[CK][CK Tile] Grouped Convolution Backward Weight set of fixes (#5387)

## Motivation

Grouped Convolution Backward Weight split k fixes for CK tile kernels

## Technical Details

- get k batch from kargs to get deduced k batch
- multiply zeroing size by data type size
- disable v6 (producing a incorrect results)

## Test Plan

test_grouped_convnd_bwd_weight_tile

## Test Result

Pass

## Submission Checklist

- [x] Look over the contributing guidelines at
https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests.

---------

Co-authored-by: Ville Pietilä <>

example/ck_tile/20_grouped_convolution/grouped_convolution_backward_weight_invoker.hpp

@@ -126,7 +126,7 @@ struct GroupedConvolutionBackwardWeightInvoker
         }
 
         auto preprocess = [&]() {
-            if(args.k_batch > 1)
+            if(kargs.k_batch > 1)
             {
                 ck_tile::hip_check_error(hipMemsetAsync(
                     kargs.wei_ptr, 0, args.template GetWeightByte<WeiDataType>(), s.stream_id_));

example/ck_tile/20_grouped_convolution/grouped_convolution_backward_weight_two_stage_invoker.hpp

@@ -180,7 +180,7 @@ struct GroupedConvolutionBackwardWeightTwoStageInvoker
         }
 
         auto preprocess = [&]() {
-            if(args.k_batch > 1)
+            if(kargs.k_batch > 1)
                 ck_tile::hip_check_error(
                     hipMemsetAsync(ws_args.wei_ptr,
                                    0,