Silicon Motion's SM2508 PCIe 5.0 NVMe SSD Controller is as Power Efficient as Promised

[TheLostSwede]

The first reviews of Silicon Motion's new PCIe 5.0 NVMe SSD controller, the SM2508 are starting to appear online, and the good news is that the controller is as power efficient as promised by the company. Tom's hardware has put up their review of a reference design M.2 SSD from Silicon Motion and in their testing, equipped with 1 TB of Kioxia's 162-layer BiCS6 TLC NAND. It easily bests the competition when it comes to power efficiency. In their file copy test, it draws nearly two watts less than its nearest competitor and as much as three watts less than the most power hungry drive. It's still using about one watt more than the best PCIe 4.0 drives, but it goes to show that the production nodes matters, as the SM2508 is produced on a 6 nm node, compared to 12 nm for Phison's E26.

We should point out that the peak power consumption did go over nine watts, but only one of the Phison E26 drives managed to stay below 10 watts here. The most power hungry PCIe 5.0 SSD controller in the test, the InnoGrit IG5666 peaks at nearly 14 watts for comparison. Idle power consumption of the SM2508 is also very good, still drawing more than the PCIe 4.0 drives it was tested against, but far less than any of the other PCIe 5.0 drives. What about performance you ask? The reference drive places itself ahead of all the Phison E26 drives when it comes to sequential file transfers, regardless if it's to or from the drive. Random read IOPS also places right at the top, but it's somewhat behind when it comes to random writes, without being a slow drive by any means. Overall we're looking at a very promising new SSD controller from Silicon Motion with the SM2508 and TPU has also received a sample that is currently undergoing testing, so expect a review here soon.



View at TechPowerUp Main Site | Source

 

[TheDeeGee]

I wonder, are there actually people who refuse to buy a storage drive because it's power usage?

 

[TheLostSwede]

I wonder, are there actually people who refuse to buy a storage drive because it's power usage?

The issue has become cooling, as the Phison and InnoGrit based drives ideally need active cooling to deliver the performance on offer, without throttling.
The SM2508 shouldn't need active cooling, just like PCIe 3.0 and 4.0 drives.
There's also an issue with using PCIe 5.0 drives in laptops, as they'll have a negative affect on the battery life.
Outside of that, I guess not.

 

[bob3002]

I wonder, are there actually people who refuse to buy a storage drive because it's power usage?

Beyond a certain thermal power, you have to put an active fan onto the SSD to prevent throttling... which precludes its usage in really thin laptops, I would think.

 

[Hagal77]

The power consumption is important in this case because the PCiE v5 NVME gets extremely hot and then throttles or crashes. @TheDeeGee

 

[CosmicWanderer]

Why are they still making them on what is essentially a 7nm node when they could have gone for at least 5nm. I bet that would have matched the efficiency of PCIe 4.0 controllers.

Cost I'm guessing, but I would have thought that given the push for 3nm, 5nm would be more cost effective by now.

 

[TheLostSwede]

Why are they still making them on what is essentially a 7nm node when they could have gone for at least 5nm. I bet that would have matched the efficiency of PCIe 4.0 controllers.

Cost I'm guessing, but I would have thought that given the push for 3nm, 5nm would be more cost effective by now.

You guessed right and it's not, as those nodes are utilised to the max and have increased quite a bit in cost.

 

[Assimilator]

What is it exactly that makes PCIe 5.0 so much more difficult/energy-intensive than 4.0, I wonder.

 

[Prince Valiant]

I wonder, are there actually people who refuse to buy a storage drive because it's power usage?

I have but it was for something intended to run most of the day every day.

 

[sLowEnd]

I wonder, are there actually people who refuse to buy a storage drive because it's power usage?

Absolutely. Power usage translates directly to heat output, and some devices don't allow for proper dissipation of said heat. (Space and/or airflow constraints)
Battery life is another possible reason.

 

[TheDeeGee]

The issue has become cooling, as the Phison and InnoGrit based drives ideally need active cooling to deliver the performance on offer, without throttling.
The SM2508 shouldn't need active cooling, just like PCIe 3.0 and 4.0 drives.
There's also an issue with using PCIe 5.0 drives in laptops, as they'll have a negative affect on the battery life.
Outside of that, I guess not.

Right, forgot that part.

Didn't think 10-15 watts was such a big deal to cool, but when you're dealing with a very small surface area it makes sense.

 

[EatingDirt]

The issue has become cooling, as the Phison and InnoGrit based drives ideally need active cooling to deliver the performance on offer, without throttling.
The SM2508 shouldn't need active cooling, just like PCIe 3.0 and 4.0 drives.
There's also an issue with using PCIe 5.0 drives in laptops, as they'll have a negative affect on the battery life.
Outside of that, I guess not.

I wouldn't be so sure this won't need active cooling to not thermally throttle at all. It still uses 30% more power than the best PCIe 4.0 drive(SN850X) on the test, and has 40% higher max power consumption.

It's definitely more efficient than the rest of the Gen5 drives, which is nice I guess. I don't think efficiency is a huge issue when it comes to most peoples use-cases, which is why I imagine companies haven't been pushing for more efficient nodes, as that will up the cost of the drives themselves.

 

[TheLostSwede]

I wouldn't be so sure this won't need active cooling to not thermally throttle at all. It still uses 30% more power than the best PCIe 4.0 drive(SN850X) on the test, and has 40% higher max power consumption.

Heatsink yes, but that's passive cooling. I very much doubt it'll need a fan.

 

[Wirko]

Phison has a less hungry controller too, the E31T. Four-channel but potentially still good performance with newer, faster flash chips. It doesn't seem popular, however. I know of a single SSD from MSI that uses it, not sure if it has come to market yet.

You guessed right and it's not, as those nodes are utilised to the max and have increased quite a bit in cost.

Also, if it's the analog parts (the PHY) that draw significant power, there would be little benefit for a much increased cost.

As a point of reference, has anyone thoroughly tested the 990 EVO's power draw?

 

[Steevo]

What is it exactly that makes PCIe 5.0 so much more difficult/energy-intensive than 4.0, I wonder.

Faster data means faster switching, and capacitive reactance to drive the memory chips themselves at high rates, and more calculations for multi level cell memory.

 

[NeuralNexus]

I'm more curious about the new in-house controller from Western Digital.

 

[GabrielLP14]

Freaking finally an efficient controller, can't wait for mine to arrive

 

[TheLostSwede]

Phison has a less hungry controller too, the E31T. Four-channel but potentially still good performance with newer, faster flash chips. It doesn't seem popular, however. I know of a single SSD from MSI that uses it, not sure if it has come to market yet.

Yes? But it hasn't been sent out to reviewers as yet, so we have no idea how well it performs. This news post was about the fact that the SM2508 actually delivered on the promised of being more power efficient.

I'm more curious about the new in-house controller from Western Digital.

And I'm sure we'll have news about it once they release it.

 

[Steevo]

What is it exactly that makes PCIe 5.0 so much more difficult/energy-intensive than 4.0, I wonder.

Faster data means faster switching, and capacitive reactance to drive the memory chips themselves at high rates, and more calculations for multi level cell memory.

 

[Wirko]

Yes? But it hasn't been sent out to reviewers as yet, so we have no idea how well it performs. This news post was about the fact that the SM2508 actually delivered on the promised of being more power efficient.

Everybody has brought their flagship products to the market first, that's quite normal, but in the case of PCIe 5.0 SSDs, even most enthusiasts have no idea what to do with them. Second tier controllers and SSDs (reaching around 10 GB/s) might be more interesting... if the power consumption is considerably lower, and price too.

There were some leaks about a 4-channel controller from SMI (SM2504XT), you too reported on that, but that was a year and a half ago. It's surprising that they haven't announced anything by now.

 

[Minus Infinity]

Good news on efficiency and power but why would I want to buy a BICS6 drive in late 2024 with BICS8 in production

 

[Hagal77]

I hope Marvell finaly bring out his Gen5 Controller "Bravera SC5"

 

[_roman_]

I wonder, are there actually people who refuse to buy a storage drive because it's power usage?

The power on consumption of the device is unnecessary too high.

More heat in the computer case and less thermal budget for graphic cards and processor. which results in higher fan speeds of any computer fan. which results in a hotter gaming room in summer.

-- I have a KC3000 2TB drive. A newer drive should use less power and be faster for any task in the 2TB size category.

 

[TheinsanegamerN]

The power on consumption of the device is unnecessary too high.

More heat in the computer case and less thermal budget for graphic cards and processor. which results in higher fan speeds of any computer fan. which results in a hotter gaming room in summer.

-- I have a KC3000 2TB drive. A newer drive should use less power and be faster for any task in the 2TB size category.

Temperature =! Heat. Yes, the SSDs run over 90c under hard load, but the power draw is still 10-15 watts. If 10 watts is eating into your thermal budget, your case needs a serious cooling upgrade. Your fans draw more power.

A GPU running at 50C and pulling 100w of power will put out 10x the heat into the chassis that the 99C 10w SSD is doing.

Right, forgot that part.

Didn't think 10-15 watts was such a big deal to cool, but when you're dealing with a very small surface area it makes sense.

Biggest issue is just requiring stupid large heatsinks or tiny fans to keep them from throttling, neither is desirable.

 

[TheLostSwede]

Good news on efficiency and power but why would I want to buy a BICS6 drive in late 2024 with BICS8 in production

I guess you wouldn't and it's unlikely to be the NAND any of Silicon Motion's partners will use when they build drives.
The drive tested is a reference design and most likely a controller validation platform, not a product that's likely to appear in retail.
As such, we should also be able to expect better performance from the controller when paired with faster/better NAND.

Everybody has brought their flagship products to the market first, that's quite normal, but in the case of PCIe 5.0 SSDs, even most enthusiasts have no idea what to do with them. Second tier controllers and SSDs (reaching around 10 GB/s) might be more interesting... if the power consumption is considerably lower, and price too.

So far, there hasn't been any tangible benefits. I haven't wasted my money on a PCIe 5.0 SSD.

There were some leaks about a 4-channel controller from SMI (SM2504XT), you too reported on that, but that was a year and a half ago. It's surprising that they haven't announced anything by now.

Good memory there, yeah, no idea what happened to that, but as we know, roadmaps changes sometimes.