Merge commit '7330ec37ee3b8cf2d54630372dfe9e86a893e4f5' into develop

This commit is contained in:
assistant-librarian[bot]
2025-09-04 21:11:23 +00:00
parent 5677205f88
commit 7f65be1b3e
51 changed files with 3709 additions and 189 deletions

View File

@@ -5,7 +5,7 @@
#include "ck_tile/host.hpp"
#include "mask.hpp"
#include "utils.hpp"
#include "json_dump.hpp"
#include "ck_tile/utility/json_dump.hpp"
#include <array>
#include <cstring>

View File

@@ -7,7 +7,7 @@
#include "mask.hpp"
#include "rotary.hpp"
#include "utils.hpp"
#include "json_dump.hpp"
#include "ck_tile/utility/json_dump.hpp"
#include <array>
#include <cstring>

View File

@@ -1,6 +1,6 @@
#include "ck_tile/host.hpp"
#include "layernorm2d_fwd.hpp"
#include "json_dump.hpp"
#include "ck_tile/utility/json_dump.hpp"
#include <algorithm>
#include <cstring>

View File

@@ -9,7 +9,7 @@
#include "ck_tile/host/kernel_launch.hpp"
#include "ck_tile/ops/epilogue.hpp"
#include "ck_tile/ops/gemm.hpp"
#include "json_dump.hpp"
#include "ck_tile/utility/json_dump.hpp"
#define CK_TILE_PIPELINE_COMPUTE_V3 1
#define CK_TILE_PIPELINE_MEMORY 2

View File

@@ -3,7 +3,7 @@
#include "ck_tile/host.hpp"
#include "ck_tile/ops/reduce.hpp"
#include "json_dump.hpp"
#include "ck_tile/utility/json_dump.hpp"
#include <cstring>
template <typename T>

View File

@@ -13,7 +13,7 @@
#include "ck_tile/core.hpp"
#include "ck_tile/ops/reduce.hpp"
#include "topk_softmax_api.hpp"
#include "json_dump.hpp"
#include "ck_tile/utility/json_dump.hpp"
#if 0
template <typename T>

View File

@@ -1,7 +1,7 @@
#include "ck_tile/host.hpp"
#include "rmsnorm2d_fwd.hpp"
#include <cstring>
#include "json_dump.hpp"
#include "ck_tile/utility/json_dump.hpp"
// different threshold for different dtype
template <typename DataType>

View File

@@ -1,7 +1,7 @@
#include "ck_tile/host.hpp"
#include "add_rmsnorm2d_rdquant_fwd.hpp"
#include <cstring>
#include "json_dump.hpp"
#include "ck_tile/utility/json_dump.hpp"
// different threshold for different dtype
template <typename InputDataType>

View File

@@ -1,6 +1,6 @@
#include "ck_tile/host.hpp"
#include "smoothquant.hpp"
#include "json_dump.hpp"
#include "ck_tile/utility/json_dump.hpp"
#include <cstring>
// different threshold for different dtype

View File

@@ -14,7 +14,7 @@
#include "ck_tile/core.hpp"
#include "ck_tile/ops/reduce.hpp"
#include "moe_sorting_api.hpp"
#include "json_dump.hpp"
#include "ck_tile/utility/json_dump.hpp"
auto create_args(int argc, char* argv[])
{

View File

@@ -1,6 +1,6 @@
#include "ck_tile/host.hpp"
#include "moe_smoothquant.hpp"
#include "json_dump.hpp"
#include "ck_tile/utility/json_dump.hpp"
#include <cstring>
#include <set>

View File

@@ -5,7 +5,7 @@
#include <set>
#include "ck_tile/host.hpp"
#include "json_dump.hpp"
#include "ck_tile/utility/json_dump.hpp"
#include "fused_moe.hpp"
// different threshold for different dtype

View File

@@ -9,7 +9,7 @@
#include "ck_tile/host/kernel_launch.hpp"
#include "ck_tile/ops/gemm/kernel/batched_gemm_kernel.hpp"
#include "ck_tile/ops/elementwise/unary_element_wise_operation.hpp"
#include <json_dump.hpp>
#include "ck_tile/utility/json_dump.hpp"
#define CK_TILE_PIPELINE_COMPUTE_V3 1
#define CK_TILE_PIPELINE_MEMORY 2

View File

@@ -9,7 +9,7 @@
#include "ck_tile/host/kernel_launch.hpp"
#include "ck_tile/ops/gemm.hpp"
#include "ck_tile/ops/elementwise/unary_element_wise_operation.hpp"
#include "json_dump.hpp"
#include "ck_tile/utility/json_dump.hpp"
#define CK_TILE_PIPELINE_COMPUTE_V3 1
#define CK_TILE_PIPELINE_MEMORY 2

View File

@@ -2,7 +2,7 @@
// Copyright (c) 2024-2025, Advanced Micro Devices, Inc. All rights reserved.
#pragma once
#include <type_traits>
#include "json_dump.hpp"
#include "ck_tile/utility/json_dump.hpp"
template <typename T>
constexpr const char* DataTypeToString()
{

View File

@@ -3,7 +3,7 @@
#pragma once
#include <cstddef>
#include "json_dump.hpp"
#include "ck_tile/utility/json_dump.hpp"
template <typename ADataType,
typename BDataType,

View File

@@ -4,7 +4,7 @@
#include "ck_tile/host.hpp"
#include "ck_tile/ops/elementwise.hpp"
#include "ck_tile/host/reference/reference_elementwise.hpp"
#include "json_dump.hpp"
#include "ck_tile/utility/json_dump.hpp"
#include "elementwise_common.hpp"
auto create_args(int argc, char* argv[])

View File

@@ -4,7 +4,7 @@
#include "ck_tile/host.hpp"
#include "ck_tile/ops/elementwise.hpp"
#include "ck_tile/host/reference/reference_elementwise.hpp"
#include "json_dump.hpp"
#include "ck_tile/utility/json_dump.hpp"
#include "elementwise_common.hpp"
auto create_args(int argc, char* argv[])

View File

@@ -4,7 +4,7 @@
#include "ck_tile/host.hpp"
#include "ck_tile/ops/elementwise.hpp"
#include "ck_tile/host/reference/reference_transpose.hpp"
#include "json_dump.hpp"
#include "ck_tile/utility/json_dump.hpp"
#include "elementwise_common.hpp"
auto create_args(int argc, char* argv[])

View File

@@ -4,7 +4,7 @@
#include "ck_tile/host.hpp"
#include "ck_tile/ops/elementwise.hpp"
#include "ck_tile/host/reference/reference_elementwise.hpp"
#include "json_dump.hpp"
#include "ck_tile/utility/json_dump.hpp"
#include "elementwise_common.hpp"
auto create_args(int argc, char* argv[])

View File

@@ -12,7 +12,7 @@
#include "batched_transpose_example.hpp"
#include "json_dump.hpp"
#include "ck_tile/utility/json_dump.hpp"
#if 0
template <typename T>
void dump_host_tensor_4d(const ck_tile::HostTensor<T>& x)