fix build for mips loongarch risc-v with simplemath

[nihui]

No description provided.

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 95.52%. Comparing base (409f9ae) to head (535e9b7).
Report is 1 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #6024      +/-   ##
==========================================
+ Coverage   95.41%   95.52%   +0.10%     
==========================================
  Files         825      825              
  Lines      269192   269647     +455     
==========================================
+ Hits       256848   257567     +719     
+ Misses      12344    12080     -264     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[Copilot]

Pull Request Overview

This PR fixes build issues in the MIPS MSA simplemath module by replacing calls to the double-precision function "atan2" with the appropriate single-precision "atan2f".

• Replaced four calls to "atan2" with "atan2f" in the vectorized math function
• Ensures consistent float precision handling for processing v4f32 data

Comments suppressed due to low confidence (1)

src/layer/mips/msa_mathfun.h:263

• Replacing 'atan2' with 'atan2f' corrects the precision for float computations; please confirm that similar changes are not needed elsewhere in the module.

tmpx[0] = atan2f(tmpx[0], tmpy[0]);

[github-actions]

The binary size change of libncnn.so (bytes)

architecture	base size	pr size	difference
x86_64	16465128	16465128	0 😘
armhf	7335260	7335260	0 😘
aarch64	10704800	10704800	0 😘

[Copilot]

Pull Request Overview

This PR fixes build issues on MIPS, Loongarch, and RISC-V platforms by updating math function calls to their float-specific versions (e.g. replacing exp with expf and tanh with tanhf).

• Updated math functions in GRU implementations and binary/element-wise ops
• Consistently replaced standard math functions with their f variants across MIPS, Loongarch, and RISC-V files

Reviewed Changes

Copilot reviewed 31 out of 31 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
src/layer/riscv/gru_riscv_zfh.cpp	Replaced exp and tanh with expf and tanhf respectively
src/layer/riscv/gru_riscv.cpp	Updated math functions to use the f variants
src/layer/riscv/binaryop_riscv_zfh.cpp	Changed pow, atan2, etc. to use powf and atan2f
src/layer/riscv/binaryop_riscv.cpp	Similar changes with math functions for binary ops
src/layer/mips/unaryop_mips.cpp	Replaced fabs, floor, ceil, sqrt, exp, log, sin, etc. with f variants
src/layer/mips/tanh_mips.cpp	Updated tanh to tanhf
src/layer/mips/swish_mips.cpp	Updated exp to expf
src/layer/mips/softmax_mips.cpp	Updated exp to expf in softmax computation
src/layer/mips/sigmoid_mips.cpp	Updated sigmoid computation to use expf
src/layer/mips/msa_mathfun.h	Replaced atan2 calls with atan2f
src/layer/mips/mish_mips.cpp	Updated mish function calls to use tanhf and logf
src/layer/mips/binaryop_mips.cpp	Updated pow, atan2, and related functions to use f variants
src/layer/loongarch/unaryop_loongarch.cpp	Updated math functions to f variants
src/layer/loongarch/tanh_loongarch.cpp	Updated tanh to tanhf
src/layer/loongarch/swish_loongarch.cpp	Updated exp to expf
src/layer/loongarch/softmax_loongarch.cpp	Updated exp to expf
src/layer/loongarch/sigmoid_loongarch.cpp	Updated sigmoid computation to use expf
src/layer/loongarch/mish_loongarch.cpp	Updated mish function calls to use tanhf and logf
src/layer/loongarch/lsx_mathfun.h	Replaced atan2 with atan2f
src/layer/loongarch/binaryop_loongarch.cpp	Updated pow, atan2, and related functions to use f variants