S�bastien Villemot dixit: >In order to fix this, we need to pre-compute so-called "architectural >defaults" for m68k. These are pre-computed timings that will speed up >dramatically the build of the package.
I wonder: *does* m68k have something like an L1 cache, at all? […] make[6]: Entering directory `/root/Xa/atlas-3.10.1/build/atlas-base/tune/sysinfo' /usr/bin/c99-gcc -c -DL2SIZE=4194304 -I/root/Xa/atlas-3.10.1/build/atlas-base/include -I/root/Xa/atlas-3.10.1/build/atlas-base/../..//include -I/root/Xa/atlas-3.10.1/build/atlas-base/../..//include/contrib -DAdd_ -DF77_INTEGER=int -DStringSunStyle -DATL_OS_Linux -DWALL -O -fomit-frame-pointer -fPIC -I./ /root/Xa/atlas-3.10.1/build/atlas-base/../..//tune/sysinfo/time.c /usr/bin/c99-gcc -c -DL2SIZE=4194304 -I/root/Xa/atlas-3.10.1/build/atlas-base/include -I/root/Xa/atlas-3.10.1/build/atlas-base/../..//include -I/root/Xa/atlas-3.10.1/build/atlas-base/../..//include/contrib -DAdd_ -DF77_INTEGER=int -DStringSunStyle -DATL_OS_Linux -DWALL -O -fomit-frame-pointer -fPIC /root/Xa/atlas-3.10.1/build/atlas-base/../..//tune/sysinfo/L1CacheSize.c /usr/bin/c99-gcc -DL2SIZE=4194304 -I/root/Xa/atlas-3.10.1/build/atlas-base/include -I/root/Xa/atlas-3.10.1/build/atlas-base/../..//include -I/root/Xa/atlas-3.10.1/build/atlas-base/../..//include/contrib -DAdd_ -DF77_INTEGER=int -DStringSunStyle -DATL_OS_Linux -DWALL -O -fomit-frame-pointer -fPIC -o xL1 L1CacheSize.o time.o /root/Xa/atlas-3.10.1/build/atlas-base/bin/ATLrun.sh /root/Xa/atlas-3.10.1/build/atlas-base/tune/sysinfo xL1 64 Calculating L1 cache size: L1CS=2, time=54.440000 (ignore=0.0e+00) L1CS=4, time=53.830000 (ignore=0.0e+00) L1CS=8, time=55.200000 (ignore=0.0e+00) L1CS=16, time=53.350000 (ignore=0.0e+00) L1CS=32, time=54.170000 (ignore=0.0e+00) L1CS=64, time=46.490000 (ignore=0.0e+00) Timing differences below tolerance, trying larger maximum cache size: L1CS=2, time=212.600000 (ignore=0.0e+00) L1CS=4, time=211.650000 (ignore=0.0e+00) L1CS=8, time=200.980000 (ignore=0.0e+00) L1CS=16, time=216.530000 (ignore=0.0e+00) L1CS=32, time=217.060000 (ignore=0.0e+00) L1CS=64, time=214.770000 (ignore=0.0e+00) L1CS=128, time=218.480000 (ignore=0.0e+00) L1CS=256, time=219.740000 (ignore=0.0e+00) Confirming result of 8kb: L1CS=2, time=210.400000 (ignore=0.0e+00) L1CS=4, time=215.380000 (ignore=0.0e+00) L1CS=8, time=206.550000 (ignore=0.0e+00) L1CS=16, time=213.900000 (ignore=0.0e+00) L1CS=32, time=204.110000 (ignore=0.0e+00) L1CS=64, time=211.100000 (ignore=0.0e+00) L1CS=128, time=208.560000 (ignore=0.0e+00) L1CS=256, time=214.300000 (ignore=0.0e+00) L1CS=2, time=209.300000 (ignore=0.0e+00) L1CS=4, time=216.740000 (ignore=0.0e+00) L1CS=8, time=211.110000 (ignore=0.0e+00) L1CS=16, time=203.680000 (ignore=0.0e+00) L1CS=32, time=221.810000 (ignore=0.0e+00) L1CS=64, time=220.010000 (ignore=0.0e+00) L1CS=128, time=214.670000 (ignore=0.0e+00) L1CS=256, time=222.550000 (ignore=0.0e+00) Calculated L1 cache size = 8kb; Correct=0 Calculating L1 cache size: L1CS=2, time=56.780000 (ignore=0.0e+00) […] bye, //mirabilos -- (gnutls can also be used, but if you are compiling lynx for your own use, there is no reason to consider using that package) -- Thomas E. Dickey on the Lynx mailing list, about OpenSSL -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org