cpu: aarch64: Expand ARM SVE support in jit_uni_softmax

[deepeshfujitsu]

Description

Enhancement: Expand ARM SVE support in jit_uni_softmax

This commit enhances the existing ARM SVE support in jit_uni_softmax to include additional vector length compatibility. The changes made are for implementation of different ARM SVE vector length.

Major Code changes:
• Introduce a new struct to accommodate the extended ARM SVE vector length, similar to the sve_512 struct.
• Modify the get_horizontal_op function accordingly for enhanced vector length.

Checklist

General

• [✓] Do all unit and benchdnn tests (make test and make test_benchdnn_*) pass locally for each commit? Yes
  Test output is same with and without this commit.

1. make test output:

   99% tests passed, 2 tests failed out of 160
   
    Total Test time (real) =  50.68 sec
    
    The following tests FAILED:
            144 - test_graph_unit_dnnl_conv_usm_cpu (Failed)
            149 - test_graph_unit_dnnl_large_partition_usm_cpu (Failed)
    Errors while running CTest
    Output from these tests are in: /home/deepesh/pull_request/oneDNN/build/Testing/Temporary/LastTest.log
    Use "--rerun-failed --output-on-failure" to re-run the failed cases verbosely.
    make: *** [Makefile:71: test] Error 8
   

2. make test_benchdnn_* output:

    make: *** No rule to make target 'test_benchdnn_*'.  Stop.
   

3. make test_benchdnn_softmax_ci_cpu/fast output:

    tests:6144 passed:1384 skipped:4736 mistrusted:24 unimplemented:0 invalid_arguments:0 failed:0 listed:0
    total: 13.70s; fill: 1.39s (10%); compute_ref: 0.47s (3%); compare: 0.87s (6%);
   

• [✓] Have you formatted the code using clang-format? Yes

cc: @kawakami-k , @davsva01

[cfRod]

Is there a reason for repeating the code for SVE256 - which is the same as SVE512 from Lines 389? Can there be just one jit_softmax_t and then there be entries for SVE512, 256 etc? (See #1445)

[deepeshfujitsu]

@cfRod, there is no specific reason for repeating the code for SVE256; it was done for faster development.

[cfRod]

Ok fair enough

[abhijain1204fujitsu]

Thanks @cfRod

[densamoilov]

@cfRod, do you have any comments from your side?

[cfRod]

Suggested for a refactor but happy to approve either way.

[abhijain1204fujitsu]

@densamoilov , Hello the changes are approved from @cfRod
Can you please confirm if this PR can be progressed for merger.

[dzarukin]

Thank you for the contribution!

[abhijain1204fujitsu]

Thank you @densamoilov @dzarukin for your quick response and merging the PR