I do however get worse benchmark scores across the board, cinebench r20 multicore as example gets 5950-6000 rather than 6100+ with no offset and -30curve optimizer, and temps go up to 83-84.
When I dial back the curve optimizer to accommodate having a +mhz offset, I don't crash even with all cores at -27 offset in curve optimizer. Now, when I start to add any +50mhz to +200mhz offset, I crash with curve optimizer -30 all core, as expected (+25mhz worked, no difference in reality though when testing). These single core cinebench r20 scores are a little worse than the 4.7ghz OC also, like 618 vs 637 and multicore is like 6105 vs 6169. HW info shows the cores hitting 4850 at max and effective core hovers around 4750 on every core. This has given me the best benchmarking scores over any other pbo/curve settings and benchmarks are just a little lower than the all-core 4.7ghz OC I already found stable. XMP is enabled & fclk is 1800 to match my 3600mhz ram. Setting PBO to auto (no 0 +mhz offset) and putting curve optimizer to -30 all core. Now, I've been messing with PBO + Curve optimizer for 2 days to take advantage of boosting a few cores higher since I dont really use all 8 cores too often while working, and I am getting worse results in all benchmarks (cinebench, superposition, 3d mark benchmarks) no matter what I do, except one thing.
After stress testing 6hrs w/ aida64, temps only hit 82 max and usually hovered at 80. I got pretty good results with an all-core overclock to 4.7 1.28v, stable and also good benchmark scores.
REASONS NOT TO USE A RAM OPTIMIZER PC
So I just got my 5800x about a week ago, after switching from a 3900x that no longer is useful 99% of my time (changed jobs), and have been messing with fine tuning voltage/heat to performance for my pc that runs about 12hrs a day.
0 kommentar(er)
