Thursday, June 13th 2024

Realtek to Join the PCIe 5.0 NVMe SSD Controller Race

Currently, only Phison offers widely available performance PCIe 5.0 NVMe controllers for consumer SSDs, although the competition is heating up and Chinese SSD controller makers MaxioTek and InnoGrit, as well as Taiwanese Silicon Motion either have or will be releasing competing controllers this year. However, it also looks like Taiwanese Realtek will be joining the fray, albeit a tad late, as their upcoming RTS5782 PCIe 5.0 NVMe controller, but the company didn't reveal a release schedule. Although Realtek is known as a more budget friendly chipmaker, regardless of the product segment, the RTS5782 will be an 8-channel controller with support for 3600 MT/s NAND, and it will have a dedicated DRAM cache using either DDR4, LPDDR3, LPDDR4 or LPDDR4x memory. It's said to deliver sequential read speeds of up to 14 GB/s and sequential write speeds of up to 12 GB/s, with random read and write IOPS hitting 2500K.

Realtek is also working on a pair of new DRAM-less models, the PCIe 4.0 RTS5776DL and the PCIe r5.0 RTS5781DL, both having support for 4-channel flash, but still supporting NAND speeds of up to 3600 MT/s. Engineering samples of the RTS5776DL are expected to be available before the end of this year, and it's said to deliver sequential read and write speeds of up to 7400 MB/s and random read and write IOPS of 1200K. The RTS5781DL will enter the engineering sample stage in the beginning of 2025 and will up the sequential read and write performance to 10 GB/s and the random read and write IOPS to 1400K. In addition to the new NVMe controllers, Realtek also had its RTS5736DL on display at Computex, which the company claims is the world's lowest power DRAM-less SATA SSD controller, that also supports size of up to 8 TB. Realtek claims the RTS5736DL is an ideal solution to pair with a USB 3.x bridge chip for external drives.
Sources: Tom's Hardware, Impress PC Watch
Add your own comment

21 Comments on Realtek to Join the PCIe 5.0 NVMe SSD Controller Race

#1
Onasi
I have a weird sense of Deja Vu for some reason. Didn’t Realtek announce their entry into the Gen5 controller race before? Am I crazy?
Posted on Reply
#2
TheLostSwede
News Editor
OnasiI have a weird sense of Deja Vu for some reason. Didn’t Realtek announce their entry into the Gen5 controller race before? Am I crazy?
Not that I'm aware of. In fact, they pulled out of the NVMe SSD controller market entirely for a while, as their products weren't very competitive.
There aren't exactly a lot of drives using their controllers. Not a single PCIe 4.0 drive in the database using Realtek. The RTS5766DL is from 2020.
www.techpowerup.com/ssd-specs/#realtek

The Transcend 245S is based on the RTS5772DL, as are some AGI AI818 drives. Another not popular controller.
At least that's a 2023 part.
Posted on Reply
#3
bonehead123
Yee haw...more competition :)

Hopefully this will help bring down the prices/up the capacity of Gen 5 drives real soon, and perhaps get things to the point of moving past the 14GB/s limit, which I know has alot to do with the speed of the NAND itself, which they already seem to be thinking about in their specs, but just sayin....
Posted on Reply
#4
Onasi
@TheLostSwede
My bad then, I guess all the announcements from Computex fried my brain a bit.

@bonehead123
That’s all more or less irrelevant in practice for daily use until random R/W actually move up significantly. And I am not sure it really will. Seems like it’s NAND itself by its very nature is the bottleneck there.
Posted on Reply
#5
TheLostSwede
News Editor
Onasi@TheLostSwede
My bad then, I guess all the announcements from Computex fried my brain a bit.
No need to apologise, not everyone can keep up with all the new things being released.
Posted on Reply
#6
LabRat 891
TheLostSwedeNot that I'm aware of. In fact, they pulled out of the NVMe SSD controller market entirely for a while, as their products weren't very competitive.
There aren't exactly a lot of drives using their controllers. Not a single PCIe 4.0 drive in the database using Realtek. The RTS5766DL is from 2020.
www.techpowerup.com/ssd-specs/#realtek

The Transcend 245S is based on the RTS5772DL, as are some AGI AI818 drives. Another not popular controller.
At least that's a 2023 part.
To be fair... RTL's NVME controllers commonly end up in 'generic' and 'scam' SSDs that no one on TPU would actively *want* (outside of morbid curiosity).
Less a rub against Realtek, and more just the 'market' they found best penetration.


(IMO) What this does mean however, is that Gen5 NVMEs are going to be (and are) getting cheaper. (By observation) with few exceptions, Realtek is a low-cost high-volume device supplier.

Forgive my lacking knowledge but...
(Using these 'affordable' Gen5 Controllers) couldn't companies take older 'surplus' NAND, double-up and double-side the stick (utilizing more NAND channels), throw in some cheap low-voltage LPDDR3/4
-and 'fufil' Gen5 speeds through parallelizing cheap parts on-drive?
Posted on Reply
#7
MachineLearning
TheLostSwedeNot that I'm aware of. In fact, they pulled out of the NVMe SSD controller market entirely for a while, as their products weren't very competitive.
There aren't exactly a lot of drives using their controllers. Not a single PCIe 4.0 drive in the database using Realtek. The RTS5766DL is from 2020.
www.techpowerup.com/ssd-specs/#realtek

The Transcend 245S is based on the RTS5772DL, as are some AGI AI818 drives. Another not popular controller.
At least that's a 2023 part.
I have had the misfortune of dealing with the RTS5763DL and one other Realtek NVMe controller, forgetting the name now. They were on TeamGroup drives (Cardea Z330 and MP33 Pro) that were reviewed as having Phison controllers IIRC... And the performance is terrible after filling SLC cache. Like 200MB/s levels of terrible. On multiple motherboards and OS installs.

I just hope their PCIe 5.0 products are actually good, having a new company that's willing to source their SSD controllers would be nice (so not Samsung, SK Hynix and Micron). At least we have Maxio and Innogrit now.
Posted on Reply
#8
TheLostSwede
News Editor
LabRat 891To be fair... RTL's NVME controllers commonly end up in 'generic' and 'scam' SSDs that no one on TPU would actively *want* (outside of morbid curiosity).
Less a rub against Realtek, and more just the 'market' they found best penetration.
Is this your way of calling Adata scammers?
LabRat 891(IMO) What this does mean however, is that Gen5 NVMEs are going to be (and are) getting cheaper. (By observation) with few exceptions, Realtek is a low-cost high-volume device supplier.
Well, so far, it's been a Phison only market, especially on the high-end, with the only other PCIe 5.0 drive being the Samsung 990 EVO afaik, unless we look at enterprise hardware.
The InnoGrit and MaxioTech controllers don't seem to have appeared in retail as yet and SK Hynix hasn't launched the Platinum P51 yet either. Silicon Motion are as we know, late as well.
So yes, PCIe 5.0 drives are likely to become cheaper soon, but it might not happen until next year.
LabRat 891Forgive my lacking knowledge but...
(Using these 'affordable' Gen5 Controllers) couldn't companies take older 'surplus' NAND, double-up and double-side the stick (utilizing more NAND channels), throw in some cheap low-voltage LPDDR3/4
-and 'fufil' Gen5 speeds through parallelizing cheap parts on-drive?
Sure, if you want a really poor performing drive, you can do that. Eight channel controllers aren't something new and there are already plenty of NVMe SSDs with four chips. There isn't really space for more than that, unless you go with a DRAM-less controller, but they don't support more than four channels.
MachineLearningI have had the misfortune of dealing with the RTS5763DL and one other Realtek NVMe controller, forgetting the name now. They were on TeamGroup drives (Cardea Z330 and MP33 Pro) that were reviewed as having Phison controllers IIRC... And the performance is terrible after filling SLC cache. Like 200MB/s levels of terrible. On multiple motherboards and OS installs.

I just hope their PCIe 5.0 products are actually good, having a new company that's willing to source their SSD controllers would be nice (so not Samsung, SK Hynix and Micron). At least we have Maxio and Innogrit now.
Yeah, I ended up buying an Apacer SSD that used to be Phison based, came with a MaxioTek controller. Sent it back without even plugging it in, since it wasn't clear they'd changed the controller. Now the MAP1602A is supposed to be good for a DRAM-less controller, unlike what you experienced, but even so.

There are a few more SSD controller makers, but Marvell seems to have thrown in the towel on consumer products, alongside Korean FADU who also only does enterprise controllers. JMicron gave up (they apparently sold it off to MaxioTek) and all the other ones are in a similar boat or focus on niche markets. There's Korean Novachips, but I haven't seen any hardware based on their chips.
www.novachips.com/products/flashStorageProcessor.asp
Posted on Reply
#9
LabRat 891
TheLostSwedeIs this your way of calling Adata scammers?
:laugh: While there was originally no intention... I do have a friend that got 'stiffed' on warranty with them. [My experience w/ RTL NVME controllers are AliExpress NVMEs]
I'd be more willing to openly call them 'generic' (along with PNY). -Both are exceptionally notorious for mid-model 'cost optimization'.
TheLostSwedeWell, so far, it's been a Phison only market, especially on the high-end, with the only other PCIe 5.0 drive being the Samsung 990 EVO afaik, unless we look at enterprise hardware.
The InnoGrit and MaxioTech controllers don't seem to have appeared in retail as yet and SK Hynix hasn't launched the Platinum P51 yet either. Silicon Motion are as we know, late as well.
So yes, PCIe 5.0 drives are likely to become cheaper soon, but it might not happen until next year.
So, basically Realtek is 'least late to the party'?
TheLostSwedeSure, if you want a really poor performing drive, you can do that.
This is what I expect, when I see 'Realtek'. I typically assume anything Realtek is 'the budget option', or the 'high value option' at best.
TheLostSwedeEight channel controllers aren't something new and there are already plenty of NVMe SSDs with four chips. There isn't really space for more than that, unless you go with a DRAM-less controller, but they don't support more than four channels.
I see.
After all, 22110 M.2s and PCIEx4 AICs are pretty much 'out of style' in the consumer space.
Posted on Reply
#10
TheLostSwede
News Editor
LabRat 891:laugh: While there was originally no intention... I do have a friend that got 'stiffed' on warranty with them. [My experience w/ RTL NVME controllers are AliExpress NVMEs]
I'd be more willing to openly call them 'generic' (along with PNY). -Both are exceptionally notorious for mid-model 'cost optimization'.
I know the head of product at XPG and he's working uphill to change the mindset at the company.
LabRat 891So, basically Realtek is 'least late to the party'?
They were late to join the part as well.
LabRat 891This is what I expect, when I see 'Realtek'. I typically assume anything Realtek is 'the budget option', or the 'high value option' at best.
Well, that's what they're mostly known for, good enough products at an affordable price point.
LabRat 891I see.
After all, 22110 M.2s and PCIEx4 AICs are pretty much 'out of style' in the consumer space.
22110 never really happened in the consumer space afaik, neither did the wider 2580/25110 form factor that was expected.
www.techpowerup.com/295276/m-2-pcie-5-0-ssds-set-to-increase-to-25-mm-in-width-might-not-fit-older-motherboards
Posted on Reply
#11
Chrispy_
Some competition is needed for sure.

Phison haven't made anything yet that will fit in a laptop without catching fire. Every Phison Gen5 drive runs at Gen4 speeds (or lower) without active cooling, so what's the point?
Posted on Reply
#12
MachineLearning
TheLostSwedeYeah, I ended up buying an Apacer SSD that used to be Phison based, came with a MaxioTek controller. Sent it back without even plugging it in, since it wasn't clear they'd changed the controller. Now the MAP1602A is supposed to be good for a DRAM-less controller, unlike what you experienced, but even so.

There are a few more SSD controller makers, but Marvell seems to have thrown in the towel on consumer products, alongside Korean FADU who also only does enterprise controllers. JMicron gave up (they apparently sold it off to MaxioTek) and all the other ones are in a similar boat or focus on niche markets. There's Korean Novachips, but I haven't seen any hardware based on their chips.
www.novachips.com/products/flashStorageProcessor.asp
Oh, we're forgetting TenaFE too.

I can understand sending the Maxio drive back on principle, although personally I'd like the MAP1602A over any other DRAMless controller and most other DRAM-using controllers too. In the latter case, purely due to low power consumption and low heat generation.
Posted on Reply
#13
TheLostSwede
News Editor
Chrispy_Some competition is needed for sure.

Phison haven't made anything yet that will fit in a laptop without catching fire. Every Phison Gen5 drive runs at Gen4 speeds (or lower) without active cooling, so what's the point?
Hopefully the upcoming Silicon Motion SM2508 drives will fix that problem.
www.techpowerup.com/323430/silicon-motions-sm2508-set-to-launch-in-q4-edging-out-phison-as-top-ssd-controller

Not enough is known with regards to thermals when it comes to InnoGrit and MaxioTech.

InnoGrits IG5666 was shown at last years Computex. Seems like they had to node shrink it from 16 to 12 nm.
www.techpowerup.com/309462/innogrit-is-readying-consumer-pcie-5-0-nvme-ssd-controller-for-q4
And Team Group apparently has a drive based on it.
www.techpowerup.com/318286/team-group-releases-t-force-ge-pro-nvme-gen-5-ssd
No-one appears to have done any kind of power draw/thermal testing on it though and Team Group doesn't provide any figures.
MachineLearningOh, we're forgetting TenaFE too.
Have you seen any products based around their controllers?
MachineLearningI can understand sending the Maxio drive back on principle, although personally I'd like the MAP1602A over any other DRAMless controller and most other DRAM-using controllers too. In the latter case, purely due to low power consumption and low heat generation.
Well, it seems to be a decent controller, but when you expect one thing and get another... It goes back. It was also supposed to be my game drive, not a laptop drive.
Posted on Reply
#14
Random_User
bonehead123Yee haw...more competition :)

Hopefully this will help bring down the prices/up the capacity of Gen 5 drives real soon, and perhaps get things to the point of moving past the 14GB/s limit, which I know has alot to do with the speed of the NAND itself, which they already seem to be thinking about in their specs, but just sayin....
I might be wrong, but the problem of current PCIE gen 5 SSDs are not the speed limit itself, but rather random write/read speeds specifically. They all push these beyond overheating, but that's all about sequential speeds, unfortunately. I'd rather buy an "older" cooler and more reliable TLC PCIE 4.0/3.0 than the any current gen 5 furnace garbage, that barely shows any marginal benefit.
Posted on Reply
#15
MachineLearning
TheLostSwedeHave you seen any products based around their controllers?
www.techpowerup.com/review/neo-forza-nfp495-4-tb/
TheLostSwedeWell, it seems to be a decent controller, but when you expect one thing and get another... It goes back. It was also supposed to be my game drive, not a laptop drive.
Right, on principle I'd send it back too unless I were curious about the performance (like my Realtek drives).

I'd honestly be fine with it as a game drive or even OS drive, DRAMless doesn't bother me with MAP1602A's performance.
Posted on Reply
#16
zmeul
if they do SSD controllers like they do the audio chips ... nah, I'm good
gen5 SSDs are already too hot for their own good
Posted on Reply
#17
TheLostSwede
News Editor
MachineLearningwww.techpowerup.com/review/neo-forza-nfp495-4-tb/
Ah, a singular drive, no wonder I had missed that. Seems like the pricing is no longer that attractive.
MachineLearningRight, on principle I'd send it back too unless I were curious about the performance (like my Realtek drives).

I'd honestly be fine with it as a game drive or even OS drive, DRAMless doesn't bother me with MAP1602A's performance.
That was in response to them power power efficient, nothing else.
zmeulif they do SSD controllers like they do the audio chips ... nah, I'm good
gen5 SSDs are already too hot for their own good
Did you bother reading this?
www.techpowerup.com/323430/silicon-motions-sm2508-set-to-launch-in-q4-edging-out-phison-as-top-ssd-controller
It's only Phison that is having thermal issues, as they seem to refuse to move to more advanced production nodes.
Considering Realtek is still developing their PCIe 5.0 SSD controllers, we have no idea about things like thermals.
Posted on Reply
#18
chrcoluk
Perhaps the lack of competition thus far, they have read the room correctly and observed many people arent in a rush to buy gen 5 drives?
Posted on Reply
#21
uuee
TheLostSwedeYeah, I ended up buying an Apacer SSD that used to be Phison based, came with a MaxioTek controller. Sent it back without even plugging it in, since it wasn't clear they'd changed the controller. Now the MAP1602A is supposed to be good for a DRAM-less controller, unlike what you experienced, but even so.
Happened exactly the same to me. Apacer AS2280Q4 came with a MAP1602A instead of Phison E16. I've sent it back without even plugging it in. Sadly i did't make any photos:(
Posted on Reply
Add your own comment
Dec 23rd, 2024 12:10 EST change timezone

New Forum Posts

Popular Reviews

Controversial News Posts