I wonder wich offset for this processor is normal. Had to go til -30 but in some situations it was unstable.
Solved! Go to Solution.
To be honest, I'm surprised that you got -25 CO all-core stable on a 5950X. Have you validated it with an all-core workload like Cinebench?
Most 6C processors will do -30 no problem. However 8C often only do -25 stable. The 12C usually need -20 or less and 16C can only do -15 all-core.
Those numbers are based on typical chips and voltages can be tweaked on a per-core basis but that involves a great deal of trial and error.
Offset was implemented for all cores.
To be honest, I'm surprised that you got -25 CO all-core stable on a 5950X. Have you validated it with an all-core workload like Cinebench?
Most 6C processors will do -30 no problem. However 8C often only do -25 stable. The 12C usually need -20 or less and 16C can only do -15 all-core.
Those numbers are based on typical chips and voltages can be tweaked on a per-core basis but that involves a great deal of trial and error.
That is why i m curious too.
Today asked Ryzen Master to make auto overclocking to find optimum offset for all cores, it found that it is -30.
Than I validated it throu running cinebench24. It is stable in this test. Results are on screenshots. One taken during test one after.
Another quastion - is there any possibility to reach higher frequences. As i see it is very unique Ryzen 9 5950x. It is a little pity, that it was found only now.
I can almost guarantee that system will not be stable at a all core -30. Just running Cinebench 24 does not validate it as stable. You will notice the greatest unstability at low load. If you have a blu ray drive, play a blue ray movie, that system will be restarting constantly. I have a couple cores at -25 and work all the way up to having 1 core at -4. Those Cinebench runs don't even look valide as the picture is not even on screen under the Ryzen Master screen.
No i dont, but i tryed to view some 2k movies and it is ok. Actually esterday I played WOW with offset -26 for all cores and it was very fine. But i totally agree that only cinebench test doesnt guarantee 100% stability.
By the way have made a stream with cinebench running and offset set on -30.
Result is a little less but i think it is because of obs.
https://www.youtube.com/live/8kzGXFzcEl4?si=3UO-ysjZ8TgW31oa