- cmucl-cvs - mailman3.common-lisp.net

[Git][cmucl/cmucl][master] 2 commits: Fix #426: Define float-modes type correctly
by Raymond Toy (＠rtoy) 14 Aug '25

14 Aug '25

Raymond Toy pushed to branch master at cmucl / cmucl Commits: d2c96f1b by Raymond Toy at 2025-08-14T07:45:18-07:00 Fix #426: Define float-modes type correctly - - - - - 2db294b1 by Raymond Toy at 2025-08-14T07:45:18-07:00 Merge branch 'issue-426-fix-float-modes-type-def' into 'master' Fix #426: Define float-modes type correctly Closes #426 See merge request cmucl/cmucl!313 - - - - - 2 changed files: - src/compiler/x86/float-sse2.lisp - tests/float.lisp Changes: ===================================== src/compiler/x86/float-sse2.lisp ===================================== @@ -1265,7 +1265,7 @@ ;;;; Float mode hackery: -(deftype float-modes () '(unsigned-byte 24)) +(deftype float-modes () '(unsigned-byte 30)) ;; For the record, here is the format of the MXCSR register. ;; ===================================== tests/float.lisp ===================================== @@ -333,3 +333,13 @@ (declare (ignore c)) (values (invoke-restart 'lisp::largest-float))))) (read-from-string string)))))) + +#+x86 +(define-test x87-set-floating-point-modes + (:tag :issues) + (let ((new-mode (setf (x86::x87-floating-point-modes) + (dpb 2 (byte 2 24) + (x86::x87-floating-point-modes))))) + (assert-true (typep new-mode 'x86::float-modes)) + (assert-equal new-mode (setf (x86::x87-floating-point-modes) new-mode)))) + View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/compare/cb874f7fa2cd9d0dd38435… -- View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/compare/cb874f7fa2cd9d0dd38435… You're receiving this email because of your account on gitlab.common-lisp.net.

1 0

[Git][cmucl/cmucl][master] Update release notes for #424
by Raymond Toy (＠rtoy) 14 Aug '25

14 Aug '25

Raymond Toy pushed to branch master at cmucl / cmucl Commits: cb874f7f by Raymond Toy at 2025-08-14T07:42:48-07:00 Update release notes for #424 - - - - - 1 changed file: - src/general-info/release-21f.md Changes: ===================================== src/general-info/release-21f.md ===================================== @@ -137,6 +137,7 @@ public domain. * #417 PCL complains about repeated aux variables in defmethod * #418 Update asdf to version 3.3.7.4 (for package-local-nicknames) + * #424 Use fdlibm `hypot` to fix bug in snapshot 2025-07 * Other changes: * Improvements to the PCL implementation of CLOS: * Changes to building procedure: View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/commit/cb874f7fa2cd9d0dd38435c… -- View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/commit/cb874f7fa2cd9d0dd38435c… You're receiving this email because of your account on gitlab.common-lisp.net.

1 0

[Git][cmucl/cmucl][master] 2 commits: Fix #424: Use fdlibm hypot
by Raymond Toy (＠rtoy) 14 Aug '25

14 Aug '25

Raymond Toy pushed to branch master at cmucl / cmucl Commits: 4660c011 by Raymond Toy at 2025-08-14T07:40:17-07:00 Fix #424: Use fdlibm hypot - - - - - 6a2028c5 by Raymond Toy at 2025-08-14T07:40:17-07:00 Merge branch 'issue-424-fix-hypot' into 'master' Fix #424: Use fdlibm hypot Closes #424 See merge request cmucl/cmucl!311 - - - - - 4 changed files: - src/code/irrat.lisp - src/lisp/GNUmakefile - + src/lisp/e_hypot.c - tests/irrat.lisp Changes: ===================================== src/code/irrat.lisp ===================================== @@ -87,7 +87,7 @@ (def-math-rtn ("__ieee754_pow" %pow) 2) #-(or x86 sparc-v7 sparc-v8 sparc-v9) (def-math-rtn "sqrt" 1) -(def-math-rtn "hypot" 2) +(def-math-rtn ("__ieee754_hypot" %hypot) 2) (def-math-rtn ("fdlibm_log1p" %log1p) 1) (def-math-rtn ("fdlibm_expm1" %expm1) 1) ===================================== src/lisp/GNUmakefile ===================================== @@ -28,6 +28,7 @@ FDLIBM = k_sin.c k_cos.c k_tan.c s_sin.c s_cos.c s_tan.c sincos.c \ e_rem_pio2.c k_rem_pio2.c \ e_log10.c s_scalbn.c \ setexception.c \ + e_hypot.c \ log2.c SRCS = lisp.c coreparse.c alloc.c monitor.c print.c interr.c \ ===================================== src/lisp/e_hypot.c ===================================== @@ -0,0 +1,144 @@ + +/* @(#)e_hypot.c 1.3 95/01/18 */ +/* + * ==================================================== + * Copyright (C) 1993 by Sun Microsystems, Inc. All rights reserved. + * + * Developed at SunSoft, a Sun Microsystems, Inc. business. + * Permission to use, copy, modify, and distribute this + * software is freely granted, provided that this notice + * is preserved. + * ==================================================== + */ + +/* __ieee754_hypot(x,y) + * + * Method : + * If (assume round-to-nearest) z=x*x+y*y + * has error less than sqrt(2)/2 ulp, than + * sqrt(z) has error less than 1 ulp (exercise). + * + * So, compute sqrt(x*x+y*y) with some care as + * follows to get the error below 1 ulp: + * + * Assume x>y>0; + * (if possible, set rounding to round-to-nearest) + * 1. if x > 2y use + * x1*x1+(y*y+(x2*(x+x1))) for x*x+y*y + * where x1 = x with lower 32 bits cleared, x2 = x-x1; else + * 2. if x <= 2y use + * t1*y1+((x-y)*(x-y)+(t1*y2+t2*y)) + * where t1 = 2x with lower 32 bits cleared, t2 = 2x-t1, + * y1= y with lower 32 bits chopped, y2 = y-y1. + * + * NOTE: scaling may be necessary if some argument is too + * large or too tiny + * + * Special cases: + * hypot(x,y) is INF if x or y is +INF or -INF; else + * hypot(x,y) is NAN if x or y is NAN. + * + * Accuracy: + * hypot(x,y) returns sqrt(x^2+y^2) with error less + * than 1 ulps (units in the last place) + */ + +#include "fdlibm.h" + +static inline void +set_hiword(double* d, int hi) +{ + union { int i[2]; double d; } ux; + + ux.d = *d; + ux.i[HIWORD] = hi; + *d = ux.d; +} + +static inline int +get_loword(double d) +{ + union { int i[2]; double d; } ux; + ux.d = d; + return ux.i[LOWORD]; +} + +static inline int +get_hiword(double d) +{ + union { int i[2]; double d; } ux; + ux.d = d; + + return ux.i[HIWORD]; +} + + +#ifdef __STDC__ + double __ieee754_hypot(double x, double y) +#else + double __ieee754_hypot(x,y) + double x, y; +#endif +{ + double a=x,b=y,t1,t2,y1,y2,w; + int j,k,ha,hb; + + ha = get_hiword(x)&0x7fffffff; /* high word of x */ + hb = get_hiword(y)&0x7fffffff; /* high word of y */ + if(hb > ha) {a=y;b=x;j=ha; ha=hb;hb=j;} else {a=x;b=y;} + set_hiword(&a,ha); /* a <- |a| */ + set_hiword(&b, hb); /* b <- |b| */ + + if((ha-hb)>0x3c00000) {return a+b;} /* x/y > 2**60 */ + k=0; + if(ha > 0x5f300000) { /* a>2**500 */ + if(ha >= 0x7ff00000) { /* Inf or NaN */ + w = a+b; /* for sNaN */ + if(((ha&0xfffff)|get_loword(a))==0) w = a; + if(((hb^0x7ff00000)|get_loword(b))==0) w = b; + return w; + } + /* scale a and b by 2**-600 */ + ha -= 0x25800000; hb -= 0x25800000; k += 600; + set_hiword(&a, ha); + set_hiword(&b, hb); + } + if(hb < 0x20b00000) { /* b < 2**-500 */ + if(hb <= 0x000fffff) { /* subnormal b or 0 */ + if((hb|(get_loword(b)))==0) return a; + t1=0; + set_hiword(&t1, 0x7fd00000); /* t1=2^1022 */ + b *= t1; + a *= t1; + k -= 1022; + } else { /* scale a and b by 2^600 */ + ha += 0x25800000; /* a *= 2^600 */ + hb += 0x25800000; /* b *= 2^600 */ + k -= 600; + set_hiword(&a, ha); + set_hiword(&b, hb); + } + } + /* medium size a and b */ + w = a-b; + if (w>b) { + t1 = 0; + set_hiword(&t1, ha); + t2 = a-t1; + w = sqrt(t1*t1-(b*(-b)-t2*(a+t1))); + } else { + a = a+a; + y1 = 0; + set_hiword(&y1, hb); + y2 = b - y1; + t1 = 0; + set_hiword(&t1, ha+0x00100000); + t2 = a - t1; + w = sqrt(t1*y1-(w*(-w)-(t1*y2+t2*b))); + } + if(k!=0) { + t1 = 1.0; + set_hiword(&t1, get_hiword(t1) + (k<<20)); + return t1*w; + } else return w; +} ===================================== tests/irrat.lisp ===================================== @@ -245,3 +245,13 @@ (tanh #c(200w0 200w0)))) +;; See bug #424 +(define-test hypot + (:tag :issues) + (assert-eql 3.8950612975366328d0 + (kernel:%hypot 2.302585092994046d0 3.141592653589793d0)) + (let ((result (expt #C(2.302585092994046d0 3.141592653589793d0) 4))) + (assert-eql -188.4466069439329d0 + (realpart result)) + (assert-eql -132.16721026205448d0 + (imagpart result)))) View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/compare/9a8fd3c477f3bbfa1dec82… -- View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/compare/9a8fd3c477f3bbfa1dec82… You're receiving this email because of your account on gitlab.common-lisp.net.

1 0

[Git][cmucl/cmucl][issue-425-correctly-rounded-math-functions] Set 64-bit precision (80-bit float) in call-out
by Raymond Toy (＠rtoy) 13 Aug '25

13 Aug '25

Raymond Toy pushed to branch issue-425-correctly-rounded-math-functions at cmucl / cmucl Commits: 5ff177cb by Raymond Toy at 2025-08-13T15:06:30-07:00 Set 64-bit precision (80-bit float) in call-out Add some tests for the special functions where we were computing the incorrect value when the precision was not set to 64-bit (80-bit float). - - - - - 2 changed files: - src/compiler/x86/sse2-c-call.lisp - tests/irrat.lisp Changes: ===================================== src/compiler/x86/sse2-c-call.lisp ===================================== @@ -41,7 +41,7 @@ #+core-math (:temporary (:sc unsigned-stack) fpu-cw) #+core-math - (:temporary (:sc unsigned-reg) temp-cw) + (:temporary (:sc unsigned-reg :offset esi-offset) temp-cw) (:node-var node) (:vop-var vop) (:save-p t) @@ -51,11 +51,14 @@ #+core-math (progn ;; Save the x87 FPU control word. Then modify it to set the - ;; precision bits to double (2). + ;; precision bits to 3 for 64-bit mantissas for 80-bit + ;; arithmetic. If we don't some of some special functions + ;; return incorrect values because the x87 precision was set to + ;; single. (inst fnstcw save-fpu-cw) (move temp-cw save-fpu-cw) (inst and temp-cw (dpb 0 (byte 2 8) #xffff)) ; Zap the precision control bits - (inst or temp-cw (dpb 2 (byte 2 8) 0)) ; Set precision control to double + (inst or temp-cw (dpb 3 (byte 3 8) 0)) ; Set precision control bits (move fpu-cw temp-cw) (inst fldcw fpu-cw) ; New CW ) ===================================== tests/irrat.lisp ===================================== @@ -245,3 +245,18 @@ (tanh #c(200w0 200w0)))) +(define-test cos-tiny + (:tag issues) + ;; This test comes from the Maxima testsuite where core-math was not + ;; computing cos(8.881784197001252d-16) correctly. It should be + ;; exactly 1. We were getting one bit less. + (assert-eql 1d0 + (cos (kernel:make-double-float 1020264448 0)))) + +(define-test log.special-value + (:tag issues) + ;; This test comes from the Maxima testsuite where core-math was not + ;; computing log(0.6899991035461426d0) => -0.37106498060016496d0. + ;; The computed result looked it had like a single-precision accuracy. + (assert-eql -0.37106498060016496d0 + (kernel:%log (kernel:make-double-float (+ 1071644672 398457) 0)))) View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/commit/5ff177cb9f69fcd1042dbdf… -- View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/commit/5ff177cb9f69fcd1042dbdf… You're receiving this email because of your account on gitlab.common-lisp.net.

1 0

[Git][cmucl/cmucl][issue-425-correctly-rounded-math-functions] Linux cross-compile needs x86_linux_clang
by Raymond Toy (＠rtoy) 12 Aug '25

12 Aug '25

Raymond Toy pushed to branch issue-425-correctly-rounded-math-functions at cmucl / cmucl Commits: 0cca89a6 by Raymond Toy at 2025-08-12T15:52:01-07:00 Linux cross-compile needs x86_linux_clang Forgot to set the build config when running bin/build.sh. - - - - - 1 changed file: - .gitlab-ci.yml Changes: ===================================== .gitlab-ci.yml ===================================== @@ -182,10 +182,10 @@ linux:cross-build: script: # Use x86_linux_clang to build because x86_linux with gcc gives an # internal compiler error building core-math. - - bin/create-target.sh xtarget x86_linux_clang - - bin/create-target.sh xcross x86_linux_clang + - bin/create-target.sh xtarget "$build_config" + - bin/create-target.sh xcross "$build_config" - bin/cross-build-world.sh -crl xtarget xcross src/tools/cross-scripts/cross-x86-x86.lisp dist/bin/lisp - - bin/build.sh -b xlinux $bootstrap -R -C "" -o "xtarget/lisp/lisp -lib xtarget/lisp" + - bin/build.sh -b xlinux $bootstrap -R -C "$build_config" -o "xtarget/lisp/lisp -lib xtarget/lisp" - bin/make-dist.sh -I xdist xlinux-4 linux:test: View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/commit/0cca89a6ac031562f71b294… -- View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/commit/0cca89a6ac031562f71b294… You're receiving this email because of your account on gitlab.common-lisp.net.

1 0

[Git][cmucl/cmucl][issue-425-correctly-rounded-math-functions] Linux cross-compile needs x86_linux_clang
by Raymond Toy (＠rtoy) 12 Aug '25

12 Aug '25

Raymond Toy pushed to branch issue-425-correctly-rounded-math-functions at cmucl / cmucl Commits: eed071ef by Raymond Toy at 2025-08-12T07:19:06-07:00 Linux cross-compile needs x86_linux_clang As we did for the linux and opensuse builders, we need to set the build_config to Config.x86_linux_clang to build the core-math routines. - - - - - 1 changed file: - .gitlab-ci.yml Changes: ===================================== .gitlab-ci.yml ===================================== @@ -174,6 +174,11 @@ linux:cross-build: - job: linux:build #- job: linux:install artifacts: true + variables: + # The CI on linux needs to use clang to build the core-math + # routines. Gcc gets an internal compiler error for the OS + # running the linux builder. + build_config: "x86_linux_clang" script: # Use x86_linux_clang to build because x86_linux with gcc gives an # internal compiler error building core-math. View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/commit/eed071ef3e30185ea44794e… -- View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/commit/eed071ef3e30185ea44794e… You're receiving this email because of your account on gitlab.common-lisp.net.

1 0

[Git][cmucl/cmucl][issue-425-correctly-rounded-math-functions] 2 commits: Document build_config variable
by Raymond Toy (＠rtoy) 11 Aug '25

11 Aug '25

Raymond Toy pushed to branch issue-425-correctly-rounded-math-functions at cmucl / cmucl Commits: 1f94e40c by Raymond Toy at 2025-08-11T06:59:03-07:00 Document build_config variable - - - - - c9414bcb by Raymond Toy at 2025-08-11T06:59:28-07:00 Clean up implementation * Remove unneeded change in definition of %hypot in irrat.lisp. * Document what :core-math feature is for in linux-os.c. * Remove old commented stuff for creating libcore-math.a which we don't use. - - - - - 4 changed files: - .gitlab-ci.yml - src/code/irrat.lisp - src/code/linux-os.lisp - src/lisp/GNUmakefile Changes: ===================================== .gitlab-ci.yml ===================================== @@ -5,6 +5,8 @@ variables: version: "$year-$month-x86" tar_ext: "xz" bootstrap: "" + # Value to use for build.sh -C option. Allows choosing a different + # config, if needed. build_config: "" # Default install configuration to download the cmucl tarballs to use ===================================== src/code/irrat.lisp ===================================== @@ -121,7 +121,7 @@ (def-math-rtn ("__ieee754_pow" %pow) 2) #-(or x86 sparc-v7 sparc-v8 sparc-v9) (def-math-rtn "sqrt" 1) -(def-math-rtn ("hypot" %hypot) 2) +(def-math-rtn "hypot" 2) (def-math-rtn ("fdlibm_log1p" %log1p) 1) (def-math-rtn ("fdlibm_expm1" %expm1) 1) ===================================== src/code/linux-os.lisp ===================================== @@ -25,6 +25,8 @@ (register-lisp-feature :linux) (register-lisp-feature :elf) (register-lisp-runtime-feature :executable) +;; Core-math is used for the special functions providing +;; correctly-rounded numerical results. (register-lisp-runtime-feature :core-math) (setq *software-type* "Linux") ===================================== src/lisp/GNUmakefile ===================================== @@ -64,8 +64,6 @@ ifneq (${EXEC_FINAL_OBJ},) $(AR) d lisp.a exec-init.o endif -#libcore-math.a : $(CORE64_OBJ) -# $(AR) crs libcore-math.a $^ version: echo 0 > version View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/compare/ed62e414bc79477722f81a… -- View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/compare/ed62e414bc79477722f81a… You're receiving this email because of your account on gitlab.common-lisp.net.

1 0

[Git][cmucl/cmucl][issue-425-correctly-rounded-math-functions] Remove unused function and directories that aren't used by cmucl
by Raymond Toy (＠rtoy) 11 Aug '25

11 Aug '25

Raymond Toy pushed to branch issue-425-correctly-rounded-math-functions at cmucl / cmucl Commits: f8ee44be by Raymond Toy at 2025-08-10T18:01:21-07:00 Remove unused function and directories that aren't used by cmucl - - - - - 92 changed files: - − src/lisp/core-math/src/binary64/cbrt/Makefile - − src/lisp/core-math/src/binary64/cbrt/cbrt.c - − src/lisp/core-math/src/binary64/cbrt/cbrt.wc - − src/lisp/core-math/src/binary64/cbrt/cbrt_mpfr.c - − src/lisp/core-math/src/binary64/cbrt/check_special.c - − src/lisp/core-math/src/binary64/cbrt/function_under_test.h - − src/lisp/core-math/src/binary64/erf/Makefile - − src/lisp/core-math/src/binary64/erf/check_special.c - − src/lisp/core-math/src/binary64/erf/erf.c - − src/lisp/core-math/src/binary64/erf/erf.sage - − src/lisp/core-math/src/binary64/erf/erf.sollya - − src/lisp/core-math/src/binary64/erf/erf.wc - − src/lisp/core-math/src/binary64/erf/erf0.sollya - − src/lisp/core-math/src/binary64/erf/erf0_accurate.sollya - − src/lisp/core-math/src/binary64/erf/erf_mpfr.c - − src/lisp/core-math/src/binary64/erf/function_under_test.h - − src/lisp/core-math/src/binary64/erfc/Makefile - − src/lisp/core-math/src/binary64/erfc/asympt0.sollya - − src/lisp/core-math/src/binary64/erfc/asympt1.sollya - − src/lisp/core-math/src/binary64/erfc/asympt2.sollya - − src/lisp/core-math/src/binary64/erfc/asympt3.sollya - − src/lisp/core-math/src/binary64/erfc/asympt4.sollya - − src/lisp/core-math/src/binary64/erfc/asympt5.sollya - − src/lisp/core-math/src/binary64/erfc/asympt_acc0.sollya - − src/lisp/core-math/src/binary64/erfc/asympt_acc1.sollya - − src/lisp/core-math/src/binary64/erfc/asympt_acc2.sollya - − src/lisp/core-math/src/binary64/erfc/asympt_acc3.sollya - − src/lisp/core-math/src/binary64/erfc/asympt_acc4.sollya - − src/lisp/core-math/src/binary64/erfc/asympt_acc5.sollya - − src/lisp/core-math/src/binary64/erfc/asympt_acc6.sollya - − src/lisp/core-math/src/binary64/erfc/asympt_acc7.sollya - − src/lisp/core-math/src/binary64/erfc/asympt_acc8.sollya - − src/lisp/core-math/src/binary64/erfc/asympt_acc9.sollya - − src/lisp/core-math/src/binary64/erfc/check_special.c - − src/lisp/core-math/src/binary64/erfc/erfc.c - − src/lisp/core-math/src/binary64/erfc/erfc.wc - − src/lisp/core-math/src/binary64/erfc/erfc_mpfr.c - − src/lisp/core-math/src/binary64/erfc/exp_accurate.sollya - − src/lisp/core-math/src/binary64/erfc/function_under_test.h - − src/lisp/core-math/src/binary64/exp10/Makefile - − src/lisp/core-math/src/binary64/exp10/check_special.c - − src/lisp/core-math/src/binary64/exp10/exp10.c - − src/lisp/core-math/src/binary64/exp10/exp10.wc - − src/lisp/core-math/src/binary64/exp10/exp10_mpfr.c - − src/lisp/core-math/src/binary64/exp10/function_under_test.h - − src/lisp/core-math/src/binary64/exp10m1/Makefile - − src/lisp/core-math/src/binary64/exp10m1/check_special.c - − src/lisp/core-math/src/binary64/exp10m1/exp10m1.c - − src/lisp/core-math/src/binary64/exp10m1/exp10m1.wc - − src/lisp/core-math/src/binary64/exp10m1/exp10m1_mpfr.c - − src/lisp/core-math/src/binary64/exp10m1/function_under_test.h - − src/lisp/core-math/src/binary64/exp2/Makefile - − src/lisp/core-math/src/binary64/exp2/check_special.c - − src/lisp/core-math/src/binary64/exp2/exp2.c - − src/lisp/core-math/src/binary64/exp2/exp2.wc - − src/lisp/core-math/src/binary64/exp2/exp2_mpfr.c - − src/lisp/core-math/src/binary64/exp2/function_under_test.h - − src/lisp/core-math/src/binary64/lgamma/Makefile - − src/lisp/core-math/src/binary64/lgamma/check_special.c - − src/lisp/core-math/src/binary64/lgamma/function_under_test.h - − src/lisp/core-math/src/binary64/lgamma/lgamma.c - − src/lisp/core-math/src/binary64/lgamma/lgamma.wc - − src/lisp/core-math/src/binary64/lgamma/lgamma_mpfr.c - − src/lisp/core-math/src/binary64/log10p1/Makefile - − src/lisp/core-math/src/binary64/log10p1/check_special.c - − src/lisp/core-math/src/binary64/log10p1/dint.h - − src/lisp/core-math/src/binary64/log10p1/function_under_test.h - − src/lisp/core-math/src/binary64/log10p1/log10p1.c - − src/lisp/core-math/src/binary64/log10p1/log10p1.wc - − src/lisp/core-math/src/binary64/log10p1/log10p1_accurate.sollya - − src/lisp/core-math/src/binary64/log10p1/log10p1_mpfr.c - − src/lisp/core-math/src/binary64/log2p1/Makefile - − src/lisp/core-math/src/binary64/log2p1/check_special.c - − src/lisp/core-math/src/binary64/log2p1/dint_log2p1.h - − src/lisp/core-math/src/binary64/log2p1/function_under_test.h - − src/lisp/core-math/src/binary64/log2p1/log2p1.c - − src/lisp/core-math/src/binary64/log2p1/log2p1.sage - − src/lisp/core-math/src/binary64/log2p1/log2p1.wc - − src/lisp/core-math/src/binary64/log2p1/log2p1_accurate.sollya - − src/lisp/core-math/src/binary64/log2p1/log2p1_mpfr.c - − src/lisp/core-math/src/binary64/rsqrt/Makefile - − src/lisp/core-math/src/binary64/rsqrt/check_special.c - − src/lisp/core-math/src/binary64/rsqrt/function_under_test.h - − src/lisp/core-math/src/binary64/rsqrt/rsqrt.c - − src/lisp/core-math/src/binary64/rsqrt/rsqrt.wc - − src/lisp/core-math/src/binary64/rsqrt/rsqrt_mpfr.c - − src/lisp/core-math/src/binary64/tgamma/Makefile - − src/lisp/core-math/src/binary64/tgamma/check_special.c - − src/lisp/core-math/src/binary64/tgamma/function_under_test.h - − src/lisp/core-math/src/binary64/tgamma/tgamma.c - − src/lisp/core-math/src/binary64/tgamma/tgamma.wc - − src/lisp/core-math/src/binary64/tgamma/tgamma_mpfr.c The diff was not included because it is too large. View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/commit/f8ee44bea614ecb49eaa40e… -- View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/commit/f8ee44bea614ecb49eaa40e… You're receiving this email because of your account on gitlab.common-lisp.net.

1 0

[Git][cmucl/cmucl][issue-425-correctly-rounded-math-functions] 22 commits: Add single-float special functions from core-math
by Raymond Toy (＠rtoy) 11 Aug '25

11 Aug '25

Raymond Toy pushed to branch issue-425-correctly-rounded-math-functions at cmucl / cmucl Commits: aed4d5af by Raymond Toy at 2025-08-10T07:12:53-07:00 Add single-float special functions from core-math Add the single-float versions of the special functions from core-math that we use in cmucl. - - - - - e8e83578 by Raymond Toy at 2025-08-10T08:21:48-07:00 Remove unused core-math routines We don't support the sinpi routines and friends in cmucl. Perhaps later, but we don't do anything with these so reduce the clutter and remove these functions. - - - - - b581c732 by Raymond Toy at 2025-08-10T11:06:23-07:00 Revert "Add single-float special functions from core-math" This reverts commit aed4d5af2d766176139af05f8069aacc460b0f33. - - - - - 67bb5022 by Raymond Toy at 2025-08-10T11:10:07-07:00 Remove unused files for acos, acosh, asin, asinh We don't need these files so remove them. - - - - - ceec2a63 by Raymond Toy at 2025-08-10T11:12:30-07:00 Remove unused files for atan, atan2, atanh We don't need these files so remove them. - - - - - 9c80fb90 by Raymond Toy at 2025-08-10T11:14:14-07:00 Remove unused files for cos, cosh We don't need these files so remove them. - - - - - 12c15a04 by Raymond Toy at 2025-08-10T11:15:40-07:00 Remove unused files for exp, exp2m1 We don't need these files so remove them. - - - - - 8f7fa8d1 by Raymond Toy at 2025-08-10T11:19:20-07:00 Remove unused files for hypot We don't need these files for anything, so remove them. - - - - - ae933c79 by Raymond Toy at 2025-08-10T11:20:30-07:00 Remove unused files for log We don't need these files for anything, so remove them. - - - - - f88b8ea8 by Raymond Toy at 2025-08-10T11:21:11-07:00 Remove unused files for log10 We don't need these files for anything, so remove them. - - - - - 9cbe6f8c by Raymond Toy at 2025-08-10T11:21:45-07:00 Remove unused files for log1p We don't need these files for anything, so remove them. - - - - - f4328e70 by Raymond Toy at 2025-08-10T17:32:00-07:00 Remove core-math object files when cleaning Forgot to remove these object file when `make clean` is run. - - - - - 4a9b17cf by Raymond Toy at 2025-08-10T17:33:14-07:00 Restore these files that were accidentially removed These were accidentally removed when we were cleaning up unneeded files. - - - - - 7e209d1e by Raymond Toy at 2025-08-10T17:45:15-07:00 Remove unused files for sin We don't need these files for anything, so remove them. - - - - - b0972c5a by Raymond Toy at 2025-08-10T17:46:45-07:00 Remove more unneeded files for sin - - - - - 6a43a49b by Raymond Toy at 2025-08-10T17:48:01-07:00 Remove unused files for sinh We don't need these files for anything, so remove them. - - - - - 626f2d22 by Raymond Toy at 2025-08-10T17:48:34-07:00 Remove more unneeded files for sinh - - - - - 9b8a03ac by Raymond Toy at 2025-08-10T17:49:12-07:00 Remove unused files for tan We don't need these files for anything, so remove them. - - - - - 37a0ceea by Raymond Toy at 2025-08-10T17:50:20-07:00 Remove unused files for tanh We don't need these files for anything, so remove them. - - - - - 34e505d1 by Raymond Toy at 2025-08-10T17:50:45-07:00 Remove more unneeded files for tanh - - - - - 44338bdb by Raymond Toy at 2025-08-10T17:51:39-07:00 Remove unused files for sincos We don't need these files for anything, so remove them. - - - - - d486ac47 by Raymond Toy at 2025-08-10T17:52:05-07:00 Remove more unneeded files for sincos - - - - - 202 changed files: - src/lisp/GNUmakefile - − src/lisp/core-math/src/binary64/acos/Makefile - − src/lisp/core-math/src/binary64/acos/acos.wc - − src/lisp/core-math/src/binary64/acos/acos_mpfr.c - − src/lisp/core-math/src/binary64/acos/check_special.c - − src/lisp/core-math/src/binary64/acos/function_under_test.h - − src/lisp/core-math/src/binary64/acosh/Makefile - − src/lisp/core-math/src/binary64/acosh/acosh.wc - − src/lisp/core-math/src/binary64/acosh/acosh_mpfr.c - − src/lisp/core-math/src/binary64/acosh/check_special.c - − src/lisp/core-math/src/binary64/acosh/function_under_test.h - − src/lisp/core-math/src/binary64/acospi/Makefile - − src/lisp/core-math/src/binary64/acospi/acos.sollya - − src/lisp/core-math/src/binary64/acospi/acos2.sollya - − src/lisp/core-math/src/binary64/acospi/acos2_0.sollya - − src/lisp/core-math/src/binary64/acospi/acos3.sollya - − src/lisp/core-math/src/binary64/acospi/acos_sqrt.sollya - − src/lisp/core-math/src/binary64/acospi/acospi.c - − src/lisp/core-math/src/binary64/acospi/acospi.sage - − src/lisp/core-math/src/binary64/acospi/acospi.wc - − src/lisp/core-math/src/binary64/acospi/acospi_mpfr.c - − src/lisp/core-math/src/binary64/acospi/check_special.c - − src/lisp/core-math/src/binary64/acospi/doit.sh - − src/lisp/core-math/src/binary64/acospi/doit2.sh - − src/lisp/core-math/src/binary64/acospi/doit2_0.sh - − src/lisp/core-math/src/binary64/acospi/doit3.sh - − src/lisp/core-math/src/binary64/acospi/doit_sqrt.sh - − src/lisp/core-math/src/binary64/acospi/function_under_test.h - − src/lisp/core-math/src/binary64/asin/Makefile - − src/lisp/core-math/src/binary64/asin/asin.wc - − src/lisp/core-math/src/binary64/asin/asin_mpfr.c - − src/lisp/core-math/src/binary64/asin/check_special.c - − src/lisp/core-math/src/binary64/asin/function_under_test.h - − src/lisp/core-math/src/binary64/asinh/Makefile - − src/lisp/core-math/src/binary64/asinh/asinh.wc - − src/lisp/core-math/src/binary64/asinh/asinh_mpfr.c - − src/lisp/core-math/src/binary64/asinh/check_special.c - − src/lisp/core-math/src/binary64/asinh/function_under_test.h - − src/lisp/core-math/src/binary64/asinpi/Makefile - − src/lisp/core-math/src/binary64/asinpi/asinpi.c - − src/lisp/core-math/src/binary64/asinpi/asinpi.sage - − src/lisp/core-math/src/binary64/asinpi/asinpi.wc - − src/lisp/core-math/src/binary64/asinpi/asinpi_mpfr.c - − src/lisp/core-math/src/binary64/asinpi/check_special.c - − src/lisp/core-math/src/binary64/asinpi/function_under_test.h - − src/lisp/core-math/src/binary64/atan/Makefile - − src/lisp/core-math/src/binary64/atan/atan.wc - − src/lisp/core-math/src/binary64/atan/atan_mpfr.c - − src/lisp/core-math/src/binary64/atan/check_special.c - − src/lisp/core-math/src/binary64/atan/function_under_test.h - − src/lisp/core-math/src/binary64/atan2/Makefile - − src/lisp/core-math/src/binary64/atan2/atan2.sage - − src/lisp/core-math/src/binary64/atan2/atan2.sollya - − src/lisp/core-math/src/binary64/atan2/atan2.wc - − src/lisp/core-math/src/binary64/atan2/atan2_0.sollya - − src/lisp/core-math/src/binary64/atan2/atan2_mpfr.c - − src/lisp/core-math/src/binary64/atan2/atan2_small.sollya - − src/lisp/core-math/src/binary64/atan2/check_special.c - − src/lisp/core-math/src/binary64/atan2/doit.sh - − src/lisp/core-math/src/binary64/atan2/function_under_test.h - − src/lisp/core-math/src/binary64/atan2/worst.c - − src/lisp/core-math/src/binary64/atan2pi/Makefile - − src/lisp/core-math/src/binary64/atan2pi/atan2.sollya - − src/lisp/core-math/src/binary64/atan2pi/atan2pi.c - − src/lisp/core-math/src/binary64/atan2pi/atan2pi.sage - − src/lisp/core-math/src/binary64/atan2pi/atan2pi.wc - − src/lisp/core-math/src/binary64/atan2pi/atan2pi_mpfr.c - − src/lisp/core-math/src/binary64/atan2pi/check_special.c - − src/lisp/core-math/src/binary64/atan2pi/doit.sh - − src/lisp/core-math/src/binary64/atan2pi/function_under_test.h - − src/lisp/core-math/src/binary64/atan2pi/tint.h - − src/lisp/core-math/src/binary64/atan2pi/worst.c - − src/lisp/core-math/src/binary64/atanh/Makefile - − src/lisp/core-math/src/binary64/atanh/atanh.wc - − src/lisp/core-math/src/binary64/atanh/atanh_mpfr.c - − src/lisp/core-math/src/binary64/atanh/check_special.c - − src/lisp/core-math/src/binary64/atanh/function_under_test.h - − src/lisp/core-math/src/binary64/atanpi/Makefile - − src/lisp/core-math/src/binary64/atanpi/atanpi.c - − src/lisp/core-math/src/binary64/atanpi/atanpi.sage - − src/lisp/core-math/src/binary64/atanpi/atanpi.wc - − src/lisp/core-math/src/binary64/atanpi/atanpi_mpfr.c - − src/lisp/core-math/src/binary64/atanpi/check_special.c - − src/lisp/core-math/src/binary64/atanpi/function_under_test.h - − src/lisp/core-math/src/binary64/cos/Makefile - − src/lisp/core-math/src/binary64/cos/check_special.c - − src/lisp/core-math/src/binary64/cos/cos.wc - − src/lisp/core-math/src/binary64/cos/cos_mpfr.c - − src/lisp/core-math/src/binary64/cos/function_under_test.h - − src/lisp/core-math/src/binary64/cosh/Makefile - − src/lisp/core-math/src/binary64/cosh/check_special.c - − src/lisp/core-math/src/binary64/cosh/cosh.wc - − src/lisp/core-math/src/binary64/cosh/cosh_mpfr.c - − src/lisp/core-math/src/binary64/cosh/function_under_test.h - − src/lisp/core-math/src/binary64/cospi/Makefile - − src/lisp/core-math/src/binary64/cospi/check_special.c - − src/lisp/core-math/src/binary64/cospi/cospi.c - − src/lisp/core-math/src/binary64/cospi/cospi.wc - − src/lisp/core-math/src/binary64/cospi/cospi_mpfr.c - − src/lisp/core-math/src/binary64/cospi/function_under_test.h - − src/lisp/core-math/src/binary64/exp/Makefile - − src/lisp/core-math/src/binary64/exp/check_special.c - − src/lisp/core-math/src/binary64/exp/exp.sage - − src/lisp/core-math/src/binary64/exp/exp.wc - − src/lisp/core-math/src/binary64/exp/exp_mpfr.c - − src/lisp/core-math/src/binary64/exp/function_under_test.h - − src/lisp/core-math/src/binary64/exp2m1/Makefile - − src/lisp/core-math/src/binary64/exp2m1/check_special.c - − src/lisp/core-math/src/binary64/exp2m1/exp2m1.c - − src/lisp/core-math/src/binary64/exp2m1/exp2m1.sage - − src/lisp/core-math/src/binary64/exp2m1/exp2m1.wc - − src/lisp/core-math/src/binary64/exp2m1/exp2m1_accurate.sollya - − src/lisp/core-math/src/binary64/exp2m1/exp2m1_fast.sollya - − src/lisp/core-math/src/binary64/exp2m1/exp2m1_mpfr.c - − src/lisp/core-math/src/binary64/exp2m1/function_under_test.h - − src/lisp/core-math/src/binary64/hypot/Makefile - − src/lisp/core-math/src/binary64/hypot/check_special.c - − src/lisp/core-math/src/binary64/hypot/function_under_test.h - − src/lisp/core-math/src/binary64/hypot/hypot.wc - − src/lisp/core-math/src/binary64/hypot/hypot_mpfr.c - − src/lisp/core-math/src/binary64/hypot/triples.c - − src/lisp/core-math/src/binary64/log/Makefile - − src/lisp/core-math/src/binary64/log/check_special.c - − src/lisp/core-math/src/binary64/log/doit.sh - − src/lisp/core-math/src/binary64/log/function_under_test.h - − src/lisp/core-math/src/binary64/log/gappa.sage - − src/lisp/core-math/src/binary64/log/log.sage - − src/lisp/core-math/src/binary64/log/log.wc - − src/lisp/core-math/src/binary64/log/log1_template.g - − src/lisp/core-math/src/binary64/log/log_mpfr.c - − src/lisp/core-math/src/binary64/log10/Makefile - − src/lisp/core-math/src/binary64/log10/P.sollya - − src/lisp/core-math/src/binary64/log10/check_special.c - − src/lisp/core-math/src/binary64/log10/function_under_test.h - − src/lisp/core-math/src/binary64/log10/log10.sage - − src/lisp/core-math/src/binary64/log10/log10.wc - − src/lisp/core-math/src/binary64/log10/log10_mpfr.c - − src/lisp/core-math/src/binary64/log1p/Makefile - − src/lisp/core-math/src/binary64/log1p/P.sollya - − src/lisp/core-math/src/binary64/log1p/check_special.c - − src/lisp/core-math/src/binary64/log1p/dint.h - − src/lisp/core-math/src/binary64/log1p/function_under_test.h - − src/lisp/core-math/src/binary64/log1p/log1p.sage - − src/lisp/core-math/src/binary64/log1p/log1p.wc - − src/lisp/core-math/src/binary64/log1p/log1p_mpfr.c - − src/lisp/core-math/src/binary64/sin/Makefile - − src/lisp/core-math/src/binary64/sin/buildSC.c - − src/lisp/core-math/src/binary64/sin/check_special.c - − src/lisp/core-math/src/binary64/sin/cos_accurate.sollya - − src/lisp/core-math/src/binary64/sin/cos_fast.sollya - − src/lisp/core-math/src/binary64/sin/function_under_test.h - − src/lisp/core-math/src/binary64/sin/sin.sage - − src/lisp/core-math/src/binary64/sin/sin.wc - − src/lisp/core-math/src/binary64/sin/sin_accurate.sollya - − src/lisp/core-math/src/binary64/sin/sin_fast.sollya - − src/lisp/core-math/src/binary64/sin/sin_mpfr.c - − src/lisp/core-math/src/binary64/sincos/Makefile - − src/lisp/core-math/src/binary64/sincos/check_special.c - − src/lisp/core-math/src/binary64/sincos/function_under_test.h - − src/lisp/core-math/src/binary64/sincos/sincos.wc - − src/lisp/core-math/src/binary64/sincos/sincos_mpfr.c - − src/lisp/core-math/src/binary64/sinh/Makefile - − src/lisp/core-math/src/binary64/sinh/Pcosh.sollya - − src/lisp/core-math/src/binary64/sinh/Pcosh2.sollya - − src/lisp/core-math/src/binary64/sinh/Psinh.sollya - − src/lisp/core-math/src/binary64/sinh/Psinh2.sollya - − src/lisp/core-math/src/binary64/sinh/buildu.c - − src/lisp/core-math/src/binary64/sinh/check_special.c - − src/lisp/core-math/src/binary64/sinh/function_under_test.h - − src/lisp/core-math/src/binary64/sinh/sinh.sage - − src/lisp/core-math/src/binary64/sinh/sinh.wc - − src/lisp/core-math/src/binary64/sinh/sinh_mpfr.c - − src/lisp/core-math/src/binary64/sinpi/Makefile - − src/lisp/core-math/src/binary64/sinpi/check_special.c - − src/lisp/core-math/src/binary64/sinpi/function_under_test.h - − src/lisp/core-math/src/binary64/sinpi/sinpi.c - − src/lisp/core-math/src/binary64/sinpi/sinpi.wc - − src/lisp/core-math/src/binary64/sinpi/sinpi_mpfr.c - − src/lisp/core-math/src/binary64/support/Makefile.bivariate - − src/lisp/core-math/src/binary64/support/Makefile.sincos - − src/lisp/core-math/src/binary64/support/Makefile.univariate - − src/lisp/core-math/src/binary64/support/check_worst.c - − src/lisp/core-math/src/binary64/support/check_worst_sincos.c - − src/lisp/core-math/src/binary64/support/check_worst_uni.c - − src/lisp/core-math/src/binary64/support/fenv_mpfr.h - − src/lisp/core-math/src/binary64/tan/Makefile - − src/lisp/core-math/src/binary64/tan/check_special.c - − src/lisp/core-math/src/binary64/tan/function_under_test.h - − src/lisp/core-math/src/binary64/tan/tan.sage - − src/lisp/core-math/src/binary64/tan/tan_mpfr.c - − src/lisp/core-math/src/binary64/tanh/Makefile - − src/lisp/core-math/src/binary64/tanh/check_special.c - − src/lisp/core-math/src/binary64/tanh/function_under_test.h - − src/lisp/core-math/src/binary64/tanh/tanh.sage - − src/lisp/core-math/src/binary64/tanh/tanh.wc - − src/lisp/core-math/src/binary64/tanh/tanh_mpfr.c - − src/lisp/core-math/src/binary64/tanpi/Makefile - − src/lisp/core-math/src/binary64/tanpi/check_special.c - − src/lisp/core-math/src/binary64/tanpi/function_under_test.h - − src/lisp/core-math/src/binary64/tanpi/tanpi.c - − src/lisp/core-math/src/binary64/tanpi/tanpi.wc - − src/lisp/core-math/src/binary64/tanpi/tanpi_mpfr.c The diff was not included because it is too large. View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/compare/431e06e7dc04a3735e36a3… -- View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/compare/431e06e7dc04a3735e36a3… You're receiving this email because of your account on gitlab.common-lisp.net.

1 0

[Git][cmucl/cmucl][issue-425-correctly-rounded-math-functions] Oops. Need to use x86_linux_clang for linux cross-compile
by Raymond Toy (＠rtoy) 10 Aug '25

10 Aug '25

Raymond Toy pushed to branch issue-425-correctly-rounded-math-functions at cmucl / cmucl Commits: 431e06e7 by Raymond Toy at 2025-08-09T21:58:19-07:00 Oops. Need to use x86_linux_clang for linux cross-compile - - - - - 1 changed file: - .gitlab-ci.yml Changes: ===================================== .gitlab-ci.yml ===================================== @@ -173,8 +173,10 @@ linux:cross-build: #- job: linux:install artifacts: true script: - - bin/create-target.sh xtarget - - bin/create-target.sh xcross + # Use x86_linux_clang to build because x86_linux with gcc gives an + # internal compiler error building core-math. + - bin/create-target.sh xtarget x86_linux_clang + - bin/create-target.sh xcross x86_linux_clang - bin/cross-build-world.sh -crl xtarget xcross src/tools/cross-scripts/cross-x86-x86.lisp dist/bin/lisp - bin/build.sh -b xlinux $bootstrap -R -C "" -o "xtarget/lisp/lisp -lib xtarget/lisp" - bin/make-dist.sh -I xdist xlinux-4 View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/commit/431e06e7dc04a3735e36a38… -- View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/commit/431e06e7dc04a3735e36a38… You're receiving this email because of your account on gitlab.common-lisp.net.

1 0