3709 is a whopping 0.9% higher than 3675. That’s (probably) within the margin of error.Wow just wow looks like i got the highest ever recorded geekbench score on the quad. Keep in mind that my lcs has been built from ground up extensively still. Previously it was 3675 with the 6.5gb ram.
I did not read the document lol. Let me try the script. I did restart tho.@Jack Neill Silly question, but after running the system tuneup script (as instructed in the Welcome to Sorbet Leopard document), have you tried rebooting a second time afterward? One of the things the script does is repair the permissions, and maybe something (a kext, probably) wasn't playing ball due to being restored to disk instead of being traditionally installed. And in that case, a permissions repair should have solved it.
Otherwise, this will require more investigation ...
I have seen scores fluctuate by a few dozen points between consecutive runs.So in my opinion, a whopping 34 point difference probably would not qualify for being within the margin of error in that regard.
I went down a rabbit hole like that on my 2020 i5 Air and thermal modding and constantly trying to get GB5 numbers up. In the end my MB works exactly the same albeit a few degrees cooler.I have seen scores fluctuate by a few dozen points between consecutive runs.
In any case, 3709 is still less than one per cent higher than 3675.
IMO, people focus way too much on these “stupid” synthetic benchmarks anyway. To hell with them. What matters is if it’s actually faster in real-world use cases.
Did you run GB several times and average the results in both cases to rule out possible fluctuations? I don’t trust a single run of GB due to my experience.Before my 2nd FX 4500 i had the chance to test it with and without on gb. If i recall correctly in my case it affected 50-100 points something
What’s the point of having more than 16 GB then?Well it supports 32gb theoratically even 64 but memory controllers nor apps can use more than 16 gbs.