[Git][cmucl/cmucl][issue-425-correctly-rounded-math-functions] Experiment to see if -march=haswell works better on CI.
Raymond Toy pushed to branch issue-425-correctly-rounded-math-functions at cmucl / cmucl Commits: c61e55e6 by Raymond Toy at 2026-01-29T18:31:13-08:00 Experiment to see if -march=haswell works better on CI. - - - - - 2 changed files: - src/lisp/Config.x86_linux - src/lisp/Config.x86_linux_clang Changes: ===================================== src/lisp/Config.x86_linux ===================================== @@ -9,7 +9,8 @@ CPPFLAGS += -m32 # Gcc 15.2.1 and possibly other version can't compile core-math # without this option. CFLAGS += -std=gnu11 -CFLAGS += -rdynamic -march=pentium4 -mfpmath=sse -mtune=generic +#CFLAGS += -rdynamic -march=pentium4 -mfpmath=sse -mtune=generic +CFLAGS += -rdynamic -march=haswell -mfpmath=sse # ANALYZER flag should be set to -fanalyzer when doing the CI builds # for the C static analyzer. CFLAGS += -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 $(ANALYZER) ===================================== src/lisp/Config.x86_linux_clang ===================================== @@ -8,7 +8,8 @@ CFLAGS += $(COPT) # Allow sse2 instructions (-msse2); and tune for pentium4 # (-mtune=pentium4), the first chip to have sse2; and finally generate # code assuming instructions can trap (-ftrapping-math). -CFLAGS += -msse2 -mtune=pentium4 -ftrapping-math +#CFLAGS += -msse2 -mtune=pentium4 -ftrapping-math +CFLAGS += -msse2 -mtune=haswell -ftrapping-math CFLAGS += -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 UNDEFSYMPATTERN = -Xlinker -u -Xlinker & View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/commit/c61e55e6984de4a9c3e380e3... -- View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/commit/c61e55e6984de4a9c3e380e3... You're receiving this email because of your account on gitlab.common-lisp.net.
participants (1)
-
Raymond Toy (@rtoy)