Update comment.

Stress the fact that the function adds not only the generic instance.

Update comment.
Stress the fact that the function adds not only the generic instance.
1354f25a · Adam Osewski · GitHub · 40fa7964 · 1354f25a
Unverified Commit 1354f25a authored Jul 07, 2023 by Adam Osewski Committed by GitHub Jul 07, 2023
Hide whitespace changes
Inline Side-by-side

Showing with 3 additions and 1 deletion

library/include/ck/library/tensor_operation_instance/gpu/grouped_convolution_forward.hpp ...or_operation_instance/gpu/grouped_convolution_forward.hpp +3 -1

No files found.
--- a/library/include/ck/library/tensor_operation_instance/gpu/grouped_convolution_forward.hpp
+++ b/library/include/ck/library/tensor_operation_instance/gpu/grouped_convolution_forward.hpp
@@ -259,7 +259,9 @@ void add_device_grouped_conv3d_fwd_xdl_ndhwgc_gkzyxc_ndhwgk_bf16_instances(
                                                              PassThrough,
                                                              PassThrough,
                                                              PassThrough>>>& instances);
-// Add a generic instance
+
+// This function adds a set of instance with the first being the most generic one. (That
+// is it supports the broadest convolution configuration spectrum)
 void add_device_grouped_conv3d_fwd_xdl_ndhwgc_gkzyxc_ndhwgk_f16_instances(
    std::vector<std::unique_ptr<DeviceGroupedConvFwdMultipleD<3,
                                                              NDHWGC,