News CPU Tower-Style Cooler Reduces M.2 SSD Temps By Over 50 Percent

Page 2 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
But you need them as close to the cpu and ram as possible for data integrity, the longer the traces the higher the risk of things going wrong.

This is very incorrect. They do it to save money when building the boards and that's all. U.2 ports (on the few motherboards that have them) are always along the right edge of the board with SATA. Higher end workstation type boards have also been known to put M.2 in the lower right area of the board as well.

The reasoning is identical to the reason we got saddled with the <Mod Edit> M.2 spec for desktop in the first place: it's cheaper for manufacturers.
 
Last edited by a moderator:
  • Like
Reactions: artk2219

LuxZg

Distinguished
Dec 29, 2007
225
42
18,710
The first generation of almost any new interface requires more power than the 2nd and 3rd generation iterations while they refine the details based on field results. There used to be a huge fuss about PCIe 4.0 burning more power than 3.0 when it first came out and it has become a non-issue since. Skip the first two generations of 5.0 hardware and you most likely won't need to think about it anymore.

BTW, due to trace length restrictions before needing retimers, the 5.0x4 NVMe is above the GPU slot on most motherboards.

This is very incorrect. They do it to save money when building the boards and that's all. U.2 ports (on the few motherboards that have them) are always along the right edge of the board with SATA. Higher end workstation type boards have also been known to put M.2 in the lower right area of the board as well.

The reasoning is identical to the reason we got saddled with the <Mod Edit> M.2 spec for desktop in the first place: it's cheaper for manufacturers.

Sort of reply to both. I remember all the fuss about PCIe 4.0 - the power, the trace length limitations, the huge cost, the issues of multilayer motherboards, data integrity - you name it!

Now we have 30cm ribbon extensions for 50$. Yes, it's "expensive", but so are the motherboards of today. Plus that's a retail ribbon price, for something that is low volume enthusiasts market, and obviously doesn't actually cost 50$. I expect similar for PCIe 5.0, al the ho-hum then suddenly it can be done even in ridiculous lengths like 1 meter for 100$.

Taking that into consideration, they could move that M.2 wherever needed, specially on 1000$ MBOs.

Likewise, looking at power, others already pointed correctly that even enthusiast level user would rarely hit any throttling temperatures, and with basic cooling shim/armor that will probably be provided with MBO, it should all be good. Sure, benchmarking SSD speeds for extended periods will eventually hit the limit, but who cares.

The tower itself looks nice and shiny, and not even expensive, if there was space I'd consider one purely for bragging rights. But it's still unlikely to happen :)
 
  • Like
Reactions: thestryker

scottscholzpdx

Honorable
Sep 14, 2017
18
13
10,515
I've got a PCI-E x16 4x NVMe slot card with active air cooling and a giant aluminum heatsink than all drive would touch, and the fan looks like it was taken from one of those old Radeon 9250 AGP cards.

Only have one drive in it right now but the temps on it are ridiculously good compared to my others.
 
pcie 5 cards might be a little warm, my pcie3 nvme hits 55c, how hot can they go? wonders what operating temp will be.
The PCI-E 4.0 SSDs I've seen have been sitting around 55C "idling" in a ~20-25C ambient environment (i.e. the OS is not really doing anything beyond running the desktop and web browser), if that helps...
 

Latest posts