Ive read some people on this sub saying that if one is always bottlenecked by gpu, a ram oc, even a good one, will be useless and show real time performance comparable to stock expo/xmp. Is this actually true? Im fairly new to overclocking, and last month i upgraded to a 64gb ddr5 6000mt/cl28 trident neo kit to try out the mem oc experience. The thing is, i have a 9800x3d and a 5070ti which means my cpu usage is almost half my gpu usage percentage in the games i play. I had a 32gb kingston fury 6000mt/cl30 kit that actually had lower latency in aida and fps was more or less the same as my current dimms while measuring boths expo profiles. I dont get it. Should i factor in the reduced capacity? And should i actually just stay away from ocing my ram at all due to my setup?
Have i literally wasted money or is there some hope?
6 Comments
That’s why they say ram oc generally isn’t worth your time. Hours of tinkering for 0.5% gains.
But it’s fun finding the limits of your hardware if you have the time.
There’s something very wrong with your write speeds there I think. Looks half of what it should be, maybe even less.
Yes it is worthwhile, less so for gaming, more for general computation. Just tightening the sub-timings you can reduce 2.5Billion Pi Calculation time from 50 seconds to 30 seconds.
That said, something is very wrong with your Write Speeds.
yeah u got it mate. u wasted your money, nothing else. lol. u already had 32gb kit with 6000cl30… stupid…
So the thing is that EXPO on a 6000CL28 and 6000CL30 kit should perform about the same.
I also have an entire video dedicated to how CAS latency on DDR5 basically doesn’t do much of anything for performance and it’s the timings that you have to tune manually that actually make a difference.
You need to benchmark with games if you want to know game performance, not AIDA64.
Most games will only gain 5-10% from RAM tuning under ideal circumstances. If you are GPU limited that number will naturally be much lower, 0% gains are possible.