manual merge

src/targets/gpu/include/migraphx/gpu/hip.hpp

@@ -29,6 +29,7 @@
 #include <migraphx/literal.hpp>
 #include <migraphx/check_shapes.hpp>
 #include <migraphx/functional.hpp>
+#include <migraphx/dyn_output.hpp>
 #include <utility>
 
 namespace migraphx {
@@ -112,7 +113,7 @@ struct hip_copy_to_gpu
     std::string name() const { return "hip::copy_to_gpu"; }
     shape compute_shape(std::vector<shape> inputs) const
     {
-        check_shapes{inputs, *this}.has(1, 2).same_type();
+        check_shapes{inputs, *this, true}.has(1, 2).same_type();
         return inputs.at(0);
     }
     argument compute(context& ctx, const shape&, const std::vector<argument>& args) const
@@ -121,6 +122,10 @@ struct hip_copy_to_gpu
         if(args.size() == 1)
             return input;
         argument result = args[1].share();
+        if(result.get_shape().dynamic())
+        {
+            result = result.reshape(args[0].get_shape());
+        }
         gpu_copy(ctx, input, result);
         // Associate the input since it was registered with hip
         return {result.get_shape(), [input, result]() mutable { return result.data(); }};
@@ -138,19 +143,24 @@ struct hip_copy_from_gpu
     std::string name() const { return "hip::copy_from_gpu"; }
     shape compute_shape(std::vector<shape> inputs) const
     {
-        check_shapes{inputs, *this}.has(1, 2).same_type();
+        check_shapes{inputs, *this, true}.has(1, 2).same_type();
         return inputs.at(0);
     }
     argument
-    compute(context& ctx, const shape& output_shape, const std::vector<argument>& args) const
+    compute(context& ctx, const dyn_output& dyn_out, const std::vector<argument>& args) const
     {
         if(args.size() == 1)
         {
-            argument result = allocate_gpu(output_shape, true);
+            argument result = allocate_gpu(dyn_out.computed_shape, true);
             gpu_copy(ctx, args[0], result);
             return result;
         }
-        copy_from_gpu(ctx, args[0], args[1]);
+        argument input = args[0].share();
+        if(input.get_shape().dynamic())
+        {
+            input = input.reshape(args[1].get_shape());
+        }
+        copy_from_gpu(ctx, input, args[1]);
         return args[1];
     }
     std::ptrdiff_t output_alias(const std::vector<shape>& args) const

src/targets/gpu/include/migraphx/gpu/oper.hpp

@@ -31,7 +31,6 @@
 #include <migraphx/argument.hpp>
 #include <migraphx/config.hpp>
 #include <migraphx/reduce_dims.hpp>
-#include <migraphx/type_name.hpp>
 #include <utility>
 #include <iostream>
 

src/targets/gpu/include/migraphx/gpu/prefix_scan_sum.hpp

@@ -33,7 +33,6 @@
 #include <migraphx/shape.hpp>
 #include <migraphx/argument.hpp>
 #include <migraphx/config.hpp>
-#include <migraphx/type_name.hpp>
 #include <utility>
 #include <iostream>
 

src/targets/gpu/include/migraphx/gpu/reduce_op.hpp

@@ -31,7 +31,6 @@
 #include <migraphx/shape.hpp>
 #include <migraphx/argument.hpp>
 #include <migraphx/config.hpp>
-#include <migraphx/type_name.hpp>
 #include <utility>
 #include <iostream>
 

src/targets/gpu/driver/include/migraphx/gpu/driver/perf.hpp → src/targets/gpu/include/migraphx/gpu/time_op.hpp

@@ -31,12 +31,10 @@
 namespace migraphx {
 inline namespace MIGRAPHX_INLINE_NS {
 namespace gpu {
-namespace driver {
 
 std::pair<double, double>
 time_op(context& ictx, operation op, const std::vector<shape>& inputs, int n = 100);
 
-} // namespace driver
 } // namespace gpu
 } // namespace MIGRAPHX_INLINE_NS
 } // namespace migraphx

src/targets/gpu/jit/ck_gemm.cpp

+/*
+ * The MIT License (MIT)
+ *
+ * Copyright (c) 2015-2022 Advanced Micro Devices, Inc. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+#include <fstream>
+#include <migraphx/filesystem.hpp>
+#include <migraphx/gpu/compiler.hpp>
+#include <migraphx/make_op.hpp>
+#include <migraphx/gpu/context.hpp>
+
+#include <migraphx/env.hpp>
+#include <migraphx/file_buffer.hpp>
+#include <migraphx/gpu/compile_gen.hpp>
+#include <migraphx/gpu/compile_hip.hpp>
+#include <migraphx/gpu/compile_hip_code_object.hpp>
+#include <migraphx/module.hpp>
+#include <migraphx/ranges.hpp>
+#include <migraphx/reduce_dims.hpp>
+#include <migraphx/stringutils.hpp>
+
+#include "ck/host/device_gemm_multiple_d.hpp"
+
+namespace migraphx {
+inline namespace MIGRAPHX_INLINE_NS {
+
+namespace gpu {
+
+using namespace migraphx::gpu::gen; // NOLINT
+
+MIGRAPHX_DECLARE_ENV_VAR(MIGRAPHX_LOG_CK_GEMM);
+MIGRAPHX_DECLARE_ENV_VAR(MIGRAPHX_CK_TUNING);
+MIGRAPHX_DECLARE_ENV_VAR(MIGRAPHX_CK_TUNING_VALUE);
+MIGRAPHX_DECLARE_ENV_VAR(MIGRAPHX_CK_DEBUG);
+
+// NOLINTNEXTLINE
+static const char* const ck_gemm_kernel = R"__migraphx__(
+#include <args.hpp>
+#include <migraphx/kernels/ck_gemm.hpp>
+#include <migraphx/kernels/pointwise.hpp>
+#include <migraphx/kernels/ops.hpp>
+#include <${include}>
+
+namespace migraphx {
+
+${preamble}
+
+extern "C" {
+
+__global__ void ${kernel}(${params})
+{
+    transform_args(make_tensors(), rotate_last())(${args})([](auto... xs) {
+        ck_gemm<${solution}, ${blocks_per_batch}>(xs...);
+    });
+}
+
+}
+
+} // namespace migraphx
+
+)__migraphx__";
+
+// NOLINTNEXTLINE
+static const char* const disable_warning_pragma = R"__migraphx__(
+#pragma clang diagnostic push
+#pragma clang diagnostic ignored "-Weverything"
+${content}
+#pragma clang diagnostic pop
+)__migraphx__";
+
+template <class P>
+static std::string ck_disable_warnings(P p)
+{
+    return interpolate_string(disable_warning_pragma,
+                              {{"content", std::string{p.first, p.second}}});
+}
+
+static std::unordered_map<std::string, std::string> create_ck_header_strings()
+{
+    std::unordered_map<std::string, std::string> result;
+    auto ck_headers = ck::host::GetHeaders();
+
+    std::transform(
+        ck_headers.begin(), ck_headers.end(), std::inserter(result, result.begin()), [&](auto&& p) {
+            return std::make_pair(p.first, ck_disable_warnings(p.second));
+        });
+    return result;
+}
+
+static std::vector<src_file> create_ck_headers()
+{
+    static const auto& header_strings = create_ck_header_strings();
+    std::vector<src_file> srcs;
+    std::transform(
+        header_strings.begin(), header_strings.end(), std::back_inserter(srcs), [&](auto&& p) {
+            return src_file{fs::path{p.first},
+                            {p.second.data(), p.second.data() + p.second.size()}};
+        });
+    return srcs;
+}
+
+static const std::vector<src_file>& ck_headers()
+{
+    static const auto& headers = create_ck_headers();
+    return headers;
+}
+
+static bool transposed_matrix(const shape& s) { return s.strides().back() != 1; }
+
+using tuning_entry = std::pair<std::vector<shape>, size_t>;
+static std::vector<tuning_entry> read_tuning(const std::string& s)
+{
+    if(not fs::exists(s))
+        return {};
+    return from_value<std::vector<tuning_entry>>(from_json_string(read_string(s)));
+}
+
+static float matrix_distance(const shape& x, const shape& y)
+{
+    if(x.type() != y.type())
+        return std::numeric_limits<float>::max();
+    if(transposed_matrix(x) != transposed_matrix(y))
+        return std::numeric_limits<float>::max();
+    auto sum_squared = std::inner_product(x.lens().rbegin(),
+                                          x.lens().rbegin() + 2,
+                                          y.lens().rbegin(),
+                                          0,
+                                          std::plus<>{},
+                                          [](auto a, auto b) { return (a - b) * (a - b); });
+    return std::sqrt(sum_squared);
+}
+
+static std::size_t get_tuning_for(const std::vector<shape>& inputs)
+{
+    static auto tuning = read_tuning(string_value_of(MIGRAPHX_CK_TUNING{}, ""));
+    if(tuning.empty())
+    {
+        std::cout << "*********** Warning: No CK tuning! for config:" << std::endl;
+        std::cout << "  " << inputs[0] << std::endl;
+        std::cout << "  " << inputs[1] << std::endl;
+        std::cout << "  " << inputs[2] << std::endl;
+    }
+    auto it = std::find_if(
+        tuning.begin(), tuning.end(), [&](const auto& p) { return p.first == inputs; });
+    if(it == tuning.end())
+    {
+        std::cout << "*********** Warning: CK tuning missing for config!" << std::endl;
+        std::cout << "  " << inputs[0] << std::endl;
+        std::cout << "  " << inputs[1] << std::endl;
+        std::cout << "  " << inputs[2] << std::endl;
+        std::vector<std::pair<float, std::size_t>> w;
+        std::transform(tuning.begin(), tuning.end(), std::back_inserter(w), [&](const auto& p) {
+            if(inputs.size() < 3 or p.first.size() < 3)
+                MIGRAPHX_THROW("Invalid CK config");
+            auto avg_distance = std::inner_product(
+                p.first.begin(),
+                p.first.begin() + 3,
+                inputs.begin(),
+                0.0f,
+                std::plus<>{},
+                [](const auto& x, const auto& y) { return matrix_distance(x, y) / 3.0f; });
+            return std::make_pair(avg_distance, p.second);
+        });
+        std::sort(w.begin(), w.end());
+        std::size_t default_value = 4;
+        if(not w.empty())
+            default_value = w.front().second;
+        auto tuning_val = value_of(MIGRAPHX_CK_TUNING_VALUE{}, default_value);
+        std::cout << "*********** Warning: CK try tuning: " << tuning_val << std::endl;
+        return tuning_val;
+    }
+    return it->second;
+}
+
+struct ck_gemm_compiler : compiler<ck_gemm_compiler>
+{
+    static std::string get_layout(const shape& s)
+    {
+        return transposed_matrix(s) ? "ck::tensor_layout::gemm::ColumnMajor"
+                                    : "ck::tensor_layout::gemm::RowMajor";
+    }
+
+    static ck::host::DataType get_type(const shape& s)
+    {
+        if(s.type() == shape::half_type)
+            return ck::host::DataType::Half;
+        else if(s.type() == shape::float_type)
+            return ck::host::DataType::Float;
+        else if(s.type() == shape::int8_type)
+            return ck::host::DataType::Int8;
+        else if(s.type() == shape::int32_type)
+            return ck::host::DataType::Int32;
+        MIGRAPHX_THROW("Unsupported ck type");
+    }
+
+    template <class Iterator, class F>
+    static std::string ck_tuple(Iterator start, Iterator last, F f)
+    {
+        std::vector<std::string> s;
+        std::transform(start, last, std::back_inserter(s), f);
+        return "ck::Tuple<" + join_strings(s, ",") + ">";
+    }
+
+    static std::vector<shape> adjust_inputs(std::vector<shape> inputs, bool& swap_inputs)
+    {
+        swap_inputs  = false;
+        auto c_shape = inputs.back();
+        if(not transposed_matrix(c_shape))
+            return inputs;
+        std::vector<int64_t> perm(c_shape.lens().size());
+        std::iota(perm.begin(), perm.end(), 0);
+        std::swap(perm[perm.size() - 1], perm[perm.size() - 2]);
+        std::transform(inputs.begin(), inputs.end(), inputs.begin(), [&](shape s) {
+            return reorder_shape(s, perm);
+        });
+        swap_inputs = true;
+        return inputs;
+    }
+
+    static std::size_t get_batch_count(const shape& s)
+    {
+        return std::accumulate(
+            s.lens().rbegin() + 2, s.lens().rend(), std::size_t{1}, std::multiplies<std::size_t>());
+    }
+
+    static void fold_batch_dims(shape& s)
+    {
+        auto lens = s.lens();
+        if(lens.size() <= 2)
+            return;
+        auto batch_count = get_batch_count(s);
+        auto m1          = lens.at(lens.size() - 2);
+        auto m2          = lens.at(lens.size() - 1);
+        if(transposed_matrix(s))
+            s = shape{s.type(), {m1, m2 * batch_count}};
+        else
+            s = shape{s.type(), {m1 * batch_count, m2}};
+    }
+
+    static void remove_batch_dims(shape& s)
+    {
+        auto lens = s.lens();
+        if(lens.size() <= 2)
+            return;
+        auto m1 = lens.at(lens.size() - 2);
+        auto m2 = lens.at(lens.size() - 1);
+        s       = shape{s.type(), {m1, m2}};
+    }
+
+    std::vector<std::string> names() const { return {"gpu::ck_gemm"}; }
+
+    static bool standard_batch(const shape& s)
+    {
+        if(s.lens().size() < 3)
+            return true;
+        std::vector<std::size_t> lens(s.lens().begin(), s.lens().end() - 2);
+        std::vector<std::size_t> strides(s.strides().begin(), s.strides().end() - 2);
+        auto base = *(s.lens().end() - 2) * *(s.lens().end() - 1);
+        std::transform(strides.begin(), strides.end(), strides.begin(), [&](auto stride) {
+            return stride / base;
+        });
+        return shape{s.type(), lens, strides}.standard();
+    }
+
+    bool can_fold_batch(const std::vector<shape>& inputs) const
+    {
+        const auto& b_shape = inputs[1];
+        if(std::any_of(inputs.begin() + 2, inputs.end() - 1, [](auto input) {
+               return not standard_batch(input);
+           }))
+            return false;
+        const auto& b_strides = b_shape.strides();
+        return std::all_of(
+            b_strides.begin(), b_strides.end() - 2, [](auto stride) { return stride == 0; });
+    }
+
+    ck::host::device_gemm_multiple_d::Problem create_problem(const std::vector<shape>& inputs,
+                                                             const value& v) const
+    {
+        const auto& a_shape = inputs[0];
+        const auto& b_shape = inputs[1];
+        const auto& c_shape = inputs.back();
+
+        auto rank = a_shape.lens().size();
+
+        auto batch_count = get_batch_count(c_shape);
+        auto m           = c_shape.lens()[rank - 2];
+        m                = can_fold_batch(inputs) ? m * batch_count : m;
+        auto n           = c_shape.lens().back();
+        auto k           = a_shape.lens().back();
+
+        const bool trans_a = transposed_matrix(a_shape);
+        const bool trans_b = transposed_matrix(b_shape);
+        const bool trans_e = transposed_matrix(c_shape);
+        const auto a_type  = get_type(a_shape);
+        const auto b_type  = get_type(b_shape);
+        const auto e_type  = get_type(c_shape);
+        std::vector<bool> ds_layout;
+        std::transform(inputs.begin() + 2,
+                       inputs.end() - 1,
+                       std::back_inserter(ds_layout),
+                       [](const auto& i) { return transposed_matrix(i); });
+        std::vector<ck::host::DataType> ds_type;
+        std::transform(inputs.begin() + 2,
+                       inputs.end() - 1,
+                       std::back_inserter(ds_type),
+                       [](const auto& i) { return get_type(i); });
+
+        std::string ck_passthrough = "ck_passthrough";
+        std::string cde_op         = ck_passthrough;
+        assert(inputs.size() < 4 or v.contains("post"));
+        if(v.contains("post"))
+        {
+            cde_op = v.at("post").to<std::string>();
+        }
+
+        return ck::host::device_gemm_multiple_d::Problem{m,
+                                                         n,
+                                                         k,
+                                                         trans_a,
+                                                         trans_b,
+                                                         trans_e,
+                                                         ds_layout,
+                                                         a_type,
+                                                         b_type,
+                                                         e_type,
+                                                         ds_type,
+                                                         ck_passthrough,
+                                                         ck_passthrough,
+                                                         cde_op};
+    }
+
+    operation compile_op(context& ctx, const std::vector<shape>& inputs, const value& v) const
+    {
+        const auto& a_shape = inputs[0];
+        const auto& b_shape = inputs[1];
+        const auto& c_shape = inputs.back();
+        auto tuning_value   = v.get("tuning_value", 4);
+        if(not v.contains("tuning_value"))
+            tuning_value = get_tuning_for({a_shape, b_shape, c_shape});
+        auto batch_count = get_batch_count(c_shape);
+        auto problem     = create_problem(inputs, v);
+
+        const auto include_header   = problem.GetIncludeHeader();
+        const auto solutions        = problem.GetSolutions(ctx.get_current_device().get_gfx_name());
+        const auto& solution        = solutions.at(tuning_value);
+        const auto template_str     = solution.template_str;
+        const auto blocks_per_batch = solution.grid_size;
+        const auto block_size       = solution.block_size;
+
+        hip_compile_options options;
+        options.additional_src_files = ck_headers();
+        auto grid_size = can_fold_batch(inputs) ? blocks_per_batch : batch_count * blocks_per_batch;
+        options.set_launch_params(v, grid_size * block_size, block_size);
+        options.inputs         = inputs;
+        options.output         = c_shape;
+        options.kernel_name    = v.get("kernel", "ck_gemm_kernel");
+        options.virtual_inputs = inputs;
+        if(can_fold_batch(inputs))
+        {
+            auto vinputs = inputs;
+            fold_batch_dims(vinputs[0]);
+            remove_batch_dims(vinputs[1]);
+            std::for_each(vinputs.begin() + 2, vinputs.end(), fold_batch_dims);
+            options.virtual_inputs = vinputs;
+        }
+
+        if(v.get("check", false) or enabled(MIGRAPHX_CK_DEBUG{}))
+            options.params += " -DMIGRAPHX_CK_CHECK=1";
+
+        auto src = interpolate_string(ck_gemm_kernel,
+                                      {{"solution", template_str},
+                                       {"include", include_header},
+                                       {"params", enum_params(inputs.size(), "void * private_p")},
+                                       {"args", enum_params(inputs.size(), "private_p")},
+                                       {"blocks_per_batch", to_string(blocks_per_batch)},
+                                       {"preamble", v.get("preamble", std::string{})},
+                                       {"kernel", options.kernel_name}});
+
+        return compile_hip_code_object(src, options);
+    }
+
+    value create_settings(instruction_ref ins, const operation& op) const
+    {
+        auto v      = op.to_value();
+        v["kernel"] = "ck_gemm_kernel";
+        if(not ins->module_inputs().empty())
+        {
+            auto* pm      = ins->module_inputs().front();
+            v["preamble"] = generate_pointwise(*pm, "post_ck_gemm_function") +
+                            "\nMIGRAPHX_LIFT_CLASS(post_ck_gemm, post_ck_gemm_function);";
+            v["post"]   = "ck_function_adaptor<post_ck_gemm>";
+            v["kernel"] = "ck_gemm_" + generate_name_from_ops(*pm) + "_kernel";
+        }
+        return v;
+    }
+
+    compiler_replace
+    compile(context& ctx, instruction_ref ins, const operation& op, const value& solution) const
+    {
+        auto shapes = to_shapes(ins->inputs());
+        auto v      = create_settings(ins, op);
+        if(solution.is_null())
+            v["tuning_value"] = 4;
+        else
+            v["tuning_value"] = solution;
+        return {compile_op(ctx, shapes, v),
+                [=](module& m, instruction_ref ins2, const operation& code_object) {
+                    if(enabled(MIGRAPHX_LOG_CK_GEMM{}))
+                    {
+                        std::vector<shape> gemm_shapes{
+                            shapes[0], shapes[1], shapes.back().with_type(shapes[0].type())};
+                        std::cout << "gpu::ck_gemm: " << to_json_string(to_value(gemm_shapes))
+                                  << std::endl;
+                    }
+                    m.replace_instruction(ins2, code_object, ins2->inputs());
+                }};
+    }
+
+    optional<tuning_config>
+    get_tuning_config(context& ctx, instruction_ref ins, const operation& op) const
+    {
+        tuning_config tc;
+        auto shapes    = to_shapes(ins->inputs());
+        auto problem   = create_problem(shapes, create_settings(ins, op));
+        auto solutions = problem.GetSolutions(ctx.get_current_device().get_gfx_name());
+        tc.solutions.resize(solutions.size());
+        std::iota(tc.solutions.begin(), tc.solutions.end(), 0);
+        std::vector<shape> gemm_shapes{shapes[0], shapes[1], shapes.back()};
+        tc.problem = to_value(gemm_shapes);
+        return tc;
+    }
+};
+
+} // namespace gpu
+} // namespace MIGRAPHX_INLINE_NS
+} // namespace migraphx

src/targets/gpu/jit/concat.cpp

@@ -78,7 +78,9 @@ struct concat_compiler : compiler<concat_compiler>
         options.params      = "-Wno-float-equal";
         options.kernel_name = v.get("kernel", "concat_kernel");
         auto axis           = find_fast_axis(options.inputs);
-        auto vec            = vectorize::elements(ctx, axis, options.inputs);
+        vectorize vec{};
+        if(axis != v.at("axis").to<std::size_t>())
+            vec = vectorize::elements(ctx, axis, options.inputs);
         options.set_launch_params(
             v, compute_global_for(ctx, get_concat_elements(options.inputs) / vec.size, 256));
         auto src = interpolate_string(
@@ -106,7 +108,7 @@ struct concat_compiler : compiler<concat_compiler>
             v["post"]          = "MIGRAPHX_LIFT(post_concat)";
             v["kernel"]        = "concat_" + generate_name_from_ops(*pm) + "_kernel";
         }
-        return replace(compile_op(ctx, to_shapes(ins->inputs()), v));
+        return compile_op(ctx, to_shapes(ins->inputs()), v);
     }
 };
 

src/targets/gpu/jit/gather.cpp

@@ -80,7 +80,7 @@ struct gather_compiler : compiler<gather_compiler>
 
     compiler_replace compile(context& ctx, instruction_ref ins, const operation& op) const
     {
-        return replace(compile_op(ctx, to_shapes(ins->inputs()), op.to_value()));
+        return compile_op(ctx, to_shapes(ins->inputs()), op.to_value());
     }
 };
 

src/targets/gpu/jit/gathernd.cpp

@@ -82,7 +82,7 @@ struct gathernd_compiler : compiler<gathernd_compiler>
 
     compiler_replace compile(context& ctx, instruction_ref ins, const operation& op) const
     {
-        return replace(compile_op(ctx, to_shapes(ins->inputs()), op.to_value()));
+        return compile_op(ctx, to_shapes(ins->inputs()), op.to_value());
     }
 };
 

src/targets/gpu/jit/layernorm.cpp

@@ -122,7 +122,7 @@ struct layernorm_compiler : compiler<layernorm_compiler>
             v["kernel"] =
                 v["layernorm"].to<std::string>() + "_" + generate_name_from_ops(*pm) + "_kernel";
         }
-        return replace(compile_op(ctx, to_shapes(ins->inputs()), v));
+        return compile_op(ctx, to_shapes(ins->inputs()), v);
     }
 };
 

src/targets/gpu/jit/mlir.cpp

@@ -32,7 +32,7 @@ namespace gpu {
 
 struct mlir_compiler : compiler<mlir_compiler>
 {
-    std::vector<std::string> names() const { return {"gpu::mlir_conv"}; }
+    std::vector<std::string> names() const { return {"gpu::mlir_op"}; }
 
     operation compile_op(context&, const std::vector<shape>&, const value&) const { return {}; }
 
@@ -45,10 +45,10 @@ struct mlir_compiler : compiler<mlir_compiler>
 
     compiler_replace insert(code_object_op co) const
     {
-        return [co = std::move(co)](module& m, instruction_ref ins) {
-            auto mlir = insert_mlir(m, ins, co, ins->inputs());
-            m.replace_instruction(ins, mlir);
-        };
+        return {std::move(co), [](module& m, instruction_ref ins, const operation& op) {
+                    auto mlir = insert_mlir(m, ins, any_cast<code_object_op>(op), ins->inputs());
+                    m.replace_instruction(ins, mlir);
+                }};
     }
 };
 

src/targets/gpu/jit/pad.cpp

@@ -92,7 +92,7 @@ struct pad_compiler : compiler<pad_compiler>
 
     compiler_replace compile(context& ctx, instruction_ref ins, const operation& op) const
     {
-        return replace(compile_op(ctx, to_shapes(ins->inputs()), op.to_value()));
+        return compile_op(ctx, to_shapes(ins->inputs()), op.to_value());
     }
 };
 } // namespace gpu

src/targets/gpu/jit/pointwise.cpp

@@ -93,10 +93,10 @@ struct pointwise_compiler : compiler<pointwise_compiler>
     {
         if(contains({"layout", "contiguous"}, op.name()))
         {
-            return replace(compile_op(
+            return compile_op(
                 ctx,
                 to_shapes(ins->inputs()),
-                {{"lambda", "[](auto x) { return x; }"}, {"kernel", op.name() + "_kernel"}}));
+                {{"lambda", "[](auto x) { return x; }"}, {"kernel", op.name() + "_kernel"}});
         }
         else
         {
@@ -105,10 +105,9 @@ struct pointwise_compiler : compiler<pointwise_compiler>
             auto pf            = generate_pointwise(*pm, "inner_pointwise");
             std::string lambda = "MIGRAPHX_LIFT(inner_pointwise)";
             auto kernel_name   = generate_name_from_ops(*pm) + "_kernel";
-            return replace(
-                compile_op(ctx,
-                           to_shapes(ins->inputs()),
-                           {{"lambda", lambda}, {"preamble", pf}, {"kernel", kernel_name}}));
+            return compile_op(ctx,
+                              to_shapes(ins->inputs()),
+                              {{"lambda", lambda}, {"preamble", pf}, {"kernel", kernel_name}});
         }
     }
 };

src/targets/gpu/jit/reduce.cpp

@@ -60,15 +60,6 @@ __global__ void reduce_kernel(void* input_p, void* output_p)
 
 )__migraphx__";
 
-static std::size_t get_reduce_elements(const std::vector<shape>& inputs)
-{
-    return inputs.front().elements() / inputs.back().elements();
-}
-static std::size_t get_reduce_elements(const std::vector<instruction_ref>& inputs)
-{
-    return get_reduce_elements(to_shapes(inputs));
-}
-
 static std::vector<std::size_t> get_reduce_lens(const std::vector<std::size_t>& input_lens,
                                                 const std::vector<std::size_t>& output_lens)
 {
@@ -86,9 +77,28 @@ static std::vector<std::size_t> get_reduce_lens(const std::vector<std::size_t>&
     return reduce_lens;
 }
 
-static std::string get_reduce_algo(const std::vector<shape>& inputs)
+template <class T>
+static shape get_reduced_shape(const shape& s, const std::vector<T>& axes)
+{
+    auto lens = s.lens();
+    std::fill(lens.begin(), lens.end(), 1);
+    for(const auto& axis : axes)
+        lens[axis] = s.lens()[axis];
+    return shape{s.type(), lens};
+}
+
+template <class T>
+static shape get_output_shape(const shape& s, const std::vector<T>& axes)
+{
+    auto lens = s.lens();
+    for(const auto& axis : axes)
+        lens[axis] = 1;
+    return shape{s.type(), lens};
+}
+
+template <class ReduceLens>
+static std::string get_reduce_algo(const std::vector<shape>& inputs, ReduceLens rlens)
 {
-    auto rlens      = get_reduce_lens(inputs.front().lens(), inputs.back().lens());
     const auto init = std::numeric_limits<std::size_t>::max();
     // The minimum stride
     auto min_stride = std::inner_product(
@@ -103,11 +113,27 @@ static std::string get_reduce_algo(const std::vector<shape>& inputs)
     return "block";
 }
 
-struct reduce_compiler : compiler<reduce_compiler>
+static std::string get_reduce_algo(const std::vector<shape>& inputs)
+{
+    auto rlens = get_reduce_lens(inputs.front().lens(), inputs.back().lens());
+    return get_reduce_algo(inputs, rlens);
+}
+
+struct simple_reduce_compiler : compiler<simple_reduce_compiler>
 {
     std::vector<std::string> names() const
     {
-        return {"reduce", "reduce_sum", "reduce_mean", "reduce_max", "reduce_min", "reduce_prod"};
+        return {"simple_reduce",
+                "reduce_sum",
+                "reduce_mean",
+                "reduce_max",
+                "reduce_min",
+                "reduce_prod"};
+    }
+
+    static std::size_t get_reduce_elements(const std::vector<shape>& inputs)
+    {
+        return inputs.front().elements() / inputs.back().elements();
     }
 
     operation compile_op(context& ctx, const std::vector<shape>& inputs, const value& v) const
@@ -157,45 +183,109 @@ struct reduce_compiler : compiler<reduce_compiler>
     compiler_replace compile(context& ctx, instruction_ref ins, const operation& op) const
     {
         value v = value::object{};
-        if(op.name() == "reduce_sum")
-        {
-            v["reduction"] = "op::sum{}";
-        }
-        else if(op.name() == "reduce_mean")
-        {
-            auto reduce_elements = get_reduce_elements(ins->inputs());
-            auto reduce_type     = ins->inputs().front()->get_shape().type();
-            v["reduction"]       = "op::sum{}";
-            std::string mean     = "op::mean<" + std::to_string(reduce_elements) + ">{}";
-            // Use float accumulator when reduction size is too large for half
-            if(reduce_type == shape::half_type and reduce_elements > 16384)
-                v["read"] = "compose(" + mean + ", op::convert_to<float>{})";
-            else if(contains({shape::float_type, shape::half_type, shape::double_type},
-                             reduce_type))
-                v["read"] = mean;
-            else
-                v["write"] = mean;
-        }
-        else if(op.name() == "reduce_max")
-        {
-            v["reduction"] = "op::max{}";
-            v["init"]      = "lowest{}";
-        }
-        else if(op.name() == "reduce_min")
+        reduce_op r{};
+        r.set(ins, op);
+        v["reduction"] = r.reduction;
+        v["read"]      = r.read;
+        v["write"]     = r.write;
+        v["init"]      = r.init;
+        return compile_op(ctx, to_shapes(ins->inputs()), v);
+    }
+};
+
+static const char* const fused_reduce_kernel = R"__migraphx__(
+#include <migraphx/kernels/index.hpp>
+#include <migraphx/kernels/reduce.hpp>
+#include <migraphx/kernels/pointwise.hpp>
+#include <migraphx/kernels/vectorize.hpp>
+#include <args.hpp>
+
+namespace migraphx {
+
+${preamble}
+
+extern "C" {
+MIGRAPHX_GLOBAL void ${kernel}(${params})
+{
+    transform_args(make_tensors(), rotate_last(), ${transformers})(${args})([](auto y, auto... xs) {
+        fused_reduce<reduce::${algo}, ${reduced}>(y, partial(${lambda})(xs...));
+    });
+}
+    
+}
+
+} // namespace migraphx
+
+)__migraphx__";
+
+struct fused_reduce_compiler : compiler<fused_reduce_compiler>
+{
+    std::vector<std::string> names() const { return {"fused_reduce"}; }
+
+    operation compile_op(context& ctx, const std::vector<shape>& inputs, const value& v) const
+    {
+        auto axes           = v.at("axes").to_vector<std::size_t>();
+        auto virtual_inputs = inputs;
+        virtual_inputs.push_back(get_reduced_shape(inputs.front(), axes));
+        virtual_inputs.push_back(get_output_shape(inputs.front(), axes));
+        virtual_inputs           = reduce_dims(virtual_inputs);
+        auto reduce_output_shape = virtual_inputs.back();
+        virtual_inputs.pop_back();
+        auto reduction_shape = virtual_inputs.back();
+        virtual_inputs.pop_back();
+
+        hip_compile_options options;
+        options.inputs         = inputs;
+        options.output         = inputs.back();
+        options.virtual_inputs = virtual_inputs;
+        auto faxis             = find_fast_axis({options.virtual_inputs.front()});
+        vectorize vec{};
+        auto nelements = reduce_output_shape.elements();
+        auto algo = v.get("algo", get_reduce_algo(options.virtual_inputs, reduction_shape.lens()));
+        if(algo == "block")
         {
-            v["reduction"] = "op::min{}";
-            v["init"]      = "highest{}";
+            // Vectorize if the axis is a reduction axis
+            if(reduce_output_shape.lens()[faxis] == 1)
+                vec = vectorize::elements(ctx, faxis, options.virtual_inputs);
+            auto relements  = reduction_shape.elements() / vec.size;
+            auto block_size = compute_block_size(relements, 256);
+            if(relements >= block_size * 256)
+                algo = "block_large";
+            options.set_launch_params(
+                v, compute_global_for(ctx, nelements * block_size, 256), block_size);
         }
-        else if(op.name() == "reduce_prod")
+        else if(algo == "lane")
         {
-            v["reduction"] = "op::product{}";
-            v["init"]      = "1";
+            options.set_launch_params(v, compute_global_for(ctx, nelements, 256));
         }
         else
         {
-            MIGRAPHX_THROW("Unsupported reduce");
+            MIGRAPHX_THROW("Unknown reduce algo: " + algo);
         }
-        return replace(compile_op(ctx, to_shapes(ins->inputs()), v));
+        options.kernel_name = v.get("kernel", "reduce_kernel");
+        auto src            = interpolate_string(
+            fused_reduce_kernel,
+            {{"kernel", options.kernel_name},
+             {"params", enum_params(inputs.size(), "void * private_p")},
+             {"args", enum_params(inputs.size(), "private_p")},
+             {"algo", algo},
+             {"reduced", "decltype(" + generate_make_shape(reduce_output_shape) + ")"},
+             {"lambda", v.at("lambda").to<std::string>()},
+             {"transformers", make_transformer_args(vec)},
+             {"preamble", v.get("preamble", std::string{})}});
+        options.params += "-Wno-float-equal";
+        return compile_hip_code_object(src, options);
+    }
+
+    compiler_replace compile(context& ctx, instruction_ref ins, const operation& op) const
+    {
+        assert(not ins->module_inputs().empty());
+        auto v        = op.to_value();
+        auto* rm      = ins->module_inputs().front();
+        v["preamble"] = generate_reduce(*rm, "fused_reduce_op");
+        v["lambda"]   = "MIGRAPHX_LIFT(fused_reduce_op)";
+        v["kernel"]   = generate_name_from_ops(*rm) + "_kernel";
+        return compile_op(ctx, to_shapes(ins->inputs()), v);
     }
 };
 } // namespace gpu

src/targets/gpu/jit/roialign.cpp

@@ -92,7 +92,7 @@ struct roialign_compiler : compiler<roialign_compiler>
 
     compiler_replace compile(context& ctx, instruction_ref ins, const operation& op) const
     {
-        return replace(compile_op(ctx, to_shapes(ins->inputs()), op.to_value()));
+        return compile_op(ctx, to_shapes(ins->inputs()), op.to_value());
     }
 };
 

src/targets/gpu/jit/scatternd.cpp

@@ -85,15 +85,15 @@ struct scatternd_compiler : compiler<scatternd_compiler>
             {{"reduction", reduction}}));
     }
 
-    compiler_replace insert(const operation& op) const
+    compiler_replace insert(const operation& co) const
     {
-        return [=](module& m, instruction_ref ins) {
-            auto args = ins->inputs();
-            args.back() =
-                m.insert_instruction(ins, make_op("hip::copy"), args.front(), args.back());
-            args.erase(args.begin());
-            return m.replace_instruction(ins, op, args);
-        };
+        return {co, [](module& m, instruction_ref ins, const operation& op) {
+                    auto args = ins->inputs();
+                    args.back() =
+                        m.insert_instruction(ins, make_op("hip::copy"), args.front(), args.back());
+                    args.erase(args.begin());
+                    return m.replace_instruction(ins, op, args);
+                }};
     }
 };
 

src/targets/gpu/jit/softmax.cpp

@@ -95,7 +95,7 @@ struct softmax_compiler : compiler<softmax_compiler>
 
     compiler_replace compile(context& ctx, instruction_ref ins, const operation& op) const
     {
-        return replace(compile_op(ctx, to_shapes(ins->inputs()), op.to_value()));
+        return compile_op(ctx, to_shapes(ins->inputs()), op.to_value());
     }
 };
 

src/targets/gpu/kernels/include/migraphx/kernels/array.hpp

@@ -272,6 +272,18 @@ struct integral_const_array : array<T, sizeof...(Xs)>
     MIGRAPHX_DEVICE_CONSTEXPR integral_const_array() : base_array({Xs...}) {}
 };
 
+template <class T, class... Ts>
+constexpr auto make_const_array(T x, Ts... xs)
+{
+    return integral_const_array<typename T::value_type, x, xs...>{};
+}
+
+template <class T, T... Xs, class F>
+constexpr auto unpack(integral_const_array<T, Xs...>, F f)
+{
+    return f(_c<Xs>...);
+}
+
 template <class T, T... Xs, class F>
 constexpr auto transform(integral_const_array<T, Xs...>, F f)
 {

src/targets/gpu/kernels/include/migraphx/kernels/ck.hpp

+/*
+ * The MIT License (MIT)
+ *
+ * Copyright (c) 2015-2022 Advanced Micro Devices, Inc. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+#ifndef MIGRAPHX_GUARD_KERNELS_CK_HPP
+#define MIGRAPHX_GUARD_KERNELS_CK_HPP
+
+#include <migraphx/kernels/debug.hpp>
+#include <migraphx/kernels/types.hpp>
+#include <migraphx/kernels/type_traits.hpp>
+#include <migraphx/kernels/tensor_view.hpp>
+#include <ck/utility/common_header.hpp>
+#include <ck/tensor_description/tensor_descriptor.hpp>
+#include <ck/tensor_description/tensor_descriptor_helper.hpp>
+#include <ck/tensor_operation/gpu/device/tensor_layout.hpp>
+
+namespace migraphx {
+
+namespace detail {
+template <class T>
+struct to_ck_type_impl
+{
+    using type = T;
+};
+template <>
+struct to_ck_type_impl<migraphx::half>
+{
+    using type = ck::half_t;
+};
+
+template <class T>
+struct to_ck_type_impl<const T>
+{
+    using type = const typename to_ck_type_impl<T>::type;
+};
+
+template <class Shape>
+constexpr bool is_row_major()
+{
+    constexpr auto strides = Shape{}.strides;
+    MIGRAPHX_ASSERT(strides.size() >= 2);
+    if(strides.back() == 1)
+    {
+        MIGRAPHX_ASSERT(not Shape{}.is_transposed());
+        return true;
+    }
+    MIGRAPHX_ASSERT(strides[strides.size() - 2] == 1);
+
+    return false;
+}
+
+} // namespace detail
+
+template <class T>
+using to_ck_type = typename detail::to_ck_type_impl<T>::type;
+
+template <class T>
+constexpr auto to_ck_pointer(T* x)
+{
+    return static_cast<to_ck_type<T>*>(x);
+}
+
+template <class T>
+constexpr auto to_ck_const_pointer(const T* x)
+{
+    return static_cast<const to_ck_type<T>*>(x);
+}
+
+template <class Shape>
+using to_ck_gemm_layout = conditional_t<detail::is_row_major<get_shape_c<Shape>>(),
+                                        ck::tensor_layout::gemm::RowMajor,
+                                        ck::tensor_layout::gemm::ColumnMajor>;
+
+template <class Tensor>
+constexpr auto to_ck_tensor()
+{
+    constexpr auto s = get_shape_c<Tensor>{};
+    return sequence(s.lens.size(), [&](auto... is) {
+        return ck::make_naive_tensor_descriptor(ck::make_tuple(s.lens[is]...),
+                                                ck::make_tuple(s.strides[is]...));
+    });
+}
+
+template <class F>
+struct ck_function_adaptor : F
+{
+    template <class... Ts>
+    constexpr ck_function_adaptor(Ts&&... xs) : F(static_cast<Ts&&>(xs)...)
+    {
+    }
+
+    template <class T, class... Ts>
+    constexpr void operator()(T& out, Ts&&... xs) const
+    {
+        out = static_cast<const F&>(*this)(static_cast<Ts&&>(xs)...);
+    }
+};
+
+struct ck_nop
+{
+    template <class T>
+    constexpr void operator()(T&) const
+    {
+    }
+};
+
+struct ck_passthrough
+{
+    template <class T, class U>
+    constexpr void operator()(T& y, U x) const
+    {
+        y = x;
+    }
+};
+
+struct ck_scale
+{
+    constexpr ck_scale(float s) : scale(s) {}
+
+    template <class T, class U>
+    constexpr void operator()(T& y, U x) const
+    {
+        y = x * static_cast<U>(scale);
+    }
+
+    float scale;
+};
+
+struct ck_add
+{
+    template <class T, class U>
+    constexpr void operator()(T& y, U x) const
+    {
+        y += x;
+    }
+};
+
+#ifdef MIGRAPHX_CK_CHECK
+#define MIGRAPHX_CK_STATIC_ASSERT static_assert
+#else
+#define MIGRAPHX_CK_STATIC_ASSERT(...)
+#endif
+
+} // namespace migraphx
+#endif // MIGRAPHX_GUARD_KERNELS_CK_HPP

src/targets/gpu/kernels/include/migraphx/kernels/ck_gemm.hpp

+/*
+ * The MIT License (MIT)
+ *
+ * Copyright (c) 2015-2022 Advanced Micro Devices, Inc. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+#ifndef MIGRAPHX_GUARD_KERNELS_CK_GEMM_HPP
+#define MIGRAPHX_GUARD_KERNELS_CK_GEMM_HPP
+
+#include <migraphx/kernels/index.hpp>
+#include <migraphx/kernels/algorithm.hpp>
+#include <migraphx/kernels/integral_constant.hpp>
+#include <migraphx/kernels/tensor_view.hpp>
+#include <migraphx/kernels/ck.hpp>
+#include <migraphx/kernels/gemm_batcher.hpp>
+
+namespace migraphx {
+
+// In CK, the B matrix is ordered as N,K instead of K,N
+template <class Dims>
+constexpr auto ck_transposeb_dims(Dims dims)
+{
+    return unpack(dims, [](auto k, auto n) { return make_const_array(n, k); });
+}
+
+template <class Tensor>
+using ck_transposeb = decltype(make_shape(ck_transposeb_dims(get_shape_c<Tensor>{}.lens),
+                                          ck_transposeb_dims(get_shape_c<Tensor>{}.strides)));
+
+template <class G, class E, class A, class B, class... Ds>
+__device__ void ck_gemm_matrix(E e, A a, B b, Ds... ds)
+{
+    constexpr auto desc = G::make_descriptor(to_ck_tensor<A>(),
+                                             to_ck_tensor<ck_transposeb<B>>(),
+                                             ck::make_tuple(to_ck_tensor<Ds>()...),
+                                             to_ck_tensor<E>());
+
+    static_assert(desc.is_valid, "Invalid ck gemm.");
+
+    G::Run(desc,
+           to_ck_const_pointer(a.data()),
+           to_ck_const_pointer(b.data()),
+           ck::make_tuple(to_ck_const_pointer(ds.data())...),
+           to_ck_pointer(e.data()));
+}
+
+template <class G, index_int BlocksPerBatch, class... Ts>
+__device__ void ck_gemm(Ts... xs)
+{
+    gemm_batch_args(make_index(), _c<BlocksPerBatch>, xs...)(
+        [](auto... ys) { ck_gemm_matrix<G>(ys...); });
+}
+
+} // namespace migraphx
+#endif