Titles that improve with more memory capacity
Titles that improve with more memory capacity
I focused on Y-cruncher Pi and Prime95 as my main benchmarks for this test. It will be intriguing to observe how CineBench 2024 performs with them. While it can be quite memory-heavy, I haven’t tested it in the same depth as when the new RAM becomes available. I also included a few other rendering benchmarks, but I doubt they’ll reveal much, making them more valuable if my expectations are incorrect. If we’re really evaluating synthetic workloads, I even ran Geekbench. Not because I prefer it, but its popularity suggests I should explore it further, particularly the various subtests. Previously, I used Aida64 for similar purposes, though it seems less common now.
Based on my experience, CB24 does respond somewhat to RAM adjustments—especially with a tuned quad rank and 5800X3D I achieved 920 points. Aida64 gives decent benchmarking but doesn’t reflect real-world RAM speed. Time Spy handles CPU DRAM tuning well, though you’ll need multiple trials because of variability. (Time Spy Extreme is better and more challenging.)
I began testing with RAM setups, but performance was extremely sluggish, causing me to lose interest quickly. I’ll include this in my notes. It matches my needs well for execution. This brings up a thought about 3DMark’s CPU Profile feature—though few seem to use it, and its relevance isn unclear. Edit: Completed six Time Spy CPU runs; the results showed a 2.6% spread from the average, with longer startup time than actual testing.
Like suggested, I should look at the games HuB played and choose a few with strong separation. It really depends on CPU specs, how the game is built, and whether it actually needs high-speed RAM. Getting very fast RAM right now often isn't worth it.
I've received the updated tracking info. The ram should arrive within the next hour. My testing remains focused on performance beyond just gaming. The changes in my system are noticeable: gaming system 7800X3D + 4070 now runs at 4800C40 to 6000C30, showing a 7% average improvement in Watch Dogs Legion 1080p Ultra. The best result was 6%, while SOTTR 1080p peaked at 0.5%. CPU performance is solid—Time Spy averaged +3.6% (6 runs) or +3.9% (6 runs). The differences between runs are significant, but the gap between results stays large. Outside gaming tests, Blender 4.2.0 saw a +2.2% boost with faster RAM versus +0.3% for others. Pov-ray slowed slightly (+0.4%) with the same speedup. Consistent outcomes highlight a power limitation issue (IMC using extra power without gain). Cinebench R15 rose +0.3%, while R23 stayed steady and Cinebench 2024 improved +1.9%. Possibly core count affects results—CPU has only 8 cores, which might limit gains. Y-cruncher Pi models showed ~5% improvement, likely due to lower core usage. Prime95 and 7-Zip compression also demonstrated expected changes. I’ll review Geekbench 6.3.0 later for comparison.
I think Victoria 3 demands significant memory resources as the game progresses, mainly because of its complex systems.
also remember those benchmark scores don’t always reflect real-world speed or suit every situation, cb is practically unimportant when it comes to gaming performance—though I understand why gamers care about these numbers. 3dmark actually provides a more accurate measure for that. in short, yes, but the main takeaway is clear: your results show what matters most.