Failing A PCIe 5.0 NVMe SSD In Less Than 3 Minutes Without Extra Cooling

[phoronix]

Phoronix: Failing A PCIe 5.0 NVMe SSD In Less Than 3 Minutes Without Extra Cooling

Following the Corsair MP700 PCIe 5.0 NVMe SSD review under Linux with many readers being surprised by file-system errors when not adding extra cooling like the motherboard's passive M.2 heatsink and being curious about the situation myself, here are more tests of ultimately how this drive will reliably hit file-system errors in three minutes or less without added cooling...

https://www.phoronix.com/news/PCIe5-NVMe-SSD-Fail-3-Minutes

[schmidtbag]

Considering the evidence this is temperature related, it shocks me that Corsair didn't implement proper thermal throttling. I can kind of understand why they wouldn't bother with better cooling, because the kind of people who would buy this would have a motherboard or AIB card that already has a SSD heatsink/shroud. But even then... just include an easily-removable heatsink.

[Michael]

Was just pointed to this that it also happens under Windows with this drive - https://www.techpowerup.com/review/c...00-2-tb/8.html

[Anux]

That certainly is an error in the firmware or even hardware? The device should start to throttle when temps get to high and eventually become slow as fuck but r/w errors are a major flaw.

[flower]

But even then... just include an easily-removable heatsink.

Would an heatsink even solve that problem?
IMHO it would just delay it. Probably Ok for longer bursts but i wouldnt trust that drive without thermal throttling.

being slower is acceptable. raising errors is not

[schmidtbag]

Would an heatsink even solve that problem?
IMHO it would just delay it. Probably Ok for longer bursts but i wouldnt trust that drive without thermal throttling.

As far as I understand, SSD peak performance only works in busts anyway since much of that performance is just on-board DRAM. So long as the heatsink is large enough to keep up with that burst, it ought to be fine. But still... it's stupid for a drive to have to thermal throttle, or need active cooling.

[RejectModernity]

Considering the evidence this is temperature related, it shocks me that Corsair didn't implement proper thermal throttling. I can kind of understand why they wouldn't bother with better cooling, because the kind of people who would buy this would have a motherboard or AIB card that already has a SSD heatsink/shroud. But even then... just include an easily-removable heatsink.

But if you throttle it it won't show you all those cosmic super speeds of PCI-E Gen 5

[bug77]

Considering the evidence this is temperature related, it shocks me that Corsair didn't implement proper thermal throttling. I can kind of understand why they wouldn't bother with better cooling, because the kind of people who would buy this would have a motherboard or AIB card that already has a SSD heatsink/shroud. But even then... just include an easily-removable heatsink.

Michael got lucky, TPU reviewers had to hard reset to even see the drive again.
Also, if you follow that thread, a SM representative stated this isn't how the drive is supposed to work and they're looking into it. There seem to be two models of this drive out there. The one without the heatsink is aimed at motherboard that include their own heatsinks (though Corsair does little to make this clear).

Considering PCIe 4 sequential speeds cannot be sustained, I would have expected nothing less from a PCIe 5 drive.

[skeevy420]

Would an heatsink even solve that problem?
IMHO it would just delay it. Probably Ok for longer bursts but i wouldnt trust that drive without thermal throttling.

being slower is acceptable. raising errors is not

Real world answer -- It might for one of the two M.2 slots in my system. The slot that's above my GPU and directly between an intake and exhaust fan. The other M.2 slot is literally in the gap between my GPU and motherboard. It'd have to use a low profile water cooled loop, which I ain't doing just for one M.2 drive, or, maybe, I could add in a 6th case fan acting as bottom exhaust.

I find it very bad for Corsair that they advertise an 85C max operating temperature, it craps out at 87C, and is a part that's usually placed next to GPUs that have parts that can run between 90-110C. How the hell were they testing these?

[Joe2021]

Okay, lets assume this is a simple firmware bug, maybe something as simple as the threshold for throttling being set to a too high value.. e.g. 87°C instead of 78°C. Fine. Shit happens.

But this should not have gone unnoticed during intense testing before mass production. They do such tests, right? Right?