[Aryansh1520]

What

PPSA28997 (God of War Sons of Sparta) imports these four functions, and none of them is exported yet.

• logb, sinhf and nearbyintf forward to the C++ standard library, like their existing neighbours (logbf, sinf).
• __powisf2, the compiler-rt helper for powi on floats, uses compiler-rt's algorithm: repeated squaring in float, then the reciprocal for a negative exponent.

Tested

• Linux (CachyOS, kernel 7.2), GCC: PPSA28997 (God of War Sons of Sparta) runs past the calls.
• tools/check_conventions.py --base origin/main and tools/pr_overlap.py --branch pass.

Checklist

• Based on current main; no other open PR implements the same functions
• One topic per PR; follow-ups go in a new PR
• No comments in code except technical debt
• Unimplemented paths throw (NotImplemented_nid_no_patch); silent stubs are listed in TechnicalDebt
• No notes, investigation .md files or images added to the repository (attach them to this PR)
• New third-party code is a submodule built from source
• Generic behaviour, not specific to one title
• Depends on: none
• AI-assisted: yes

[MarcZX10]

Please add focused tests to guest_math for logb, sinhf, nearbyintf and powisf2, including exponents 0, -1 and INT_MIN, plus special values. The for-loop increment squares base after remaining has become zero; exponent 1 therefore computes an unused FLT_MAX * FLT_MAX, which can raise overflow. LLVM compiler-rt stops before that final square (reference: https://github.com/llvm/llvm-project/blob/main/compiler-rt/lib/builtins/powisf2.c). I built exact head 23c84b2 on Linux with GCC 12; the existing guest_math passed 1/1 in 0.02 s and conventions passed, but that test does not cover these new exports. No GitHub checks are reported for this head.

[oneandonlydean]

Which of these does PPSA28997 stop at on main? Please add that log line; with it and the tests MarcZX100 asked for, this can go in.

[Aryansh1520]

Updated in a046106.

• __powisf2 now follows compiler-rt exactly: it halves the exponent and stops before squaring once it reaches zero, so exponent 1 no longer computes the unused FLT_MAX * FLT_MAX. With GCC 12+ at -O2 the old dead square was already optimized away (the new overflow-flag check passed on both versions here), but the code no longer depends on that.
• guest_math now covers logb (powers of two, negatives, denormal min, max, ±0, -inf, NaN), sinhf (±0 sign, known values, overflow to ±inf, ±inf, NaN), nearbyintf (ties to even, -0 sign, large values, no FE_INEXACT, upward and toward-zero rounding modes) and __powisf2 (exponents 0, 1, -1, 10, 127, 128, -126, INT_MAX, INT_MIN; zero, infinity and NaN bases; no FE_OVERFLOW for FLT_MAX^1 and FLT_MAX^-1).

On where PPSA28997 stops: I don't have a log line for these four. I found them statically, among the unresolved imports of the title's modules (logb from Il2cppUserAssemblies.prx, nearbyintf from eboot.elf, sinhf and __powisf2 in the libc import table), and implemented them before any run reached them. So no recorded run stopped at one of them, and I won't claim a line I didn't see. On main (as of Oct 8) the title stops earlier, at LTov9gMEqCU (std::ios_base::_Addstd, covered by #545).

The build workflows on this head are waiting for approval to run (fork PR).

[Aryansh1520]

To avoid the GuestMath.cpp overlap with #1818 and #2038, the new checks now live in their own test, GuestFloatMath.cpp (guest_float_math), in 7550255. The checks are unchanged, and this PR no longer edits GuestMath.cpp.

[popcoder565]

Native Windows run, since no Windows results were posted. Setup: Windows 10 (19045), MinGW-w64 GCC 15.2.0 (WinLibs r7), Release. Head 7550255d, merged onto main d70b899 in one build with 38 other PRs that had no Windows run.

• This PR's tests pass: guest_float_math. None of them reported skipping a case.
• Full ctest in that build: 519 of 519. That needed local fixes to three other PRs in the batch that don't build or pass on Windows: feat(libc): export scriptingGetMem and scriptingFreeMem from application heap #1611, feat(posix): implement gethostbyname #1796 and feat(zlib): implement sequential libSceZlib decompression #1824 (reported there).

No title was run.

🤖 Addressed by Claude Code

[oneandonlydean]

Closing: on main PPSA28997 stops at LTov9gMEqCU (#545) before it reaches any of these, and library functions go in when a title stops on them. Please reopen with the log line once one of them is the stop.