Apply blocked format for deconv

[byungilm]

Description of the issue

• iGPU execution showed worse performance than CPU
  benchmark_app -m irsmodel.onnx -d GPU -ip f16 Result is 160 FPS
benchmark_app -m irsmodel.onnx -d CPU -ip f16 Result is 250 FPS
• Root cause is that deconvolution with simple format execution showed worse performance.
• cpu :
• igpu :
  

The code and line that caused this issue

• Convolution layers are only counted while selecting blocked formats even though Deconvolution shows lower performance with simple format :
  program.cpp:1622

Reproduction step and snapshot

• Reproduced by benchmark
  .\benchmark_app.exe -m models\irsmodel\irsmodel.onnx -d GPU.0 -t 10 -b 1

Checklist

• Is it a proper fix?
• Did you include test case for this fix, if necessary?
• Did you review existing test that can be extended to cover this scenario? Passed llm_bench

Tickets:

• CVS-170498