the new single rank must be really recent stuff like last 3-4 weeks manufacturing unless they started to be mixing it up.
mine is week 39 2020 which is end of september
Correct...single sided = single rankMy ram is dated 2043. I haven't had it in a machine yet to check it but looking at it it only has ram chips one side of the PCB, leading me to think it's single rank.
Crucial make the best ones for clocking. There are some Samsung and Hynix sticks but they have very little headroom in them and cost around the same. Just stick to Crucial for 32gb sticks.Any other alternatives to crucial for 3600mhz c16 32gb dual ranked sticks?
Crucial make the best ones for clocking. There are some Samsung and Hynix sticks but they have very little headroom in them and cost around the same. Just stick to Crucial for 32gb sticks.
Thanks. Just have to now wait for stock to reappear. I contacted crucial a few weeks back but they didn't know when new stock would be back in.
Is your signature up to date? The X99 bioses are definitely not well optimised for 32gb sticks of any type. I have an R5E, MSi Godlike and a GB X99 SOC and none of them do well with 32gb sticks even dropping down to triple or dual channel. I would be especially cautious over Micron Rev B as its less than a year old. Zero chance of ever getting bios optimisation for it on the X99 platform.
The memory is a for an upgrade. I have a 5900X ordered and will be picking up a X570 Asus Dark Hero when it is released.
M16FE1
This means there are 16 1gb ram chips on the PCB so it will be dual rank. Not as good as single rank as it will put more load on the memory controller if using Ryzen. Not really an issue if you are only using the two sticks and I don't know how much of a performance hit it would take if using four sticks together, maybe someone else can chime in.
Thank you. I'm only intending to use two slots on the MB so hopefully I haven't made a bad buying decision
This video might help https://www.youtube.com/watch?v=AGux0pANft0
https://youtu.be/0oL-SedTy8YThank you for the link. I won't pretend I understand everything that was mentioned in the video but it does seem that I would have had a slight performance increase if I had gone for 4 sticks instead of two. That's a lesson learned for the next time I buy.
Guys, I ordered a white 32GB set (2x16GB) in mid October and they are still sitting sealed in the box.
I've taken a look at the labels on each of the two sticks and can see the following: 2033 BL16G36C16U4WL.M16FE1
I know next to nothing about ram but if I've understood the thread information correctly, I have the good to use version?
I think you've received some fluffy answers here from others!
If you only intend having 32gb, then yes, you have the best memory. With only 2 sticks, dual rank will give better performance than single rank sticks. You might have an issue if you put more memory in your system (ie 64gb), but it sounds like this is unlikely.
OK, thanks Stu. I've struggled to take in the heavier technical descriptions in the videos I've seen and got myself a little lost to be honest. All the channel and rank stuff goes way over my old head! The main thing I noticed were higher scores when using 4 sticks of the same memory over 2 and I drew my conclusion from that.
I guess I'll be OK if I stick with what I have now and be mindful that I could introduce strain on the memory controller if I ever try to add another two 16GB sticks, which I'm unlikely to do.
Thank you to everyone who came back to me with answers and guidance
This has been a hot topic lately! The easiest way to explain, in most cases, you get best performance with 4 ranks of chips... This can be achieved with 4 sticks of single rank (4 X 1 = 4) or 2 sticks of dual rank (2 X 2 = 4).
4 ranks gives better performance than 2 ranks because you have more bandwidth due to the additional ranks.
You might ask, why not go higher than 4 for even more bandwidth? In many systems going above 4 ranks puts more strain on the memory controller... It will still work but you must use slower timings and/or speed, thereby losing performance.