mindspore 2.4.10__cp39-cp39-manylinux1_x86_64.whl → 2.5.0__cp39-cp39-manylinux1_x86_64.whl

mindspore/lib/plugin/ascend/ms_kernels_internal/internal_kernel/include/apply_rotary_pos_emb_nz/kernel/apply_rotary_pos_emb_nz_fp32.h

@@ -1,230 +0,0 @@
-/**
- * Copyright 2024 Huawei Technologies Co., Ltd
- *
- * 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 ROTARY_POS_EMB_NZ_FP32
-#define ROTARY_POS_EMB_NZ_FP32
-#include "apply_rotary_pos_emb_nz_base.h"
-template <typename QK_DTYPE, typename COS_DTYPE, bool IF_COS_BROADCAST>
-class RopeFp32 : public RopeBase<QK_DTYPE, COS_DTYPE, IF_COS_BROADCAST> {
- public:
-  __aicore__ inline RopeFp32(RopeTilingNzData *tilingData) : RopeBase<QK_DTYPE, COS_DTYPE, IF_COS_BROADCAST>(tilingData) {
-    this->repeatSize_ = 64;                   // 64 = 256B / sizeof(float)
-    this->maxProcessNum_ = 3 * MAX_LEN_FP16;  // 3 for fp16 space needed
-    this->repeatTimesQ_ = (this->tilingData_->hiddenSizeQ + this->repeatSize_ - 1) / this->repeatSize_;
-    this->repeatTimesK_ = (this->tilingData_->hiddenSizeK + this->repeatSize_ - 1) / this->repeatSize_;
-    headDimAlign_ = ((this->tilingData_->headDim + ELE_NUM_FP32 - 1) / ELE_NUM_FP32) * ELE_NUM_FP32;
-    this->alignHalfHeadDim_ = (this->rotateStride_ * NUM_TWO) % ELE_NUM_FP32;
-    this->hiddenSizeAlign_ = ((this->hiddenSize_ + this->repeatSize_ - 1) / this->repeatSize_) * this->repeatSize_;
-    this->syncOffset_ =
-      (this->tilingData_->headDim % ELE_NUM_FP16 == 0) ? this->hiddenSizeAlign_ : this->headNum_ * headDimAlign_;
-    this->offsetExtraGm_ = NUM_TWO * block_idx * this->syncOffset_;
-
-    sliceSizeTmp_ = (SLICE_SIZE / this->tilingData_->headDim) * this->tilingData_->headDim;  // 向下取整
-
-    // fp16
-    this->oriPos_ = 0;
-    this->removeBefore_ = this->oriPos_ + sliceSizeTmp_;
-    this->padBefore_ = this->removeBefore_ + sliceSizeTmp_;
-    resOut_ = this->padBefore_ + sliceSizeTmp_;
-
-    // fp32
-    this->cosPad_ = 0;
-    this->sinPad_ = this->cosPad_ + sliceSizeTmp_;
-    this->negOne_ = this->sinPad_ + sliceSizeTmp_;
-    oriPosF32_ = this->negOne_ + sliceSizeTmp_;
-    PadBeforeF32_ = oriPosF32_ + sliceSizeTmp_;
-    removeBeforeF32_ = PadBeforeF32_ + sliceSizeTmp_;
-    posOneF32_ = removeBeforeF32_ + sliceSizeTmp_;
-    resOutFp32_ = posOneF32_ + sliceSizeTmp_;
-
-    this->pipe_.InitBuffer(qkfp32QueueCO2_, 1,
-                           (this->tilingData_->maxUbSize - this->batchSize_ * NUM_TWO -
-                            this->maxProcessNum_ * sizeof(QK_DTYPE)));  // 留給fp32的
-    AscendC::LocalTensor<COS_DTYPE> qkfp32_perloop_ub = qkfp32QueueCO2_.AllocTensor<COS_DTYPE>();
-    qkfp32Ubuf_ = (__ubuf__ COS_DTYPE *)qkfp32_perloop_ub.GetPhyAddr();
-    this->pipe_.InitBuffer(outQueueCO2_, 1, ((this->maxProcessNum_) * sizeof(QK_DTYPE)));
-    AscendC::LocalTensor<QK_DTYPE> cache_perloop_ub1 = outQueueCO2_.AllocTensor<QK_DTYPE>();
-    commonUbuf_ = (__ubuf__ QK_DTYPE *)cache_perloop_ub1.GetPhyAddr();
-
-    // 判断
-    if (this->tilingData_->hiddenSizeQ > sliceSizeTmp_) {
-      sliceTimeQ_ = (this->tilingData_->hiddenSizeQ + sliceSizeTmp_ - 1) / sliceSizeTmp_;    // 向上取整
-      lastSliceSizeQ_ = this->tilingData_->hiddenSizeQ - (sliceTimeQ_ - 1) * sliceSizeTmp_;  // 向上取整
-    } else {
-      sliceTimeQ_ = 1;
-      lastSliceSizeQ_ = this->tilingData_->hiddenSizeQ;
-    }
-
-    if (this->tilingData_->hiddenSizeK > sliceSizeTmp_) {
-      sliceTimeK_ = (this->tilingData_->hiddenSizeK + sliceSizeTmp_ - 1) / sliceSizeTmp_;  // 向上取整
-      lastSliceSizeK_ = this->tilingData_->hiddenSizeK - (sliceTimeK_ - 1) * sliceSizeTmp_;
-    } else {
-      sliceTimeK_ = 1;
-      lastSliceSizeK_ = this->tilingData_->hiddenSizeK;
-    }
-  }
-
-  template <typename T>
-  __aicore__ inline void CastQKDType2F32(uint32_t repeatTimes) {
-    vconv_f162f32(qkfp32Ubuf_ + oriPosF32_, commonUbuf_ + this->oriPos_, repeatTimes, 1, 1, DEFAULT_REPEAT_STRIDE,
-                  DEFAULT_REPEAT_STRIDE / NUM_TWO);
-    vconv_f162f32(qkfp32Ubuf_ + removeBeforeF32_, commonUbuf_ + this->removeBefore_, repeatTimes, 1, 1,
-                  DEFAULT_REPEAT_STRIDE, DEFAULT_REPEAT_STRIDE / NUM_TWO);
-    vconv_f162f32(qkfp32Ubuf_ + PadBeforeF32_, commonUbuf_ + this->padBefore_, repeatTimes, 1, 1, DEFAULT_REPEAT_STRIDE,
-                  DEFAULT_REPEAT_STRIDE / NUM_TWO);
-  }
-
-#if defined(__CCE_KT_TEST__) || (__CCE_AICORE__ == 220)
-  template <>
-  __aicore__ inline void CastQKDType2F32<bfloat16_t>(uint32_t repeatTimes) {
-    vconv_bf162f32(qkfp32Ubuf_ + oriPosF32_, commonUbuf_ + this->oriPos_, repeatTimes, 1, 1, DEFAULT_REPEAT_STRIDE,
-                   DEFAULT_REPEAT_STRIDE / NUM_TWO);
-    vconv_bf162f32(qkfp32Ubuf_ + removeBeforeF32_, commonUbuf_ + this->removeBefore_, repeatTimes, 1, 1,
-                   DEFAULT_REPEAT_STRIDE, DEFAULT_REPEAT_STRIDE / NUM_TWO);
-    vconv_bf162f32(qkfp32Ubuf_ + PadBeforeF32_, commonUbuf_ + this->padBefore_, repeatTimes, 1, 1,
-                   DEFAULT_REPEAT_STRIDE, DEFAULT_REPEAT_STRIDE / NUM_TWO);
-  }
-#endif
-
-  template <typename T>
-  __aicore__ inline void CastF322QKDType(__gm__ QK_DTYPE *dst, __ubuf__ QK_DTYPE *src1, __ubuf__ float *src,
-                                         uint32_t repeatTimes, uint32_t hiddenSize1) {
-    vconv_f322f16(src1, src, repeatTimes, 1, 1, DEFAULT_REPEAT_STRIDE / NUM_TWO, DEFAULT_REPEAT_STRIDE);
-    set_flag(PIPE_V, PIPE_MTE3, EVENT_ID1);
-
-    wait_flag(PIPE_V, PIPE_MTE3, EVENT_ID1);
-    copy_ubuf_to_gm(dst, src1, 0, 1, hiddenSize1 / ELE_NUM_FP16, 0, 0);
-  }
-
-#if defined(__CCE_KT_TEST__) || (__CCE_AICORE__ == 220)
-  template <>
-  __aicore__ inline void CastF322QKDType<bfloat16_t>(__gm__ QK_DTYPE *dst, __ubuf__ QK_DTYPE *src1, __ubuf__ float *src,
-                                                     uint32_t repeatTimes, uint32_t hiddenSize1) {
-    vconv_f322bf16r(src1, src, repeatTimes, 1, 1, DEFAULT_REPEAT_STRIDE / NUM_TWO, DEFAULT_REPEAT_STRIDE);
-    set_flag(PIPE_V, PIPE_MTE3, EVENT_ID1);
-
-    wait_flag(PIPE_V, PIPE_MTE3, EVENT_ID1);
-    copy_ubuf_to_gm(dst, src1, 0, 1, hiddenSize1 / ELE_NUM_FP16, 0, 0);
-  }
-#endif
-
-
-  __aicore__ inline void Process(__gm__ uint8_t *extraGm) {
-    if (this->tilingData_->cosFormat == 1) {
-      pipe_barrier((PIPE_ALL));
-      this->ExpandCosSin(qkfp32Ubuf_, this->cosGm_, (__gm__ COS_DTYPE *)extraGm);
-      this->cosGm_ = (__gm__ COS_DTYPE *)extraGm;
-      pipe_barrier((PIPE_ALL));
-      this->ExpandCosSin(qkfp32Ubuf_, this->sinGm_,
-                         (__gm__ COS_DTYPE *)extraGm + this->tilingData_->ntokens * this->tilingData_->headDim);
-      this->sinGm_ = (__gm__ COS_DTYPE *)extraGm + this->tilingData_->ntokens * this->tilingData_->headDim;
-      extraGm =
-        extraGm + this->tilingData_->ntokens * this->tilingData_->headDim * 4;  // sizeof(uint8_t) * 2 = sizeof(half)
-      pipe_barrier((PIPE_ALL));
-    }
-    uint32_t headNumTempQ = this->tilingData_->hiddenSizeQ > sliceSizeTmp_
-                              ? (sliceSizeTmp_ / this->tilingData_->headDim)
-                              : this->tilingData_->headNumQ;
-    uint32_t dynamicSliceQ =
-      this->tilingData_->hiddenSizeQ > sliceSizeTmp_ ? sliceSizeTmp_ : this->tilingData_->hiddenSizeQ;
-    uint32_t headNumTempK = this->tilingData_->hiddenSizeK > sliceSizeTmp_
-                              ? (sliceSizeTmp_ / this->tilingData_->headDim)
-                              : this->tilingData_->headNumK;
-    uint32_t dynamicSliceK =
-      this->tilingData_->hiddenSizeK > sliceSizeTmp_ ? sliceSizeTmp_ : this->tilingData_->hiddenSizeK;
-    uint32_t repeatTemp = (dynamicSliceQ + this->repeatSize_ - 1) / this->repeatSize_;
-    this->ExpandNeg(qkfp32Ubuf_, posOneF32_, headNumTempQ, repeatTemp);
-    for (uint32_t zz = 0; zz < this->dynamicRound_; ++zz) {  // 每个核 核内只需拷贝一次cos sin（每个头cos sin都一样）
-      this->CosSinBroadcast(extraGm, zz, qkfp32Ubuf_,
-                            dynamicSliceQ);  // 一次放得下：hiddensize， 一次放不下：sliceSizeTmp
-      for (uint32_t perSlice = 0; perSlice < sliceTimeQ_; ++perSlice) {  // 核内每块
-        uint32_t dynamicSliceQTemp = (perSlice == sliceTimeQ_ - 1) ? lastSliceSizeQ_ : sliceSizeTmp_;
-        headNumTempQ = dynamicSliceQTemp / this->tilingData_->headDim;
-        uint32_t repeatTimeOnce = (dynamicSliceQTemp + this->repeatSize_ - 1) / this->repeatSize_;
-        this->QkComm(this->qGm_ + block_idx * this->nlCoreRun_ * this->tilingData_->hiddenSizeQ +
-                       zz * this->tilingData_->hiddenSizeQ + perSlice * sliceSizeTmp_,
-                     extraGm, dynamicSliceQTemp, commonUbuf_, headNumTempQ);
-        if (this->alignRotary_ == 0) {
-          pipe_barrier((PIPE_V));
-          CastQKDType2F32<QK_DTYPE>(repeatTimeOnce);
-
-          pipe_barrier((PIPE_V));
-          this->CalcRopeAlign(qkfp32Ubuf_, repeatTimeOnce, oriPosF32_, removeBeforeF32_, resOutFp32_);
-        } else {
-          set_flag(PIPE_MTE2, PIPE_V, EVENT_ID1);
-          wait_flag(PIPE_MTE2, PIPE_V, EVENT_ID1);
-
-          CastQKDType2F32<QK_DTYPE>(repeatTimeOnce);
-          pipe_barrier((PIPE_V));
-          this->CalcRope(qkfp32Ubuf_, repeatTimeOnce, oriPosF32_, removeBeforeF32_, PadBeforeF32_, posOneF32_,
-                         resOutFp32_);
-
-        }  // 帶PIPE_V
-        CastF322QKDType<QK_DTYPE>(this->outQGm_ + block_idx * this->nlCoreRun_ * this->tilingData_->hiddenSizeQ +
-                                    zz * this->tilingData_->hiddenSizeQ + perSlice * sliceSizeTmp_,
-                                  commonUbuf_ + resOut_, qkfp32Ubuf_ + resOutFp32_, repeatTimeOnce, dynamicSliceQTemp);
-        pipe_barrier(PIPE_ALL);
-      }
-      for (uint32_t perSlice = 0; perSlice < sliceTimeK_; ++perSlice) {  // 核内每块
-        uint32_t dynamicSliceKTemp = (perSlice == sliceTimeK_ - 1) ? lastSliceSizeK_ : sliceSizeTmp_;
-        headNumTempK = dynamicSliceKTemp / this->tilingData_->headDim;
-        uint32_t repeatTimeOnce = (dynamicSliceKTemp + this->repeatSize_ - 1) / this->repeatSize_;
-        this->QkComm(this->kGm_ + block_idx * this->nlCoreRun_ * this->tilingData_->hiddenSizeK +
-                       zz * this->tilingData_->hiddenSizeK + perSlice * sliceSizeTmp_,
-                     extraGm, dynamicSliceKTemp, commonUbuf_, headNumTempK);
-        if (this->alignRotary_ == 0) {
-          pipe_barrier((PIPE_V));
-          CastQKDType2F32<QK_DTYPE>(repeatTimeOnce);
-
-          pipe_barrier((PIPE_V));
-          this->CalcRopeAlign(qkfp32Ubuf_, repeatTimeOnce, oriPosF32_, removeBeforeF32_, resOutFp32_);
-        } else {
-          set_flag(PIPE_MTE2, PIPE_V, EVENT_ID1);
-          wait_flag(PIPE_MTE2, PIPE_V, EVENT_ID1);
-          CastQKDType2F32<QK_DTYPE>(repeatTimeOnce);
-
-          pipe_barrier((PIPE_V));
-          this->CalcRope(qkfp32Ubuf_, repeatTimeOnce, oriPosF32_, removeBeforeF32_, PadBeforeF32_, posOneF32_,
-                         resOutFp32_);
-        }  // 帶PIPE_V
-
-        CastF322QKDType<QK_DTYPE>(this->outKGm_ + block_idx * this->nlCoreRun_ * this->tilingData_->hiddenSizeK +
-                                    zz * this->tilingData_->hiddenSizeK + perSlice * sliceSizeTmp_,
-                                  commonUbuf_ + resOut_, qkfp32Ubuf_ + resOutFp32_, repeatTimeOnce, dynamicSliceKTemp);
-        pipe_barrier(PIPE_ALL);
-      }
-    }
-  }
-
- private:
-  AscendC::TQue<AscendC::QuePosition::VECIN, 1> qkfp32QueueCO2_;
-  AscendC::TQue<AscendC::QuePosition::VECIN, 1> outQueueCO2_;
-  __ubuf__ QK_DTYPE *commonUbuf_{nullptr};
-  __ubuf__ COS_DTYPE *qkfp32Ubuf_{nullptr};
-  uint32_t oriPosF32_{0};        // fp32的buf中qk的位置
-  uint32_t PadBeforeF32_{0};     // fp32的buf中保存qk[-x : hiddensize - x]
-  uint32_t removeBeforeF32_{0};  // fp32的buf中保存qk[x : hiddensize + x]
-  uint32_t posOneF32_{0};        // fp32的buf中0 0 0 1 1 1的位置
-  uint32_t headDimAlign_;        // 对齐的headDim
-  uint32_t sliceTimeQ_;          // 切分块的次数
-  uint32_t lastSliceSizeQ_;      // 最后一块的大小
-  uint32_t sliceTimeK_;
-  uint32_t lastSliceSizeK_;
-  uint32_t sliceSizeTmp_;
-  uint32_t resOut_;
-  uint32_t resOutFp32_;
-};
-
-#endif

mindspore/lib/plugin/ascend/ms_kernels_internal/internal_kernel/include/apply_rotary_pos_emb_nz/kernel/apply_rotary_pos_emb_nz_tiling.h

@@ -1,43 +0,0 @@
-/**
- * Copyright 2024 Huawei Technologies Co., Ltd
- *
- * 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 MS_KERNELS_INTERNAL_KERNEL_ASCENDC_ROPE_NZ_TILING_DATA_H_
-#define MS_KERNELS_INTERNAL_KERNEL_ASCENDC_ROPE_NZ_TILING_DATA_H_
-
-#include <stdint.h>
-
-struct RopeTilingNzData {
-  uint32_t hiddenSizeQ{16};
-  uint32_t hiddenSizeK{16};
-  uint32_t headDim{1};  // qk头长度的最大值
-  uint32_t headNumQ{1};
-  uint32_t headNumK{1};
-  uint32_t rotaryCoeff{4};  // 旋转系数
-  uint32_t ntokens{1};      // 总token数
-  uint32_t realCore{0};     // 实际用到核数
-  uint32_t cosFormat{0};    // 是否复用cos sin
-  uint32_t batch{32};       // 几个batch
-  uint32_t maxUbSize{0};    // 最大UB内存
-  uint32_t tilingId{0};
-
-  uint32_t seqLen;
-  uint32_t broadCastCos{0};
-  uint32_t posDtype;
-  uint32_t posSize;
-  uint32_t maxSeqLen;
-};
-
-#endif

mindspore/lib/plugin/ascend/ms_kernels_internal/internal_kernel/include/apply_rotary_pos_emb_nz/kernel/apply_rotary_pos_emb_nz_value.h

@@ -1,27 +0,0 @@
-/**
- * Copyright (c) Huawei Technologies Co., Ltd. 2024. All rights reserved.
- *
- * 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 COMMON_NZ_VAL_H
-#define COMMON_NZ_VAL_H
-const constexpr uint32_t NUM_TWO = 2; // 2
-const constexpr uint32_t BLK_SIZE = 32; // 一个block字节数
-const constexpr uint32_t ELE_NUM_FP16 = 16; // 一个block fp16元素个数
-const constexpr uint32_t ELE_NUM_FP32 = 8; // 一个block字节数 fp32元素个数
-const constexpr uint32_t MAX_LEN_FP16 = 8192; // 非fp16情况下最大长度（hiddensize）
-const constexpr uint8_t DEFAULT_REPEAT_STRIDE = 8; // 默认stride, 8 * 32 = 256
-const constexpr int64_t REG_910B = 48; // 饱和模式寄存器位置
-const constexpr int64_t REG_310P = 53; // 饱和模式寄存器位置
-const constexpr int64_t SLICE_SIZE = 4096; // 切片大小
-#endif

mindspore/lib/plugin/ascend/ms_kernels_internal/internal_kernel/include/asdop/asd_op_impl.h

@@ -1,74 +0,0 @@
-/**
- * Copyright 2023-2024 Huawei Technologies Co., Ltd
- *
- * 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 MS_KERNELS_INTERNAL_KERNEL_ASDOP_IMPL_H_
-#define MS_KERNELS_INTERNAL_KERNEL_ASDOP_IMPL_H_
-#include "asdops/op_desc.h"
-#include "asdops/operation.h"
-#include "asdops/run_info.h"
-#include "asdops/tactic.h"
-#include "asdops/tensor.h"
-#include "internal_kernel.h"
-#include <unordered_map>
-namespace mindspore {
-namespace internal {
-
-class AsdOpsImpl : public InternelKernelImpl {
- public:
-  AsdOpsImpl(const OpParamPtr &param) : InternelKernelImpl(param){};
-  virtual ~AsdOpsImpl() = default;
-  bool Init(const ValidateInfo &info) override;
-  bool InitPagedAttention910(const ValidateInfo &info);
-  void SetInputs(const std::vector<Tensor *> &inputs) override;
-  void SetAsd910PagedAttentionC8Inputs(const std::vector<Tensor *> &inputs);
-  void SetOutputs(const std::vector<Tensor *> &outputs) override;
-  void SetWorkSpace(const std::vector<DeviceRawBuf> &workspace) override;
-  void SetStream(const void *stream_ptr) override;
-  void SetDeviceTilingBuf(const DeviceRawBuf &tilingBuf) override;
-  int Launch() override;
-  size_t GetTilingBufSize() override;
-  int Tiling(HostRawBuf &tilingBuf) override;
-  std::vector<uint64_t> GetWorkSpaceSize() override;
-  int InferShape(const std::vector<DIMS> &input_shapes, std::vector<DIMS> &output_shapes) override;
-  std::string GetOpName() override { return tactic_->GetName(); }
-  uint32_t GetLaunchCoreNum() override {
-    auto &kernelInfo = cache_info_.run_info_.GetKernelInfo();
-    return kernelInfo.GetBlockDim();
-  }
-  void UpdateParam(const OpParamPtr &param) override;
-
- private:
-  AsdOps::Tactic *InitAndGetTactic();
-
- protected:
-  AsdOps::Tactic *tactic_ = nullptr;
-  AsdOps::Operation *op_ = nullptr;
-  AsdOps::LaunchParam launch_param_;
-  AsdOps::OpDesc op_desc_;
-  bool validated_ = false;
-  std::string soc_{"Ascend910B4"};
-};
-
-class AsdOps310PImpl : public AsdOpsImpl {
- public:
-  AsdOps310PImpl(const OpParamPtr &param) : AsdOpsImpl(param){};
-  virtual ~AsdOps310PImpl() = default;
-  void SetInputs(const std::vector<Tensor *> &inputs) override;
-  void SetOutputs(const std::vector<Tensor *> &outputs) override;
-};
-
-}  // namespace internal
-}  // namespace mindspore
-#endif

mindspore/lib/plugin/ascend/ms_kernels_internal/internal_kernel/include/backend_param.h

@@ -1,74 +0,0 @@
-/**
- * Copyright 2023-2024 Huawei Technologies Co., Ltd
- *
- * 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 BACKEND_PARAM_H_
-#define BACKEND_PARAM_H_
-namespace mindspore {
-namespace internal {
-
-struct HardwareInfo {
-  uint32_t coreNum{0};
-  uint32_t l2Size{0};
-  uint32_t l1Size{0};
-  uint32_t l0aSize{0};
-  uint32_t l0bSize{0};
-  uint32_t l0cSize{0};
-  uint32_t hbmBandWidth{1};
-  uint32_t l2BandWidth{5};
-  uint32_t ubSize{0};
-};
-
-static void GetHardwareInfoPPMatmul910B1(HardwareInfo &hwInfo) {
-  hwInfo.coreNum = 24;
-  hwInfo.l2Size = 201326592;
-  hwInfo.l1Size = 524288;
-  hwInfo.l0aSize = 65536;
-  hwInfo.l0bSize = 65536;
-  hwInfo.l0cSize = 131072;
-  hwInfo.ubSize = 196608;
-}
-
-static void GetHardwareInfoPPMatmul910B2(HardwareInfo &hwInfo) {
-  hwInfo.coreNum = 24;
-  hwInfo.l2Size = 201326592;
-  hwInfo.l1Size = 524288;
-  hwInfo.l0aSize = 65536;
-  hwInfo.l0bSize = 65536;
-  hwInfo.l0cSize = 131072;
-  hwInfo.ubSize = 196608;
-}
-
-static void GetHardwareInfoPPMatmul910B3(HardwareInfo &hwInfo) {
-  hwInfo.coreNum = 20;
-  hwInfo.l2Size = 201326592;
-  hwInfo.l1Size = 524288;
-  hwInfo.l0aSize = 65536;
-  hwInfo.l0bSize = 65536;
-  hwInfo.l0cSize = 131072;
-  hwInfo.ubSize = 196608;
-}
-
-static void GetHardwareInfoPPMatmul910B4(HardwareInfo &hwInfo) {
-  hwInfo.coreNum = 20;
-  hwInfo.l2Size = 100663296;
-  hwInfo.l1Size = 524288;
-  hwInfo.l0aSize = 65536;
-  hwInfo.l0bSize = 65536;
-  hwInfo.l0cSize = 131072;
-  hwInfo.ubSize = 196608;
-}
-}  // namespace internal
-}  // namespace mindspore
-#endif

mindspore/lib/plugin/ascend/ms_kernels_internal/internal_kernel/include/cast/cast_impl.h

@@ -1,48 +0,0 @@
-/**
- * Copyright 2024 Huawei Technologies Co., Ltd
- *
- * 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 MS_KERNELS_INTERNAL_KERNEL_CAST_IMPL_H_
-#define MS_KERNELS_INTERNAL_KERNEL_CAST_IMPL_H_
-
-#include <vector>
-#include "include/internal_kernel.h"
-
-namespace mindspore {
-namespace internal {
-class CastImpl : public InternelKernelImpl {
- public:
-  CastImpl(const OpParamPtr &param) : InternelKernelImpl(param) {}
-  virtual ~CastImpl() {}
-  bool Init(const ValidateInfo &info) override;
-  void SetStream(const void *stream_ptr) override;
-  void SetDeviceTilingBuf(const DeviceRawBuf &tilingBuf) override;
-  int Launch() override;
-  uint64_t GetTilingBufSize() override;
-  int Tiling(HostRawBuf &tilingBuf) override;
-  std::vector<uint64_t> GetWorkSpaceSize() override;
-  int InferShape(const std::vector<DIMS> &input_shapes, std::vector<DIMS> &output_shapes) override;
-
- private:
-  int32_t GetCastDType();
-  int32_t GetMaxUbCount(uint32_t cast_type);
-  void DoUbTiling(uint32_t factor, uint32_t total_num, uint32_t &ub_num, uint32_t &ub_loop, uint32_t &ub_tail);
-
- private:
-  DeviceRawBuf tiling_buf_;
-  void *stream_ptr_ = nullptr;
-};
-}  // namespace internal
-}  // namespace mindspore
-#endif  // MS_KERNELS_INTERNAL_KERNEL_CAST_IMPL_H_

mindspore/lib/plugin/ascend/ms_kernels_internal/internal_kernel/include/cast/kernel/cast_kernel.h

@@ -1,21 +0,0 @@
-/**
- * Copyright 2024 Huawei Technologies Co., Ltd
- *
- * 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 MS_KERNELS_INTERNAL_KERNEL_ASCENDC_CAST_KERNEL_H_
-#define MS_KERNELS_INTERNAL_KERNEL_ASCENDC_CAST_KERNEL_H_
-void cast_do(uint32_t block_dim, void *l2ctrl, void *stream, uint8_t *in, uint8_t *out, uint8_t *tiling,
-             uint32_t cast_type);
-#endif  // MS_KERNELS_INTERNAL_KERNEL_ASCENDC_CAST_KERNEL_H_

mindspore/lib/plugin/ascend/ms_kernels_internal/internal_kernel/include/compare/compare_impl.h

@@ -1,55 +0,0 @@
-/**
- * Copyright 2024 Huawei Technologies Co., Ltd
- *
- * 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 MS_KERNELS_INTERNAL_KERNEL_COMPARE_IMPL_H_
-#define MS_KERNELS_INTERNAL_KERNEL_COMPARE_IMPL_H_
-
-#include <vector>
-#include "include/internal_kernel.h"
-#include "src/utils/elewise_utils.h"
-
-namespace mindspore {
-namespace internal {
-class CompareImpl : public InternelKernelImpl {
- public:
-  CompareImpl(const OpParamPtr &param) : InternelKernelImpl(param) { GetCompareMode(); }
-  virtual ~CompareImpl() {}
-  bool Init(const ValidateInfo &info) override;
-  void SetStream(const void *stream_ptr) override;
-  void SetDeviceTilingBuf(const DeviceRawBuf &tilingBuf) override;
-  int Launch() override;
-  uint64_t GetTilingBufSize() override;
-  int Tiling(HostRawBuf &tilingBuf) override;
-  std::vector<uint64_t> GetWorkSpaceSize() override;
-  int InferShape(const std::vector<DIMS> &input_shapes, std::vector<DIMS> &output_shapes) override;
-  bool IsSupported() override;
-
- private:
-  int32_t GetMaxUbCount(uint32_t in_dtype);
-  void GetCompareMode();
-
- private:
-  DeviceRawBuf tiling_buf_;
-  void *stream_ptr_ = nullptr;
-  uint32_t compare_mode_{0};
-  uint32_t broadcast_mode_{0};
-  size_t ndims_{0};
-  int64_t in0_shape_[MAX_COMPARE_SHAPE_LEN];
-  int64_t in1_shape_[MAX_COMPARE_SHAPE_LEN];
-  int64_t out_shape_[MAX_COMPARE_SHAPE_LEN];
-};
-}  // namespace internal
-}  // namespace mindspore
-#endif  // MS_KERNELS_INTERNAL_KERNEL_COMPARE_IMPL_H_

mindspore/lib/plugin/ascend/ms_kernels_internal/internal_kernel/include/compare/compare_tiling.h

@@ -1,27 +0,0 @@
-/**
- * Copyright 2024 Huawei Technologies Co., Ltd
- *
- * 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 MS_KERNELS_INTERNAL_KERNEL_ASCENDC_COMPARE_TILING_H_
-#define MS_KERNELS_INTERNAL_KERNEL_ASCENDC_COMPARE_TILING_H_
-#include "utils/elewise_tiling.h"
-namespace mindspore::internal {
-struct CompareTilingData : public ElewiseTailTilingData {
-  uint32_t input_dtype{0};
-  uint32_t broadcast_mode{0};
-  uint32_t compare_mode{0};
-};
-}  // namespace mindspore::internal
-#endif  // MS_KERNELS_INTERNAL_KERNEL_ASCENDC_COMPARE_TILING_H_

mindspore/lib/plugin/ascend/ms_kernels_internal/internal_kernel/include/compare/kernel/compare_kernel.h

@@ -1,23 +0,0 @@
-/**
- * Copyright 2024 Huawei Technologies Co., Ltd
- *
- * 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 MS_KERNELS_INTERNAL_KERNEL_ASCENDC_COMPARE_KERNEL_H_
-#define MS_KERNELS_INTERNAL_KERNEL_ASCENDC_COMPARE_KERNEL_H_
-#include <stdint.h>
-#define BIT_SIZE 8
-void compare(uint32_t blockdim, void *l2ctrl, void *stream, uint8_t *in1, uint8_t *in2, uint8_t *out, uint8_t *tiling,
-             uint32_t dtype, uint32_t broadcast_mode, uint32_t compare_mode);
-#endif  // MS_KERNELS_INTERNAL_KERNEL_ASCENDC_NOT_EQUAL_KERNEL_H_

mindspore/lib/plugin/ascend/ms_kernels_internal/internal_kernel/include/elewise_binary/and_impl.h

@@ -1,29 +0,0 @@
-/**
- * Copyright 2024 Huawei Technologies Co., Ltd
- *
- * 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 MS_KERNELS_INTERNAL_KERNEL_AND_IMPL_H_
-#define MS_KERNELS_INTERNAL_KERNEL_AND_IMPL_H_
-#include "elewise_binary_impl.h"
-namespace mindspore {
-namespace internal {
-class AndImpl : public ElewiseBinaryImpl {
- public:
-  AndImpl(const OpParamPtr &param) : ElewiseBinaryImpl(param) {}
-  virtual ~AndImpl() {}
-  int Launch() override;
-};
-}  // namespace internal
-}  // namespace mindspore
-#endif  // MS_KERNELS_INTERNAL_KERNEL_AND_IMPL_H_