Read more.RGB memory modules will be available in kit speeds up to 3600MHz, CL14-15-15-35 timings.
Read more.RGB memory modules will be available in kit speeds up to 3600MHz, CL14-15-15-35 timings.
Ooft, that's nice but it ain't gonna be cheap!
Now we just need to know if faster than default RAM speeds effect performance in a similar manner as they did with previous generations, i assume it does but i would be nice to have that confirmed before spending more than necessary.
It might be late but I don't really understand. I thought that with Zen 2 there was no need for tuned memory for a specific brand of cpus anymore. Does that mean that memory kits are usually tuned for intel cpus in general(by that I mean that shouldn't it work "optimum-ly" with either brand from the start?)? Or is this only due to infinity fabric process...? I don't recall needing specifically tuned memory kits for intel or amd when I built one of my 1st pcs which was an athlon xp barton 2500 when it just got out at that time. If someone could explain, it would be greatly appreciated.
This is all i've found so far.
https://translate.google.com/transla...13&prev=search
Looks like it may make a few fps difference at 1080p running 3600 over 3200, not sure it would at resolutions above that. All of the measurements were at CL18 too so it's not really in depth.
Grab that. Get that. Check it out. Bring that here. Grab anything useful. Take anything good.
Corky34 (09-07-2019)
This is the thing. Faster RAM, tighter timings, etc. are pointless outside of benchmarks. The cost/improvement ratio is nuts! I get wanting the best, and in the past I've definitely gotten parts that I didn't need just to get a bit more oomph, but bleeding edge RAM is just one of those things that will never make sense to me.
If you're just talking about faster RAM then I'd be inclined to agree as most of the advantage gained is lost from having to cross a clock domain, you can only send data as fast as the slowest link, however with Zen you're not just talking about faster RAM as the data fabric connecting everything shares the same clock domain as the memory controller (that changed with Zen2 but the 1:2 divider is rather pointless IMO).
By running faster RAM you're essentially overclocking the bus that connects the chiplets together.
..the only thing this is optimised for is G-Skill's profit margin.
Have you seen this? TechPowerUp has a whole slew of very interesting articles on the new Ryzen 3000 series (and X570 chipset).
I asked Scan to explain a while back and never got an answer, and worse, they seem to mix up Intel and AMD specific terms so I don't think any descriptions can be trusted (Ryzen optimised ram talking about intel motherboard chipsets for eg.)
It's possible that a particular combination of latencies would suit access patterns from a particular type of CPU over another, but very unlikely as that's what more local CPU cache is used for.
Ok. Time will tell))
Apparently my Corsair LPX was meant to be for Intel boards, strange as its been powering my Ryzen very happily :/
Jon
Yea i was looking through all their testing yesterday, it seems there's not as much scaling in performance with Zen2 when it comes to memory, same with PCIe 4.0, and with X570, we're talking single digit performance gains in most situations, unfortunately i can't stop myself looking at the more expensive stuff.
There are currently 1 users browsing this thread. (0 members and 1 guests)