![EmuCR: N64 EmuCR: N64](https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhxblspv-rsAdnc0N-Wy-nQpgbj8_7Q8tJAn5Axpmmt4G-7xRW3ha0Lo8wKxfuWozHcEttguAvsV_oRy-st8lisZv2PN5k6sbC1m8D5cJg8n8Em2mBTZ7rkBFp6sIgKOKPhUU1BXSOKft0/s1600/images.jpg)
N64 Git Changelog:
* Merge pull request #39 from Dillonb/accurate_fpu
* Fix FPU on Windows
* Fix CVT overflow checks
* set_cause_cvt_l_d takes a double
* trunc.l, round.l, ceil.l, floor.l, cvt.l wip
* trunc.w, round.w, ceil.w, floor.w, cvt.w complete
* updates to trunc.w, round.w, ceil.w, floor.w, cvt.w. Not quite done yet
* cvt_w_s, cvt_w_d, remove last remaining NaN asserts
* cvt.s.fmt, cvt.d.fmt
* MFC1/DMFC1/MTC1/DMTC1 preserve cause
* actually, this is the behavior of all invalid FPU operations
* DCFC1/DCTC1 throw unimplemented exception
* all compare instructions
* exceptions and failure cases for mul/div/sqrt/abs/neg + fpu mov preserves cause
* handle FE_UNDERFLOW better
* macro for FPU ops, use for add.s/d, sub.s/d
* exceptions and failure cases for add.d
* exceptions and failure cases for add.s
* fix FPU exceptions - unimplemented operation should always be enabled
* implement ceil.l.d, ceil.w.d, floor.l.d, floor.w.d
* fix 64 bit floating point register accesses
* fix 32 bit floating point register accesses
* FPU accuracy updates
Download:
N64 Git (2023/03/19) : 1cloudfile bayfiles gofile pixeldrain send zippyshare
Source: Here
0 Comments
Post a Comment