Upgraded from 1Rx16 to 1Rx8 or 2Rx8 – performance improvements confirmed. - updated with results before and after
Upgraded from 1Rx16 to 1Rx8 or 2Rx8 – performance improvements confirmed. - updated with results before and after
I bought my laptop with 8GB modules (four chips, single side), later discovering they weren’t the best quality. Perhaps because I haven’t done a fresh build in a while, I’m eager to upgrade. This might be the right time. Asus seems to minimize the issue, claiming a typical 5% variation when switching to 1Rx8. Another source suggests a much larger difference—possibly tens of percent—based on system factors. They didn’t specify their baseline specs. Another article referenced a 11370H (4c8t) plus 3050, while my machine is 5800H (8c16t) plus 3070. I’m analyzing this in two levels: rank and bank. My opinion is that if I were to purchase new 16GB modules now, they’d likely be 1Rx8 (usually eight chips, four on each side). If I locate older models, they might be 2Rx8—photos show eight chips on one side, but no image of the other. In the past, switching from 1R to 2R could boost performance significantly, even up to 25% for compute tasks. A quick search suggests buying used units on eBay could match the cost of brand new 1Rx8 kits, without any warranty. It’s unclear if my laptop can handle 32GB modules. I don’t require that amount, but it could help ensure a fresh RAM upgrade in a 2Rx8 setup. Thoughts? Focusing on the technical side.
2xr8 exceeds 1rx8, once the battery for my laptop functions properly I’ll concentrate on filling all slots with 2gb modules. Laptops operate at RAM speeds far below CPU limits, so even a quad-core processor will suffice. Still, dual-rank is preferable over single or half-rank, as it significantly boosts performance. Most systems are tightly restricted on RAM, making a jump from half or single to dual the most effective upgrade.
I discovered Kingston's site now lets you pick your RAM setup. I was weighing 2Rx8 32GB modules but wasn't sure if my laptop could handle them or if I really needed that amount. Instead, I opted for 2Rx8 16GB modules. The order has been placed. It costs more than Crucial, but Crucial only offers 1R at 16GB right now.
Collected the results and completed the evaluation. Prior changes showed minor adjustments: before: 2x Samsung 3200C22 1Rx16 → after: 2x Kingston (Samsung chip) 3200C22 2Rx8 Games at 1080p Final Fantasy XIV Endwalker Benchmark (max): +7% Watchdogs Legion (very high): +34% Call of Duty Modern Warfare II (extreme): +4% Shadow of the Tomb Raider (highest): +12%. For gaming, I opted for the highest feasible settings with native rendering. Variations might have been larger if lower resolutions were used. FFXIV EW also displayed a loading time improvement. I didn’t prioritize this analysis heavily, but performance could improve by 2-5% in that area with the new memory. I ran Y-cruncher 0.7.10 build 9513 (25m +41%) and Prime95 (30.8b17) – throughput jumped significantly. I expected a boost from Prime95, but the gain was much greater than anticipated. Previously, I noticed about a 25% improvement switching from 1R to 2R on similar hardware. The system features a 5800H CPU with strong Zen 3 design, but it only contains half the L3 cache compared to desktop models. Memory feeding has always been a challenge; this faster RAM might help unlock more performance. Aida64 6.25.5400 showed a +9% read speed, +38% write speed, and +19% latency – no change. This version predates the Zen 3 release, so I didn’t update my license. The latency test doesn’t fully reflect CPU performance, but the outcome remained stable. Bandwidth metrics improved generally, though reads only increased by 9%. Writes surged by 38%, with copy performance moderate. I couldn’t link Aida64 bandwidth results to compute outcomes earlier; it seems a mix of bank and rank interleaving enhances overall efficiency.