From 2aea22cae18b369a09dd024b8da1baac3fb7c22b Mon Sep 17 00:00:00 2001
From: KateJu <153474223+kateju12@users.noreply.github.com>
Date: Thu, 23 Apr 2026 22:08:50 +0800
Subject: [PATCH] Fix per-layer conv2d int8 CPU verification reference path
 (#6656)

case example_conv2d_fwd_xdl_perlayer_quantization_int8.exe 1 0

## Motivation

<!-- Explain the purpose of this PR and the goals it aims to achieve.
-->

## Technical Details

<!-- Explain the changes along with any relevant GitHub links. -->

## Test Plan

<!-- Explain any relevant testing done to verify this PR. -->

## Test Result

<!-- Briefly summarize test outcomes. -->

## Submission Checklist

- [ ] Look over the contributing guidelines at
https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests.
---
 ...n_conv2d_fwd_perlayer_quantization_example.inc | 15 +++++++++++----
 1 file changed, 11 insertions(+), 4 deletions(-)

diff --git a/example/40_conv2d_fwd_quantization/run_conv2d_fwd_perlayer_quantization_example.inc b/example/40_conv2d_fwd_quantization/run_conv2d_fwd_perlayer_quantization_example.inc
index 02228d7654..26c3165446 100644
--- a/example/40_conv2d_fwd_quantization/run_conv2d_fwd_perlayer_quantization_example.inc
+++ b/example/40_conv2d_fwd_quantization/run_conv2d_fwd_perlayer_quantization_example.inc
@@ -108,28 +108,35 @@ bool run_grouped_conv_fwd(bool do_verification,
 
     if(do_verification)
     {
+        Tensor<AccDataType> c_host(out_g_n_k_wos_desc);
+
         auto ref_conv = ck::tensor_operation::host::ReferenceConvFwd<NDimSpatial,
                                                                      InDataType,
                                                                      WeiDataType,
-                                                                     OutDataType,
+                                                                     AccDataType,
                                                                      InElementOp,
                                                                      WeiElementOp,
-                                                                     OutElementOp>();
+                                                                     PassThrough>();
 
         auto ref_invoker  = ref_conv.MakeInvoker();
         auto ref_argument = ref_conv.MakeArgument(in,
                                                   wei,
-                                                  out_host,
+                                                  c_host,
                                                   conv_param.conv_filter_strides_,
                                                   conv_param.conv_filter_dilations_,
                                                   conv_param.input_left_pads_,
                                                   conv_param.input_right_pads_,
                                                   in_element_op,
                                                   wei_element_op,
-                                                  out_element_op);
+                                                  PassThrough{});
 
         ref_invoker.Run(ref_argument);
 
+        out_host.ForEach([&](auto&, auto idx)
+        {
+            out_element_op(out_host(idx), c_host(idx));
+        });
+
         out_device_buf.FromDevice(out_device.mData.data());
 
         pass &=