Friday, May 12th 2023

Crucial T700 Gen 5 SSD Throttles Down to HDD Performance Levels Without a Cooler

Crucial T700, the company's flagship M.2 NVMe Gen 5 SSD, runs hot—like every other drive based on the Phison E26-series controller (such as the Corsair MP700). ComputerBase.de discovered what the drive does without some sort of cooling. The E26 controller has a Tjmax value of around 86°C, and what happens when it's reached depends on the drive in question. The Corsair MP700 can turn itself off to protect the controller—something that will definitely cause your machine to hang with a BSOD.

The Crucial T700, on the other hand, aggressively throttles down the controller in an attempt to lower temperatures. While the drive won't stop (and your machine won't hang), its performance drops to hard drive levels, with CrystalDiskMark (CDM) measurements pointing to around 101 MB/s (of course, with much lower access times than a HDD). Both Crucial and Corsair offer the drive with large heatsinks, and recommend users to use them. This should severely limit the adoption of Gen 5 NVMe SSDs among notebooks, where the notebook chassis has room for only bare drives. However, some OEMs specializing in larger high-end gaming notebooks and desktop-replacement workstations, can find ways to connect the drives to the notebook's main cooling system using flattened heatpipes. You can catch ComputerBase's review of the MP700 in the source link below.
Sources: ComputerBase.de (Twitter), ComputerBase.de
Add your own comment

70 Comments on Crucial T700 Gen 5 SSD Throttles Down to HDD Performance Levels Without a Cooler

#51
csendesmark
A&P211I'm only having issues now because of 4K videos, they are much bigger than 720/1080. I started doing 4K videos last year. I never had throttling problems with a sata 2.5in drive in my last laptop.
Do you need the portability?
Can't you do you editing in a workstation instead a laptop?
Posted on Reply
#52
Mussels
Freshwater Moderator
lemoncarbonateWhy is everything getting ridiculously hotter?

Next gen CPU and GPU are getting hotter, now SSD.
I thought next gen should also mean better power efficiency.
When they need a new product to release and don't have one ready, they just overclock the old one.

Refinement takes years, and they're not giving themselves that time - by the time PCI-E gen 4 NVME's had 2-3 revisions under their belt, gen 5 was out and it's back to testing out brand new inefficient stuff again.
Aleksandar_038OK, Gen5 drives are problematic... but, M2 format is also part of the problem.

I know that M2 is slim and fit for laptop. But, proper 2.5" would for sure be easier to cool down properly? Or I am mistaken?
With NVME drives it's usually the controller that overheats, which works fine with a very small amount of cooling material - many just use a thin bit of copper tape

It's usually just an assumption that
"motherboards have NVME coolers already, so let's not include one and save money" leads to people running without cooling at all
Or
"The average user doesnt run into this in our testing based on a bunch of assumptions, so it doesnt matter"
Posted on Reply
#53
ypsylon
One: nobody sane should buy Gen.5 NVMe for gaming and other utterly trivial tasks. Just stay with ordinary SSD or gen3 or 4 NVMe at most and don't waste time with 5.

Two: if you know you'll use the potential then get yourself plenty of airflow in the case.

People are so aghast its so hot and throttles like mad. And how it can be not? Gen5 NVMe controller power draw is on par with enterprise grade RAID controller. If you leave say 8xxx series card from Adaptec/Microsemi/whoever bought them next/ without enough airflow controller will reach 100C in space of few seconds. With each RAID ctrl -usually- comes a memo in the box that you need 200 CFM of airflow for efficient cooling and you have 50x50 radiator installed on the chip.. That's 17W of power draw in a nutshell. Gen5 M.2 NVMe stick can only be worse because it has much smaller surface area to evaporate the heat.

Need a beefy card with huge radiator and plenty of unobstructed airflow. Something along the lines of Accelsior 4M2 or 8M2 form OWC, but these are not Gen 5.
Posted on Reply
#54
Mussels
Freshwater Moderator
ypsylonTwo: if you know you'll use the potential then get yourself plenty of airflow in the case
They're not even a 10 watt device... please dont give advice like that if you dont actually know anything about the topic at hand

None of what you've said has *anything* to do with NVME drives

Posted on Reply
#55
W1zzard
MusselsThey're not even a 10 watt device... please dont give advice like that if you dont actually know anything about the topic at hand

None of what you've said has *anything* to do with NVME drives

Check my mp700 pro review, these drives can exceed 10 w when tested properly. However, in nearly all real life tasks they stay well below that because those tasks don’t run at the drive‘s maximum speed. Like that file copy test is qd1 only, 50% write, possibly out of slc cache, possibly thermal throttled
Posted on Reply
#56
csendesmark
MusselsThey're not even a 10 watt device...
You also need to remember that those devices less than 10g heavy, so a small part don't need a lot of Watts to get really hot.
I have a Samsung 980PRO which is 5g and can use 6W of power, while my 7900 XT Pulse is 1360g where the cooler is about ~1200g of the total weight.
That ~100g PCB uses "only" ~330W, and I cannot tell when I seen a top tier card with passive cooler, so I would say airflow is matters - a lot
So if someone made a ~100W passive card, they would rely on the pc case's airflow to carry away the heat.
Posted on Reply
#57
bug
R-T-BThe controller is the heat source not the NAND, so yeah node shrinks will help.
The controller is already on TSMC's 12nm. Not exactly the latest process, but not too far behind either. (Controllers are traditionally 2 or more nodes behind the cutting edge, do not expect them on N3 anytime soon.)
Posted on Reply
#58
lemoncarbonate
ypsylonOne: nobody sane should buy Gen.5 NVMe for gaming and other utterly trivial tasks. Just stay with ordinary SSD or gen3 or 4 NVMe at most and don't waste time with 5.
I stay with SATA for gaming and daily task. It was tested by Techspot that gaming doesn't benefit from NVME at all.
Posted on Reply
#59
trparky
lemoncarbonateI stay with SATA for gaming and daily task. It was tested by Techspot that gaming doesn't benefit from NVME at all.
Except that most modern motherboards only come with at most four or six SATA ports but at the same time they come with two or three m.2 slots.
Posted on Reply
#60
Mussels
Freshwater Moderator
W1zzardCheck my mp700 pro review, these drives can exceed 10 w when tested properly. However, in nearly all real life tasks they stay well below that because those tasks don’t run at the drive‘s maximum speed. Like that file copy test is qd1 only, 50% write, possibly out of slc cache, possibly thermal throttled
I mean yes, above 10W
But... by this much?


That's not something you need case fans for, or talk about high end cooling

It's something a very small heatsink can handle, or the lightest airflow.
Posted on Reply
#61
chrcoluk
Whats the difference between this and the MP700 TPU reviewed on spec? Curious which is the highest model.

Not a great start, products seem rushed to market, one turns itself off when too hot, the other throttles excessively. Why they trying so bad to save a few $ on not providing with a heatsink?
lemoncarbonateWhy is everything getting ridiculously hotter?

Next gen CPU and GPU are getting hotter, now SSD.
I thought next gen should also mean better power efficiency.
The whole PC tech industry has decided they going all in on performance. Heat, power consumption have clearly gone down the priority list.

It seems utterly ridiculous this path has been chosen on SSDs though as the difference will only be seen in benchmarking.
InVasManiWe'll probably see a node shrink resolve all or most of the larger issues with Gen 5. I think the quick transition from Gen 4 to Gen 5 probably hasn't allowed for as much maturity on the node side of things and this kind of the end result of all of that. Give it time and I'm sure it'll reach better maturity though this early hardware for Gen 5 is looking rather subpar.
Node shrinks harm nand, thats why nand stacking became a thing.
TheinsanegamerNTBF, the biggest source of heat on these drives is not the NAND, it's the controller, and as speeds jack up every gen so does heat. Node improvements will help fix that particular issue.

2.5" would help, so long as the metal casing was used as a heatsink. But that would require some kind of NVMe cable.
One already exists, called U.2, but seems they kept it of consumer machines for the same reason they dont get SAS.

So PCI Express slot drives would be the logical solution, but wait... a lot of board vendors now ship with barely any PCIE slots. (Current example, p460 DC SSD in my machine right now idling at 22C, load temp in 30s, is a 12W SSD under load).
Posted on Reply
#63
chrcoluk
Jismin 2023 it's completely in to start delidding the controller of your SSD and direct-die cool it.
Noctua SSD cooler, coming soon in 2024. :)
bugWelp, I was wrong thinking 10W power draw approaches that of physical drives: documents.westerndigital.com/content/dam/doc-library/en_us/assets/public/western-digital/product/internal-drives/wd-blue-hdd/product-brief-western-digital-wd-blue-mobile-sata-hdd.pdf

It absolutely destroys it :(
Yeah some SATA HDD's can hit above 7w, but the use of helium has lowered the power draw again when its used. Makes helium 7200rpm drives have similar power draw to 5400rpm.
Posted on Reply
#64
bug
chrcolukYeah some SATA HDD's can hit above 7w, but the use of helium has lowered the power draw again when its used. Makes helium 7200rpm drives have similar power draw to 5400rpm.
Seeing how simply shifting bits around draws more power than physically moving stuff kinda puts thing in perspective, doesn't it?
Posted on Reply
#65
Mussels
Freshwater Moderator
And it's important to note that SATA SSD's internally don't fill out that 2.5" space, most of it is empty air - its not used for cooling or anything.

Early 2.5" SSD's looked like this, with multiple flash chips spread out everywhere


Vs a modern WD green 1TB

These can have the exact same issues with a controller overheating, most just use a single thermal pad to connect to whatever enclosure they're in, even if its plastic.
Since flash memory speeds are way above what SATA can achieve, they use as few memory chips as possible making them rather small.
bugWelp, I was wrong thinking 10W power draw approaches that of physical drives: documents.westerndigital.co...f-western-digital-wd-blue-mobile-sata-hdd.pdf

It absolutely destroys it :(
2W to 10W?
5x power consumption for 80x the performance is hardly destroyed, especially when the NVME would finish in a single second what that mech drive would take 10 minutes to complete, if it was best-case for the mech drive (single large write to an empty drive) thus idling faster and using much much less power for the same task.

That WD document shows "average" wattages, not maximum or peak. Averages will always be in an NVME's favour since it's so much faster it'll finish the task and idle while the mech drive is still spinning up. Literally in many cases.
Posted on Reply
#66
csendesmark
@Mussels it would be fantastic to have the m.2 speeds with that 2.5" size drives, so you could put there some heatsinks for the fast ones.
Yes I know about u.2, but for some reason those are not really spreading, because it could give us what I asking for in my last sentence.
Posted on Reply
#67
Mussels
Freshwater Moderator
csendesmark@Mussels it would be fantastic to have the m.2 speeds with that 2.5" size drives, so you could put there some heatsinks for the fast ones.
Yes I know about u.2, but for some reason those are not really spreading, because it could give us what I asking for in my last sentence.
Why? Small heatsinks already get the job done. It's not a not of heat at all, it's just barely too much for no cooling at all.
You'd need PCI-E x4 riser cables to the SSD, which is a bit impractical - you could always use riser cards/RAID cards if you really wanted to, but when i tried that they ran way hotter and i returned it (WD AN1500)
Posted on Reply
#68
csendesmark
MusselsWhy? Small heatsinks already get the job done. It's not a not of heat at all, it's just barely too much for no cooling at all.
You'd need PCI-E x4 riser cables to the SSD, which is a bit impractical - you could always use riser cards/RAID cards if you really wanted to, but when i tried that they ran way hotter and i returned it (WD AN1500)
Hope you are right, but I see a bad trend where SSD-s becoming more and more hot
Posted on Reply
#69
bug
csendesmarkHope you are right, but I see a bad trend where SSD-s becoming more and more hot
Look at it this way: until PCIe 6 makes its way into home PCs, you're nor going to see worse.
Posted on Reply
#70
Mussels
Freshwater Moderator
csendesmarkHope you are right, but I see a bad trend where SSD-s becoming more and more hot
Stop looking at temperatures and look at wattages.
temperatures mean it needs more cooling, wattage shows how hard it will be to cool.

10 watts is nothing, it's just at the edge of being unable to run without cooling when the controller is so small. A heatsink the size of a coin would make a massive difference.
Posted on Reply
Add your own comment
Nov 21st, 2024 11:14 EST change timezone

New Forum Posts

Popular Reviews

Controversial News Posts