fix(opencl): enable image on mali(cl2.1)

GitOrigin-RevId: 0c670fba807e9bf25e7825e7de5ce8c04d30dae8

dnn/include/megdnn/handle.h

@@ -38,6 +38,17 @@ class Handle {
             CAMBRICON = 12,
         };
 
+        //! Device vendor
+        enum class HandleVendorType : uint32_t {
+            NOT_SPEC = 0,
+            MALI = 1,
+            ADRENO = 2,
+            CUDA = 3,
+            INTEL = 4,
+            POWERVR = 5,
+            AMD = 6,
+        };
+
     protected:
         Handle(megcoreComputingHandle_t computing_handle, HandleType type);
 
@@ -130,6 +141,9 @@ class Handle {
         //! get alignment in bytes for rows of image 2D tensor format
         virtual size_t image2d_pitch_alignment() const;
 
+        //! get vendor type
+        virtual HandleVendorType vendor_type() const;
+
         HandleType type() const {
             return m_handle_type;
         }

dnn/include/megdnn/tensor_format.h

@@ -12,6 +12,7 @@
 #pragma once
 
 #include "megdnn/basic_types.h"
+#include "megdnn/handle.h"
 
 #include "megdnn/internal/visibility_prologue.h"
 namespace megdnn {
@@ -43,7 +44,7 @@ public:
 
 protected:
     ImplBase(Type type) : m_type{type} {}
-    ~ImplBase() = default;
+    virtual ~ImplBase() = default;
 
     static TensorFormat impl_to_tensor_format(ImplBase* impl) { return {impl}; }
 
@@ -93,8 +94,8 @@ namespace detail {
  *
  * \p align_axis is the axis to be aligned, also the first axis of image width.
  * More precisely speaking, `stride[align_axis-1] * dtype.size()` must divide \p
- * align_size_in_byte. Axes from 0 to align_axis-1 would be considered as the
- * height of the image, and other axes are the width.
+ * align_size_in_elements. Axes from 0 to align_axis-1 would be considered as
+ * the height of the image, and other axes are the width.
  *
  * Empty tensors and negative strides are not allowed. Only contiguous or
  * broadcasted cases are allowed.
@@ -103,41 +104,32 @@ namespace detail {
  * considered as contiguous.
  */
 class Image2DTensorFormatBase : public TensorFormat::ImplBase {
-    size_t m_align_axis, m_align_size_in_byte_log2;
+    size_t m_align_axis, m_align_size_in_elements_log2;
 
 protected:
     Image2DTensorFormatBase(Type type, size_t align_axis,
-                            size_t align_size_in_byte);
-    ~Image2DTensorFormatBase() = default;
+                            size_t align_size_in_elements);
+    virtual ~Image2DTensorFormatBase() = default;
 
 public:
     /*!
-     * \brief get alignment requirement in bytes
+     * \brief get alignment requirement in elements
      * \param div_log2 the result would be divided by `(1 << div_log2)`
      */
-    size_t align_size_in_byte(size_t div_log2 = 0) const {
-        return 1 << (m_align_size_in_byte_log2 > div_log2
-                             ? m_align_size_in_byte_log2 - div_log2
+    size_t align_size_in_elements(size_t div_log2 = 0) const {
+        return 1 << (m_align_size_in_elements_log2 > div_log2
+                             ? m_align_size_in_elements_log2 - div_log2
                              : 0);
     }
 
     size_t align_axis() const { return m_align_axis; }
 
-    size_t init_contiguous_stride(TensorLayout& layout) const override;
-
-    bool is_contiguous_spec(const TensorLayout& layout) const override;
-
-    TensorLayout collapse_contiguous_spec(
-            const TensorLayout& layout) const override;
-
-    //! span for image must include the padding at the last row
-    TensorLayout::Span span_spec(const TensorLayout& layout) const override;
+    size_t align_size_in_elements_log2() const {
+        return m_align_size_in_elements_log2;
+    }
 
     std::string to_string() const override;
 
-    //! raise exception if preconditions violated
-    virtual void assert_valid(const TensorLayout& layout) const;
-
     //! modify the align axis and return a new TensorFormat
     virtual TensorFormat change_axis(size_t axis) const = 0;
 
@@ -147,9 +139,6 @@ public:
     //! number of rows
     size_t image_height(const TensorLayout& layout) const;
 
-    //! delta of addresses of consecutive rows (in bytes)
-    size_t image_row_pitch(const TensorLayout& layout) const;
-
     void serialize_append(std::string& result) const override;
 protected:
     struct SerializePack {
@@ -159,9 +148,27 @@ protected:
 
 template <size_t PIXEL_SIZE>
 class Image2DPackedTensorFormatBase : public Image2DTensorFormatBase {
+    Handle::HandleVendorType m_vendor_type = Handle::HandleVendorType::NOT_SPEC;
+    /*!
+     * \brief get fix alignment requirement in bytes, consider m_vendor_type,
+     * for example on MALI, CL_DEVICE_IMAGE_PITCH_ALIGNMENT means image_width
+     * align COUNT, but mdl needs align size in byte, which equal to
+     * (image_width algin count) * sizeof(data_type) * pixel_size
+     */
+    size_t image_pitch_alignment_in_bytes(size_t align_size_in_elements,
+                                          const TensorLayout& layout) const;
+
 protected:
-    using Image2DTensorFormatBase::Image2DTensorFormatBase;
-    ~Image2DPackedTensorFormatBase() = default;
+    Image2DPackedTensorFormatBase(Type type, size_t align_axis,
+                                  size_t align_size_in_elements,
+                                  Handle::HandleVendorType vendor_type)
+            : detail::Image2DTensorFormatBase(type, align_axis,
+                                              align_size_in_elements),
+              m_vendor_type(vendor_type) {}
+
+    virtual ~Image2DPackedTensorFormatBase() = default;
+
+    Handle::HandleVendorType vendor() const { return m_vendor_type; }
 
 public:
     /*!
@@ -173,7 +180,20 @@ public:
      */
     size_t image_width(const TensorLayout& layout) const;
 
-    void assert_valid(const TensorLayout& layout) const override;
+    //! raise exception if preconditions violated
+    void assert_valid(const TensorLayout& layout) const;
+
+    size_t image_row_pitch(const TensorLayout& layout) const;
+
+    //! span for image must include the padding at the last row
+    TensorLayout::Span span_spec(const TensorLayout& layout) const override;
+
+    size_t init_contiguous_stride(TensorLayout& layout) const override;
+
+    bool is_contiguous_spec(const TensorLayout& layout) const override;
+
+    TensorLayout collapse_contiguous_spec(
+            const TensorLayout& layout) const override;
 };
 using Image2DPack4TensorFormatBase = Image2DPackedTensorFormatBase<4>;
 }  // namespace detail
@@ -190,7 +210,10 @@ public:
     static constexpr Type TYPE = Type::IMAGE2D_PACK4;
 
     //! for internal usage or test purposes
-    static TensorFormat make_raw(size_t align_axis, size_t align_size_in_byte);
+    static TensorFormat make_raw(size_t align_axis,
+                                 size_t align_size_in_elements,
+                                 Handle::HandleVendorType vendor_type =
+                                         Handle::HandleVendorType::NOT_SPEC);
 
     static TensorFormat make(size_t align_axis, const Handle* handle);
 
@@ -215,9 +238,10 @@ public:
     TensorFormat change_axis(size_t axis) const override;
 
 private:
-    Image2DPack4TensorFormat(size_t align_axis, size_t align_size_in_byte)
-            : detail::Image2DPack4TensorFormatBase(TYPE, align_axis,
-                                                   align_size_in_byte) {}
+    Image2DPack4TensorFormat(size_t align_axis, size_t align_size_in_elements,
+                             Handle::HandleVendorType vendor_type)
+            : detail::Image2DPack4TensorFormatBase(
+                      TYPE, align_axis, align_size_in_elements, vendor_type) {}
 };
 
 }  // namespace megdnn

dnn/src/common/handle.cpp

@@ -147,6 +147,10 @@ std::unique_ptr<Handle> Handle::make(megcoreComputingHandle_t computing_handle,
         megdnn_throw("image2d tensor format not supported on this handle");
     }
 
+    megdnn::HandleImplHelper::HandleVendorType Handle::vendor_type() const {
+        return HandleVendorType::NOT_SPEC;
+    }
+
     bool Handle::check_cross_dev_copy_constraint(const TensorLayout& src) {
         return src.is_contiguous();
     }

dnn/src/common/relayout_format.cpp

@@ -236,6 +236,7 @@ void RelayoutFormat::deduce_layout(const TensorLayout& src, TensorLayout& dst) {
 
 void RelayoutFormat::deduce_format(TensorFormat src, TensorFormat& dst) {
     size_t align = handle()->image2d_pitch_alignment();
+    auto vendor_type = handle()->vendor_type();
     using Param = param::RelayoutFormat;
 #define CHECK_SRC(_expect)                                                \
     megdnn_assert(src == _expect, "invalid src format: expect=%s got=%s", \
@@ -251,7 +252,7 @@ void RelayoutFormat::deduce_format(TensorFormat src, TensorFormat& dst) {
             break;
         case Param::Mode::NHWC_NHWCD4I:
             CHECK_SRC(DefaultTensorFormat::make());
-            dst = Image2DPack4TensorFormat::make_raw(2, align);
+            dst = Image2DPack4TensorFormat::make_raw(2, align, vendor_type);
             break;
         case Param::Mode::NCHW_NHWCD4:
             CHECK_SRC(DefaultTensorFormat::make());
@@ -263,10 +264,10 @@ void RelayoutFormat::deduce_format(TensorFormat src, TensorFormat& dst) {
             break;
         case Param::Mode::NCHW_NHWCD4I:
             CHECK_SRC(DefaultTensorFormat::make());
-            dst = Image2DPack4TensorFormat::make_raw(2, align);
+            dst = Image2DPack4TensorFormat::make_raw(2, align, vendor_type);
             break;
         case Param::Mode::NHWCD4I_NCHW:
-            CHECK_SRC(Image2DPack4TensorFormat::make_raw(2, align));
+            CHECK_SRC(Image2DPack4TensorFormat::make_raw(2, align, vendor_type));
             dst = DefaultTensorFormat::make();
             break;
         case Param::Mode::NHWCD4_NCHW:
@@ -280,7 +281,7 @@ void RelayoutFormat::deduce_format(TensorFormat src, TensorFormat& dst) {
         case Param::Mode::INTER_WEIGHT_DENSEI:
         case Param::Mode::INTER_WEIGHT_DENSEI_DOT:
             CHECK_SRC(DefaultTensorFormat::make());
-            dst = Image2DPack4TensorFormat::make_raw(3, align);
+            dst = Image2DPack4TensorFormat::make_raw(3, align, vendor_type);
             break;
         case Param::Mode::INTER_WEIGHT_GROUP:
             CHECK_SRC(DefaultTensorFormat::make());
@@ -289,7 +290,7 @@ void RelayoutFormat::deduce_format(TensorFormat src, TensorFormat& dst) {
         case Param::Mode::INTER_WEIGHT_GROUPI:
         case Param::Mode::INTER_WEIGHT_GROUPI_DOT:
             CHECK_SRC(DefaultTensorFormat::make());
-            dst = Image2DPack4TensorFormat::make_raw(4, align);
+            dst = Image2DPack4TensorFormat::make_raw(4, align, vendor_type);
             break;
         case Param::Mode::INTER_WEIGHT_CHAN:
             CHECK_SRC(DefaultTensorFormat::make());
@@ -297,7 +298,7 @@ void RelayoutFormat::deduce_format(TensorFormat src, TensorFormat& dst) {
             break;
         case Param::Mode::INTER_WEIGHT_CHANI:
             CHECK_SRC(DefaultTensorFormat::make());
-            dst = Image2DPack4TensorFormat::make_raw(1, align);
+            dst = Image2DPack4TensorFormat::make_raw(1, align, vendor_type);
             break;
         case Param::Mode::NCHW4_CHWN4:
             CHECK_SRC(DefaultTensorFormat::make());

dnn/src/common/tensor_format.cpp

@@ -185,23 +185,134 @@ TensorFormat DefaultTensorFormat::make() {
 /* ===================== Image2DTensorFormatBase ===================== */
 
 Image2DTensorFormatBase::Image2DTensorFormatBase(Type type, size_t align_axis,
-                                                 size_t align_size_in_byte)
-        : ImplBase(type) {
-    megdnn_assert(align_size_in_byte && align_axis);
-    m_align_axis = align_axis;
-    m_align_size_in_byte_log2 = __builtin_ctz(align_size_in_byte);
-    megdnn_assert((1u << m_align_size_in_byte_log2) == align_size_in_byte,
-                  "align size not power of 2: %zu", align_size_in_byte);
+                                                 size_t align_size_in_elements)
+        : ImplBase(type), m_align_axis(align_axis) {
+    megdnn_assert(align_size_in_elements && align_axis);
+    m_align_size_in_elements_log2 = __builtin_ctz(align_size_in_elements);
+    megdnn_assert(
+            (1u << m_align_size_in_elements_log2) == align_size_in_elements,
+            "align size not power of 2: %zu", align_size_in_elements);
 }
 
-size_t Image2DTensorFormatBase::init_contiguous_stride(
+void Image2DTensorFormatBase::serialize_append(std::string& result) const {
+    SerializePack pack;
+    pack.align_axis = m_align_axis;
+    megdnn_assert(pack.align_axis == m_align_axis);  // detect overflow
+    result.append(reinterpret_cast<char*>(&pack), sizeof(pack));
+}
+
+size_t Image2DTensorFormatBase::image_height(const TensorLayout& layout) const {
+    size_t accum = 1;
+    for (int i = m_align_axis - 1; i >= 0; --i) {
+        if (layout.stride[i] == 0) {
+            // this dimension is broadcasted
+        } else {
+            accum *= layout.shape[i];
+        }
+    }
+    return accum;
+}
+
+size_t Image2DTensorFormatBase::image_width_elems(
+        const TensorLayout& layout) const {
+    size_t high_elem = 0;
+    for (size_t i = m_align_axis; i < layout.ndim; ++i) {
+        high_elem += (layout.shape[i] - 1) * layout.stride[i];
+    }
+    return high_elem + 1;
+}
+
+std::string Image2DTensorFormatBase::to_string() const {
+    return ssprintf("I2D{%zu,%d}", m_align_axis,
+                    1 << m_align_size_in_elements_log2);
+}
+
+/* ===================== Image2DPackedTensorFormatBase ===================== */
+
+template <size_t PIXEL_SIZE>
+size_t Image2DPackedTensorFormatBase<PIXEL_SIZE>::image_width(
+        const TensorLayout& layout) const {
+    auto ret = image_width_elems(layout);
+    megdnn_assert(ret % PIXEL_SIZE == 0);
+    return ret / PIXEL_SIZE;
+}
+
+template <size_t PIXEL_SIZE>
+void Image2DPackedTensorFormatBase<PIXEL_SIZE>::assert_valid(
+        const TensorLayout& layout) const {
+    auto m_align_axis = align_axis();
+    megdnn_assert(!(layout.shape[layout.ndim - 1] % PIXEL_SIZE),
+                  "bad shape: %zu", layout.shape[layout.ndim - 1]);
+    megdnn_assert(layout.dtype.valid() && layout.ndim > m_align_axis);
+    ptrdiff_t first_non_zero_stride = 0;
+    for (int i = layout.ndim - 1; i >= 0; --i) {
+        megdnn_assert(layout.shape[i] && layout.stride[i] >= 0);
+        if (i < static_cast<int>(m_align_axis) && !first_non_zero_stride) {
+            first_non_zero_stride = layout.stride[i];
+        }
+    }
+    size_t mask =
+            image_pitch_alignment_in_bytes(
+                    align_size_in_elements(layout.dtype.size_log()), layout) -
+            1;
+
+    megdnn_assert(!(first_non_zero_stride & mask),
+                  "first stride is %d, but alignment is %zu",
+                  static_cast<int>(first_non_zero_stride), mask + 1);
+}
+
+template <size_t PIXEL_SIZE>
+size_t Image2DPackedTensorFormatBase<PIXEL_SIZE>::image_row_pitch(
+        const TensorLayout& layout) const {
+    for (int i = align_axis() - 1; i >= 0; --i) {
+        // find a non-broadcast axis
+        if (auto s = layout.stride[i]) {
+            return layout.dtype.size(s);
+        }
+    }
+    // use width for all broadcasted case
+    size_t alignment_in_bytes_log2 = align_size_in_elements_log2();
+    if (m_vendor_type == Handle::HandleVendorType::MALI) {
+        alignment_in_bytes_log2 +=
+                __builtin_ctz(layout.dtype.size() * PIXEL_SIZE);
+    }
+
+    return get_aligned_power2<size_t>(
+            layout.dtype.size(image_width_elems(layout)),
+            1 << alignment_in_bytes_log2);
+}
+
+template <size_t PIXEL_SIZE>
+size_t
+Image2DPackedTensorFormatBase<PIXEL_SIZE>::image_pitch_alignment_in_bytes(
+        size_t align_size_in_elements, const TensorLayout& layout) const {
+    return m_vendor_type == Handle::HandleVendorType::MALI
+                   ? (align_size_in_elements * layout.dtype.size() * PIXEL_SIZE)
+                   : align_size_in_elements;
+}
+
+template <size_t PIXEL_SIZE>
+TensorLayout::Span Image2DPackedTensorFormatBase<PIXEL_SIZE>::span_spec(
+        const TensorLayout& layout) const {
+    assert_valid(layout);
+    size_t size = image_height(layout) * image_row_pitch(layout);
+    auto mask = (1 << layout.dtype.size_log()) - 1;
+    megdnn_assert(!(size & mask), "unaligned size: %zu", size);
+    return {0, 0, size >> layout.dtype.size_log(), size};
+}
+
+template <size_t PIXEL_SIZE>
+size_t Image2DPackedTensorFormatBase<PIXEL_SIZE>::init_contiguous_stride(
         TensorLayout& layout) const {
+    auto m_align_axis = align_axis();
     if (!layout.ndim)
         return 0;
     megdnn_assert(layout.dtype.valid() && layout.ndim > m_align_axis,
                   "dtype=%s ndim=%zu align=%zu", layout.dtype.name(),
                   layout.ndim, m_align_axis);
-    size_t align_size = align_size_in_byte(layout.dtype.size_log());
+    size_t align_size = image_pitch_alignment_in_bytes(
+            align_size_in_elements(layout.dtype.size_log()), layout);
+
     size_t accum = 1;
     SafeMultiplies<size_t> mul;
     for (size_t i = layout.ndim; i; --i) {
@@ -216,12 +327,15 @@ size_t Image2DTensorFormatBase::init_contiguous_stride(
     return accum;
 };
 
-bool Image2DTensorFormatBase::is_contiguous_spec(
+template <size_t PIXEL_SIZE>
+bool Image2DPackedTensorFormatBase<PIXEL_SIZE>::is_contiguous_spec(
         const TensorLayout& layout) const {
     megdnn_assert(layout.dtype.valid());
-    size_t align_size = align_size_in_byte(layout.dtype.size_log());
+    size_t align_size = image_pitch_alignment_in_bytes(
+            align_size_in_elements(layout.dtype.size_log()), layout);
+
     ptrdiff_t expected = 1;
-    int height_axis = static_cast<int>(m_align_axis - 1);
+    int height_axis = static_cast<int>(align_axis() - 1);
     for (int i = layout.ndim - 1; i >= 0; --i) {
         if (i == height_axis) {
             expected = megdnn::get_aligned_power2<size_t>(expected, align_size);
@@ -235,7 +349,12 @@ bool Image2DTensorFormatBase::is_contiguous_spec(
                     return false;
                 }
 
-                size_t mask = align_size_in_byte(layout.dtype.size_log()) - 1;
+                size_t mask =
+                        image_pitch_alignment_in_bytes(
+                                align_size_in_elements(layout.dtype.size_log()),
+                                layout) -
+                        1;
+
                 megdnn_assert(s > expected && !(s & mask),
                               "invalid row pitch: %d; layout: %s",
                               static_cast<int>(s), layout.to_string().c_str());
@@ -250,11 +369,12 @@ bool Image2DTensorFormatBase::is_contiguous_spec(
     return expected != 0;
 }
 
-TensorLayout Image2DTensorFormatBase::collapse_contiguous_spec(
+template <size_t PIXEL_SIZE>
+TensorLayout Image2DPackedTensorFormatBase<PIXEL_SIZE>::collapse_contiguous_spec(
         const TensorLayout& layout) const {
     assert_valid(layout);
     TensorLayout res{layout};
-    int new_axis = m_align_axis;
+    int new_axis = align_axis();
     // remove all dims with shape 1
     for (int i = static_cast<int>(res.ndim) - 1; i >= 0 && res.ndim >= 3; --i) {
         if (i == new_axis && static_cast<int>(res.ndim) == new_axis + 1) {
@@ -302,95 +422,6 @@ TensorLayout Image2DTensorFormatBase::collapse_contiguous_spec(
     return res;
 }
 
-TensorLayout::Span Image2DTensorFormatBase::span_spec(
-        const TensorLayout& layout) const {
-    assert_valid(layout);
-    size_t size = image_height(layout) * image_row_pitch(layout);
-    auto mask = (1 << layout.dtype.size_log()) - 1;
-    megdnn_assert(!(size & mask), "unaligned size: %zu", size);
-    return {0, 0, size >> layout.dtype.size_log(), size};
-}
-
-void Image2DTensorFormatBase::serialize_append(std::string& result) const {
-    SerializePack pack;
-    pack.align_axis = m_align_axis;
-    megdnn_assert(pack.align_axis == m_align_axis);  // detect overflow
-    result.append(reinterpret_cast<char*>(&pack), sizeof(pack));
-}
-
-size_t Image2DTensorFormatBase::image_height(const TensorLayout& layout) const {
-    size_t accum = 1;
-    for (int i = m_align_axis - 1; i >= 0; --i) {
-        if (layout.stride[i] == 0) {
-            // this dimension is broadcasted
-        } else {
-            accum *= layout.shape[i];
-        }
-    }
-    return accum;
-}
-
-size_t Image2DTensorFormatBase::image_row_pitch(
-        const TensorLayout& layout) const {
-    for (int i = m_align_axis - 1; i >= 0; --i) {
-        // find a non-broadcast axis
-        if (auto s = layout.stride[i]) {
-            return layout.dtype.size(s);
-        }
-    }
-    // use width for all broadcasted case
-    return get_aligned_power2<size_t>(
-            layout.dtype.size(image_width_elems(layout)),
-            1 << m_align_size_in_byte_log2);
-}
-
-void Image2DTensorFormatBase::assert_valid(const TensorLayout& layout) const {
-    megdnn_assert(layout.dtype.valid() && layout.ndim > m_align_axis);
-    ptrdiff_t first_non_zero_stride = 0;
-    for (int i = layout.ndim - 1; i >= 0; --i) {
-        megdnn_assert(layout.shape[i] && layout.stride[i] >= 0);
-        if (i < static_cast<int>(m_align_axis) && !first_non_zero_stride) {
-            first_non_zero_stride = layout.stride[i];
-        }
-    }
-    size_t mask = align_size_in_byte(layout.dtype.size_log()) - 1;
-    megdnn_assert(!(first_non_zero_stride & mask),
-                  "first stride is %d, but alignment is %zu",
-                  static_cast<int>(first_non_zero_stride), mask + 1);
-}
-
-size_t Image2DTensorFormatBase::image_width_elems(
-        const TensorLayout& layout) const {
-    size_t high_elem = 0;
-    for (size_t i = m_align_axis; i < layout.ndim; ++i) {
-        high_elem += (layout.shape[i] - 1) * layout.stride[i];
-    }
-    return high_elem + 1;
-}
-
-std::string Image2DTensorFormatBase::to_string() const {
-    return ssprintf("I2D{%zu,%d}", m_align_axis,
-                    1 << m_align_size_in_byte_log2);
-}
-
-/* ===================== Image2DPackedTensorFormatBase ===================== */
-
-template <size_t PIXEL_SIZE>
-size_t Image2DPackedTensorFormatBase<PIXEL_SIZE>::image_width(
-        const TensorLayout& layout) const {
-    auto ret = image_width_elems(layout);
-    megdnn_assert(ret % PIXEL_SIZE == 0);
-    return ret / PIXEL_SIZE;
-}
-
-template <size_t PIXEL_SIZE>
-void Image2DPackedTensorFormatBase<PIXEL_SIZE>::assert_valid(
-        const TensorLayout& layout) const {
-    Image2DTensorFormatBase::assert_valid(layout);
-    megdnn_assert(!(layout.shape[layout.ndim - 1] % PIXEL_SIZE),
-                  "bad shape: %zu", layout.shape[layout.ndim - 1]);
-}
-
 namespace megdnn {
 namespace detail {
 template class Image2DPackedTensorFormatBase<4>;
@@ -398,26 +429,29 @@ template class Image2DPackedTensorFormatBase<4>;
 }  // namespace megdnn
 
 /* ===================== Image2DPack4TensorFormat  ===================== */
-TensorFormat Image2DPack4TensorFormat::make_raw(size_t align_axis,
-                                                size_t align_size_in_byte) {
+TensorFormat Image2DPack4TensorFormat::make_raw(
+        size_t align_axis, size_t align_size_in_elements,
+        Handle::HandleVendorType vendor_type) {
     static std::mutex mtx;
     static std::unordered_map<uint64_t,
                               std::unique_ptr<Image2DPack4TensorFormat>>
             cache;
-    megdnn_assert(std::max(align_axis, align_size_in_byte) <=
+    megdnn_assert(std::max(align_axis, align_size_in_elements) <=
                   std::numeric_limits<uint32_t>::max());
     MEGDNN_LOCK_GUARD(mtx);
     auto&& ptr = cache[(static_cast<uint64_t>(align_axis) << 32) |
-                       align_size_in_byte];
+                       align_size_in_elements];
     if (!ptr) {
-        ptr.reset(new Image2DPack4TensorFormat{align_axis, align_size_in_byte});
+        ptr.reset(new Image2DPack4TensorFormat{
+                align_axis, align_size_in_elements, vendor_type});
     }
     return impl_to_tensor_format(ptr.get());
 }
 
 TensorFormat Image2DPack4TensorFormat::make(size_t align_axis,
                                             const Handle* handle) {
-    return make_raw(align_axis, handle->image2d_pitch_alignment());
+    return make_raw(align_axis, handle->image2d_pitch_alignment(),
+                    handle->vendor_type());
 }
 
 TensorFormat Image2DPack4TensorFormat::deserialize(const Handle* handle,
@@ -429,7 +463,7 @@ TensorFormat Image2DPack4TensorFormat::deserialize(const Handle* handle,
 }
 
 TensorFormat Image2DPack4TensorFormat::change_axis(size_t axis) const {
-    return make_raw(axis, align_size_in_byte());
+    return make_raw(axis, align_size_in_elements(), vendor());
 }
 
 // vim: syntax=cpp.doxygen

dnn/src/cuda/handle.cpp

@@ -123,6 +123,10 @@ size_t HandleImpl::image2d_pitch_alignment() const {
     return align;
 }
 
+HandleImpl::HandleVendorType HandleImpl::vendor_type() const {
+    return HandleVendorType::CUDA;
+}
+
 } // namespace cuda
 } // namespace megdnn
 

dnn/src/cuda/handle.h

@@ -123,6 +123,7 @@ class HandleImpl: public HandleImplHelper {
         TypeCvt* typecvt_opr() { return get_helper_opr<TypeCvt, 0>(this); }
 
         size_t image2d_pitch_alignment() const override;
+        HandleVendorType vendor_type() const override;
     private:
         bool m_is_tegra_k1;
         int m_device_id;

dnn/src/naive/handle.cpp

@@ -118,6 +118,10 @@ size_t HandleImpl::image2d_pitch_alignment() const {
     return g_image2d_pitch_alignment;
 }
 
+HandleImpl::HandleVendorType HandleImpl::vendor_type() const {
+    return HandleVendorType::NOT_SPEC;
+}
+
 size_t HandleImpl::exchange_image2d_pitch_alignment(size_t alignment) {
     auto ret = g_image2d_pitch_alignment;
     g_image2d_pitch_alignment = alignment;

dnn/src/naive/handle.h

@@ -169,6 +169,7 @@ public:
      * \param alignment the new alignment value to set
      */
     static size_t exchange_image2d_pitch_alignment(size_t alignment);
+    HandleVendorType vendor_type() const override;
 };
 
 }  // namespace naive

dnn/test/common/test_basic_types.cpp

@@ -175,6 +175,30 @@ namespace {
     }
 }
 
+TEST(BASIC_TYPES, TENSOR_LAYOUT_FMT_WITH_VENDOR_MALI) {
+    TensorFormat fmt = Image2DPack4TensorFormat::make_raw(
+            1, 512, Handle::HandleVendorType::MALI);
+    TensorLayout layout{{5, 3, 8}, dtype::Float32{}, fmt};
+
+    ASSERT_EQ(layout.stride[2], 1);
+    ASSERT_EQ(layout.stride[1], 8);
+    ASSERT_EQ(layout.stride[0], 2048);
+    ASSERT_EQ(8192u, image_row_pitch(layout));
+    ASSERT_EQ(6u, image_width(layout));
+    ASSERT_EQ(5u, image_height(layout));
+
+    fmt = Image2DPack4TensorFormat::make_raw(1, 512,
+                                             Handle::HandleVendorType::MALI);
+    TensorLayout layout_s{{5, 3, 8}, dtype::Float16{}, fmt};
+
+    ASSERT_EQ(layout_s.stride[2], 1);
+    ASSERT_EQ(layout_s.stride[1], 8);
+    ASSERT_EQ(layout_s.stride[0], 2048);
+    ASSERT_EQ(4096u, image_row_pitch(layout_s));
+    ASSERT_EQ(6u, image_width(layout_s));
+    ASSERT_EQ(5u, image_height(layout_s));
+}
+
 TEST(BASIC_TYPES, TENSOR_LAYOUT_FMT) {
     TensorFormat fmt = Image2DPack4TensorFormat::make_raw(1, 1024);
     ASSERT_FALSE(fmt.is_default());
@@ -233,7 +257,7 @@ TEST(BASIC_TYPES, TENSOR_LAYOUT_FMT) {
         auto&& impl = contig.format.as_impl<Image2DPack4TensorFormat>();
         ASSERT_EQ(make_layout({1, 8}, {32, 1}, layout.dtype), contig);
         ASSERT_EQ(1u, impl.align_axis());
-        ASSERT_EQ(64u, impl.align_size_in_byte());
+        ASSERT_EQ(64u, impl.align_size_in_elements());
     }
 }
 
@@ -258,7 +282,7 @@ TEST(BASIC_TYPES, TENSOR_LAYOUT_FMT_COLLAPSE_H) {
         auto&& impl = contig.format.as_impl<Image2DPack4TensorFormat>();
         ASSERT_EQ(make_layout({v0, 8}, {32, 1}, layout.dtype), contig);
         ASSERT_EQ(1u, impl.align_axis());
-        ASSERT_EQ(64u, impl.align_size_in_byte());
+        ASSERT_EQ(64u, impl.align_size_in_elements());
     }
 }
 
@@ -274,7 +298,7 @@ TEST(BASIC_TYPES, TENSOR_LAYOUT_FMT_COLLAPSE_W) {
                               layout.dtype),
                   contig);
         ASSERT_EQ(1u, impl.align_axis());
-        ASSERT_EQ(64u, impl.align_size_in_byte());
+        ASSERT_EQ(64u, impl.align_size_in_elements());
     }
 }