCAL)
 static thread_local void* gCache = nullptr;
 void* GrAtlasTextOp::operator new(size_t s) {
     if (gCache != nullptr) {
diff --git a/src/gpu/ops/GrAtlasTextOp.h b/src/gpu/ops/GrAtlasTextOp.h
index 3a2406d..c586888 100644
--- a/src/gpu/ops/GrAtlasTextOp.h
+++ b/src/gpu/ops/GrAtlasTextOp.h
@@ -30,12 +30,10 @@
         }
     }
 
-#if !defined(GR_OP_ALLOCATE_USE_POOL) && defined(GR_HAS_THREAD_LOCAL)
+#if defined(GR_HAS_THREAD_LOCAL)
     void* operator new(size_t s);
     void operator delete(void* b) noexcept;
     static void ClearCache();
-#else
-    static void ClearCache() {}
 #endif
 
     static const int kVerticesPerGlyph = GrAtlasSubRun::kVerticesPerGlyph;
diff --git a/src/gpu/ops/GrOp.cpp b/src/gpu/ops/GrOp.cpp
index 9b1fda3..994cddf 100644
--- a/src/gpu/ops/GrOp.cpp
+++ b/src/gpu/ops/GrOp.cpp
@@ -10,29 +10,6 @@
 std::atomic<uint32_t> GrOp::gCurrOpClassID {GrOp::kIllegalOpID + 1};
 std::atomic<uint32_t> GrOp::gCurrOpUniqueID{GrOp::kIllegalOpID + 1};
 
-#if defined(GR_OP_ALLOCATE_USE_POOL)
-    void GrOp::DeleteFromPool::operator() (GrOp* op) {
-        if (op != nullptr) {
-            op->~GrOp();
-            fPool->release(op);
-        }
-    }
-#endif
-
-#if defined(GR_OP_ALLOCATE_USE_POOL) && defined(SK_DEBUG)
-    void* GrOp::operator new(size_t size) {
-        // All GrOp-derived class should be allocated in a GrMemoryPool
-        SkASSERT(0);
-        return ::operator new(size);
-    }
-
-    void GrOp::operator delete(void* target) {
-        // All GrOp-derived class should be released from their owning GrMemoryPool
-        SkASSERT(0);
-        ::operator delete(target);
-    }
-#endif
-
 GrOp::GrOp(uint32_t classID) : fClassID(classID) {
     SkASSERT(classID == SkToU32(fClassID));
     SkASSERT(classID);
diff --git a/src/gpu/ops/GrOp.h b/src/gpu/ops/GrOp.h
index 3ae4e3d..34acb43 100644
--- a/src/gpu/ops/GrOp.h
+++ b/src/gpu/ops/GrOp.h
@@ -68,28 +68,11 @@
 
 class GrOp : private SkNoncopyable {
 public:
-    #if defined(GR_OP_ALLOCATE_USE_POOL)
-        struct DeleteFromPool {
-            DeleteFromPool() : fPool{nullptr} {}
-            DeleteFromPool(GrMemoryPool* pool) : fPool{pool} {}
-            void operator() (GrOp* op);
-            GrMemoryPool* fPool;
-        };
-        using Owner =  std::unique_ptr<GrOp, DeleteFromPool>;
-    #else
         using Owner = std::unique_ptr<GrOp>;
-    #endif
 
     template<typename Op, typename... Args>
     static Owner Make(GrRecordingContext* context, Args&&... args) {
-        #if defined(GR_OP_ALLOCATE_USE_POOL)
-            GrMemoryPool* pool = context->priv().opMemoryPool();
-            void* mem = pool->allocate(sizeof(Op));
-            GrOp* op = new (mem) Op(std::forward<Args>(args)...);
-            return Owner{op, pool};
-        #else
-            return Owner{new Op(std::forward<Args>(args)...)};
-        #endif
+        return Owner{new Op(std::forward<Args>(args)...)};
     }
 
     template<typename Op, typename... Args>
@@ -97,23 +80,12 @@
             GrRecordingContext* context, const SkPMColor4f& color,
             GrPaint&& paint, Args&&... args);
 
-    #if defined(GR_OP_ALLOCATE_USE_POOL)
-        template<typename Op, typename... Args>
-        static Owner MakeWithExtraMemory(
-                GrRecordingContext* context, size_t extraSize, Args&&... args) {
-            GrMemoryPool* pool = context->priv().opMemoryPool();
-            void* mem = pool->allocate(sizeof(Op) + extraSize);
-            GrOp* op = new (mem) Op(std::forward<Args>(args)...);
-            return Owner{op, pool};
-        }
-    #else
-        template<typename Op, typename... Args>
-        static Owner MakeWithExtraMemory(
-                GrRecordingContext* context, size_t extraSize, Args&&... args) {
-            void* bytes = ::operator new(sizeof(Op) + extraSize);
-            return Owner{new (bytes) Op(std::forward<Args>(args)...)};
-        }
-    #endif
+    template<typename Op, typename... Args>
+    static Owner MakeWithExtraMemory(
+            GrRecordingContext* context, size_t extraSize, Args&&... args) {
+        void* bytes = ::operator new(sizeof(Op) + extraSize);
+        return Owner{new (bytes) Op(std::forward<Args>(args)...)};
+    }
 
     virtual ~GrOp() = default;
 
@@ -169,20 +141,7 @@
         return SkToBool(fBoundsFlags & kZeroArea_BoundsFlag);
     }
 
-    #if defined(GR_OP_ALLOCATE_USE_POOL) && defined(SK_DEBUG)
-        // All GrOp-derived classes should be allocated in and deleted from a GrMemoryPool
-        void* operator new(size_t size);
-        void operator delete(void* target);
-
-        void* operator new(size_t size, void* placement) {
-            return ::operator new(size, placement);
-        }
-        void operator delete(void* target, void* placement) {
-            ::operator delete(target, placement);
-        }
-    #else
-        void operator delete(void* p) { ::operator delete(p); }
-    #endif
+    void operator delete(void* p) { ::operator delete(p); }
 
     /**
      * Helper for safely down-casting to a GrOp subclass
diff --git a/src/gpu/ops/GrSimpleMeshDrawOpHelper.h b/src/gpu/ops/GrSimpleMeshDrawOpHelper.h
index a4910eb..8ea2390 100644
--- a/src/gpu/ops/GrSimpleMeshDrawOpHelper.h
+++ b/src/gpu/ops/GrSimpleMeshDrawOpHelper.h
@@ -196,18 +196,10 @@
 GrOp::Owner GrOp::MakeWithProcessorSet(
         GrRecordingContext* context, const SkPMColor4f& color,
         GrPaint&& paint, Args&&... args) {
-#if defined(GR_OP_ALLOCATE_USE_POOL)
-    GrMemoryPool* pool = context->priv().opMemoryPool();
-    char* bytes = (char*)pool->allocate(sizeof(Op) + sizeof(GrProcessorSet));
-    char* setMem = bytes + sizeof(Op);
-    GrProcessorSet* processorSet = new (setMem)  GrProcessorSet{std::move(paint)};
-    return Owner{new (bytes) Op(processorSet, color, std::forward<Args>(args)...), pool};
-#else
     char* bytes = (char*)::operator new(sizeof(Op) + sizeof(GrProcessorSet));
     char* setMem = bytes + sizeof(Op);
     GrProcessorSet* processorSet = new (setMem)  GrProcessorSet{std::move(paint)};
     return Owner{new (bytes) Op(processorSet, color, std::forward<Args>(args)...)};
-#endif
 }
 
 template <typename Op, typename... OpArgs>