Notice how they set up tests to be lighter on the GPU than normal, for the single dGPU testing, to showcase what differences there may be, as well, yet still end up with pretty minor improvements from major speed increases, or timings getting tighter. Until you're overclocking enough to want faster RAM, using multiple GPUs, or doing something that the CPU can't predict well enough (say, a large physics sim), the differences tend to be negligible. When they can't do that well enough, RAM sped and timings make a difference.įor most people, with most builds, even programs that get measured benefits tend to get <5% either way, often with greater variations than CL9 to CL11. The programmable CAS latency is listed as per the title. Cases where you might care are also cases where you'd be better off spending on faster RAM, as well.ĬPUs are very good at predicting their futures, and putting what they need in their caches. RAM TIMING Programmable CAS Latency: 11, 10, 9, 8, 7, 6 Build Help. Trust me when I say the only thing you should really worry about for RAM is capacity, and then it's literally just get as much as you need. I haven't built my first rig yet but I'm planning on it soon and have been researching literally for years. I know that it seems super complicated and I know I was super confused as to whether it mattered or not when I first started researching. And from the first reply you cna see it doesn't impact it all that much.ĭon't worry too much about it. For all intents and purposes, I wouldn't worry too much about the other timings the first value gives you a pretty good sense of performance. The values don't have to be the same, and the last one is usually a sum of different values, both reported and not reported. When companies advertise their timing they usually refer to just the first value. DDR4-2666) but the CAS latencies are different (i.e.
#CAS LATENCY 10 VS 11 HOW TO#
Those are different timings that are related to nuances in RAM module state changes, so they relate to how it responds to different things. Find out more about CAS latency, RAM speed, and how to measure memory.
#CAS LATENCY 10 VS 11 UPGRADE#
Buildzoid did a video about this just recently:Įarly DDR5 likely won't be any upgrade at all over a highly-tuned kit of DDR4. The XMP timings of many really fast DDR4 kits are awful and will provide lower real world performance than a slower, but highly tuned kit (but of course a faster kit that's also tuned will be even better). There's no point pumping a ton of extra bandwidth if it means your timings go to crap. Some games care a lot more about system memory than others though, and you'll see by far the biggest benefit when you're CPU-bound. You can easily see from testing of current DDR4 that both higher speeds and tightened timings help gaming performance.
0 kommentar(er)
