feat(gpu): speed up packing KS for levels==1

[andrei-stoian-zama]

Optimize packing ks for level count == 1

• Adds an optimized GEMM kernel
• Adds an alternative packing keyswitch implementation using the gemm kernel
• Implements a fast path for packing keyswitch for levels==1 using the gemm kernel
• Adds the fast path to the integer compression GPU code and adds a test with custom parameters that use the fast path

Achieves ~4x speedup for packing keyswitch

[agnesLeroy]

Hei @andrei-stoian-zama! It looks good to me overall, but I think it would be good that @pdroalves take a look since he was the one to implement the packing keyswitch on GPU.

[pdroalves]

Heyy @andrei-stoian-zama , thanks for this! You followed very closely our conventions and the code is clean and easy to read. Very impressive.

I added a few comments and suggestions to improve it, but nothing serious. Check to see wdyt. As long as Hyperstack allow us to see the CI green we can merge it.

About benchmarks, have you measured the impact for compression?

[pdroalves]

Looks good to me. It would be good to run the tests, though. I will try to rerun then before merge.

Thank you very much @andrei-stoian-zama !