aboutsummaryrefslogtreecommitdiff
path: root/src/dynamic_fusion/sketch/gpu/template_writer/cl/ClTemplateWriter.h
diff options
context:
space:
mode:
authorSiCong Li <sicong.li@arm.com>2022-08-29 18:25:51 +0100
committerSiCong Li <sicong.li@arm.com>2022-11-01 10:38:21 +0000
commitf44bbc5c697de841dce97c0f2fa39bae391a8174 (patch)
tree56468ef833726318e545043f4abcd16ad3775094 /src/dynamic_fusion/sketch/gpu/template_writer/cl/ClTemplateWriter.h
parent3394f3e3df7fd2d924c41822a8564493fc06473a (diff)
downloadComputeLibrary-f44bbc5c697de841dce97c0f2fa39bae391a8174.tar.gz
Rewrite dynamic fusion
The new version introduces the following major changes: * Change public interface to simplify and standardize the user experience - Use the term "Workload" uniformly - Simplify operator interface to be a set of static methods: validate_op(), create_op() * Separate the kernel writing into its own component (template_writer). This is to allow the co-development of GpuKernelWriter, and to allow easy replacement once GpuKernelWriter is mature. * Optimize the core fusion algorithm used by the component graph. The details can be found in GpuKernelComponentGraph::fuse() * Use Gpu instead of Cl prefixes for most of the Workload interfaces (except for runtime and kernel components, which have to be language specific) This allows the potential extension to other Gpu langauges in the future. * Refactor runtime memory interface so that auxiliary tensor handling is separate from the user tensor passing. This is because the former is less stable and may require extension in the future. * Hide source code object from the user as it is not required at the moment * Deprecate the old prototype entirely by disabling it in SCons build Resolves COMPMID-5510, COMPMID-5512, COMPMID-5513 Change-Id: If69d2362856f2de4503546b7b6cf48a525cf3079 Signed-off-by: SiCong Li <sicong.li@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/8406 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com> Reviewed-by: Jakub Sujak <jakub.sujak@arm.com> Reviewed-by: Viet-Hoa Do <viet-hoa.do@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Benchmark: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'src/dynamic_fusion/sketch/gpu/template_writer/cl/ClTemplateWriter.h')
-rw-r--r--src/dynamic_fusion/sketch/gpu/template_writer/cl/ClTemplateWriter.h92
1 files changed, 92 insertions, 0 deletions
diff --git a/src/dynamic_fusion/sketch/gpu/template_writer/cl/ClTemplateWriter.h b/src/dynamic_fusion/sketch/gpu/template_writer/cl/ClTemplateWriter.h
new file mode 100644
index 0000000000..83f617b6c6
--- /dev/null
+++ b/src/dynamic_fusion/sketch/gpu/template_writer/cl/ClTemplateWriter.h
@@ -0,0 +1,92 @@
+/*
+ * Copyright (c) 2022 Arm Limited.
+ *
+ * SPDX-License-Identifier: MIT
+ *
+ * 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 SRC_DYNAMIC_FUSION_SKETCH_GPU_TEMPLATE_WRITER_CL_CLTEMPLATEWRITER
+#define SRC_DYNAMIC_FUSION_SKETCH_GPU_TEMPLATE_WRITER_CL_CLTEMPLATEWRITER
+
+#include "src/dynamic_fusion/sketch/gpu/GpuKernelArgument.h"
+#include "src/dynamic_fusion/sketch/gpu/GpuKernelComponentGroup.h"
+#include "src/dynamic_fusion/sketch/gpu/IGpuKernelWriter.h"
+#include "src/dynamic_fusion/sketch/gpu/template_writer/GpuKernelVariableTable.h"
+
+#include <map>
+
+namespace arm_compute
+{
+namespace experimental
+{
+namespace dynamic_fusion
+{
+/** Use a templated-string-based method to write kernel code
+ * It stitches the component code templates together based on the valid fusion configuration.
+ * It then instantiates the actual kernel code from the template and the generated tag lookup table.
+ */
+class ClTemplateWriter : public IGpuKernelWriter
+{
+public:
+ /** Instantiates a kernel code string from the kernel code template
+ * @note: some tags can be unused since they could be used only for the macros, or only for the component code
+ *
+ * @param[in] code_template Kernel code template
+ * @param[in] tags Tag lookup table
+ *
+ * @return std::string Instantiated kernel string
+ */
+ static std::string replace_tags(const std::string &code_template, const TagLUT &tags);
+ /** Default constructor */
+ ClTemplateWriter() = default;
+ /** Constructor
+ *
+ * @param[in] components Kernel component group from which the kernel will be generated
+ */
+ ClTemplateWriter(const GpuKernelComponentGroup &components);
+ /** Destructor */
+ ~ClTemplateWriter() override;
+ /** Generate kernel name */
+ std::string get_name() override;
+ /** Generate kernel code */
+ std::string get_code() override;
+ /** Generate build options */
+ CLBuildOptions get_build_options() override;
+ /** Generate config id string of the entire kernel. This is used for tuning */
+ std::string get_config_id() override;
+ /** Generate execution window */
+ Window get_window() const override;
+ /** Get the kernel argument lists of the kernel*/
+ std::map<ITensorInfo::Id, GpuKernelArgument> get_tensors() override;
+
+private:
+ std::string write_kernel_name() const;
+ std::string write_code();
+ std::string write_global_section() const;
+ std::string write_argument_declaration(const GpuKernelVariableTable::TensorVariable &var) const;
+ std::string write_kernel_signature(const GpuKernelVariableTable::VariableList &argument_list) const;
+
+private:
+ GpuKernelComponentGroup _components{};
+ GpuKernelVariableTable _vtable{};
+};
+} // namespace dynamic_fusion
+} // namespace experimental
+} // namespace arm_compute
+#endif /* SRC_DYNAMIC_FUSION_SKETCH_GPU_TEMPLATE_WRITER_CL_CLTEMPLATEWRITER */