Hi there,
I have various machines that have a copy of IM installed. Recently I installed version 6.8.6-6 using macports on a mac running 10.8.4 (mountain Lion). I also have a mac running the same version of IM on 10.6.8 (Snow Leopard). I ran a couple of tests on each machine and noticed that the newer mac running ML & IM ran slower than my older mac running SL & IM. On closer inspection I noticed than when a convert command was issued only 1 thread was being used on the mac running ML, where all 8 threads where being used on the mac running SL. I did a bit of searching here and found that OpenMP had not been enabled on the ML mac when being installed using macports. I then followed some advice here and manually installed the delegates and configured IM manually. After the install I could see that OpenMP had been installed and when I ran the tests again all 8 threads where being used but it was running even slower.
The tests times where:
SL 8 threads = 18 seconds
ML 1 thread = 20 seconds
ML 8 threads = 28 seconds
SL mac has a 2.2 GHz Intel Core i7 and 8 GB of ram
ML mac has a 2.6GHz Quad-core Intel i7 and 16 GB and a SSD
Could you please tell me why the convert process runs faster on 1 thread than on 8 in ML and why OpenMP does not get installed when configured with macports?
Many Thanks,
Nik