onednn-src 0.1.13

/*******************************************************************************
* Copyright 2021 Intel Corporation
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
*     http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*******************************************************************************/

#ifndef GPU_INTEL_CONCAT_XE_HPP
#define GPU_INTEL_CONCAT_XE_HPP

#include "common/primitive.hpp"
#include "gpu/intel/concat/config.hpp"
#include "gpu/intel/primitive.hpp"

namespace dnnl {
namespace impl {
namespace gpu {
namespace intel {
namespace concat {

struct xe_t : public primitive_t {
    using primitive_t::primitive_t;
    struct pd_t : public concat::pd_t {
        pd_t(const primitive_attr_t *attr, const memory_desc_t *dst_md, int n,
                int concat_dim, const memory_desc_t *const *src_mds)
            : concat::pd_t(attr, dst_md, n, concat_dim, src_mds) {}

        pd_t(const pd_t &rhs) = default;
        ~pd_t() override = default;

        DECLARE_CONCAT_PD_T("xe:any", xe_t);

        status_t init(impl::engine_t *engine) {

            using sm = primitive_attr_t::skip_mask_t;

            VDISPATCH_CONCAT(n_inputs() <= 16, VERBOSE_BAD_PARAM, "n_inputs");
            VDISPATCH_CONCAT(attr()->has_default_values(sm::scales),
                    VERBOSE_UNSUPPORTED_ATTR);
            VDISPATCH_CONCAT_SC(set_default_params(), VERBOSE_UNSUPPORTED_TAG);
            VDISPATCH_CONCAT(memory_desc_ndims_ok(dst_md()), VERBOSE_BAD_NDIMS,
                    "dst", dst_md()->ndims);
            CHECK(init_conf(engine));
            return status::success;
        }

        status_t init_conf(impl::engine_t *engine);
        status_t init_kernel_ctx(compute::kernel_ctx_t &kernel_ctx) const;

        conf_t conf;

    protected:
        bool can_use_sub_group_size(
                const intel::engine_t *intel_engine, int sub_group_size) const;
        int calculate_sub_group_size(const intel::engine_t *intel_engine) const;
        std::pair<int, int> calculate_iter_dim_idx_chunk(int num_threads) const;
    };

    status_t init(impl::engine_t *engine) override {
        compute::kernel_ctx_t kernel_ctx;

        status_t status = pd()->init_kernel_ctx(kernel_ctx);
        CHECK(status);

        status = create_kernel(engine, &kernel, "xe_concat", kernel_ctx);
        CHECK(status);

        return status::success;
    }

    status_t execute(const exec_ctx_t &ctx) const override;

private:
    const pd_t *pd() const { return (const pd_t *)primitive_t::pd().get(); }

    compute::kernel_t kernel;
};

} // namespace concat
} // namespace intel
} // namespace gpu
} // namespace impl
} // namespace dnnl

#endif // GPU_INTEL_CONCAT_XE_HPP