Skip to content

[CPU: Bugfix] fix the problem mentioned in #4454 #4552

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
Sherlockzhangjinge wants to merge 3 commits into
base: master
Choose a base branch
from
+8 −3
Open

[CPU: Bugfix] fix the problem mentioned in #4454 #4552

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
700bacf
[CPU:Bugfix] supplement missing headfile, and fix a bug in MNNLineDep…
Sherlockzhangjinge May 18, 2026
cd3a9bf
[CPU:Bugfix] Fix the bug in MNNLineDepthWiseInt8AddBiasScaleUnit
Sherlockzhangjinge May 24, 2026
860017d
[CPU:Bugfix] Fix remaining conflict markers
Sherlockzhangjinge Jun 19, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions source/backend/cpu/riscv/rvv/MNNAvgPoolInt8.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,6 +9,7 @@
#include <stdint.h>
#include <sys/types.h>
#include <limits.h>
#include "../../compute/Int8FunctionsOpt.h"

void MNNAvgPoolInt8_RVV(int8_t* dst, int8_t* src, size_t outputWidth, size_t inputWidth, size_t kernelx, size_t kernely,
size_t stridesx, ssize_t paddingx, ssize_t factor) {
Expand Down
4 changes: 4 additions & 0 deletions source/backend/cpu/riscv/rvv/MNNFloat2Int8.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7,7 +7,11 @@
//
#include <riscv_vector.h>
#include <stdint.h>
<<<<<<< HEAD
#include <sys/types.h>
=======
#include "../../compute/Int8FunctionsOpt.h"
>>>>>>> fc26945b ([CPU:Bugfix] supplement missing headfile, and fix a bug in MNNLineDepthWiseInt8AddBiasScaleUnit which causse the error below:)
Comment thread
Sherlockzhangjinge marked this conversation as resolved.
Outdated

void MNNFloat2Int8_RVV(const float* src, int8_t* dst, size_t sizeQuad, const float* scalep, ssize_t minValue,
ssize_t maxValue, const float* zeroPoint, ssize_t quanParamVec) {
Expand Down
1 change: 1 addition & 0 deletions source/backend/cpu/riscv/rvv/MNNInt8ScaleToFloat.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -10,6 +10,7 @@
#include <stdint.h>
#include <sys/types.h>
#include <string.h>
#include "../../compute/Int8FunctionsOpt.h"

void MNNInt8ScaleToFloat_RVV(float* dst, const int8_t* src, const float* scale, size_t size, const float* zeroPoint,
ssize_t quantParamVec) {
Expand Down
10 changes: 8 additions & 2 deletions source/backend/cpu/riscv/rvv/MNNLineDepthWiseInt8AddBiasScaleUnit.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -20,7 +20,11 @@ void MNNLineDepthWiseInt8AddBiasScaleUnit_RVV(int8_t* dst, const int8_t* src, co
const int8_t* srcPtr = src;
const int8_t* weightPtr = weight;

<<<<<<< HEAD
Comment thread
Sherlockzhangjinge marked this conversation as resolved.
Outdated
const float* bias_z = (const float*)parameters->bias;
=======
const int32_t* bias_z = parameters->bias;
>>>>>>> d7f9d697 ([CPU:Bugfix] Fix the bug in MNNLineDepthWiseInt8AddBiasScaleUnit )
const float* scale_z = parameters->scale;
const int32_t max_val = parameters->maxValue;
const int32_t min_val = parameters->minValue;
Expand Down Expand Up @@ -55,7 +59,9 @@ void MNNLineDepthWiseInt8AddBiasScaleUnit_RVV(int8_t* dst, const int8_t* src, co
}
}

vfloat32m4_t vec_bias = __riscv_vle32_v_f32m4(bias_z, vl);
// vfloat32m4_t vec_bias = __riscv_vle32_v_f32m4(bias_z, vl);
vint32m4_t vec_bias_int = __riscv_vle32_v_i32m4(bias_z, vl);
vfloat32m4_t vec_bias = __riscv_vfcvt_f_x_v_f32m4(vec_bias_int, vl);
vfloat32m4_t f_sum = __riscv_vfcvt_f_x_v_f32m4(vec_sum, vl);
f_sum = __riscv_vfadd_vv_f32m4(f_sum, vec_bias, vl);

Expand All @@ -71,4 +77,4 @@ void MNNLineDepthWiseInt8AddBiasScaleUnit_RVV(int8_t* dst, const int8_t* src, co

__riscv_vse8_v_i8m1(dst_x, v_out, vl);
}
}
}
Loading