• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

SK Hynix Unveils 4D NAND Flash Memory Concept

btarunr

Editor & Senior Moderator
Staff member
Joined
Oct 9, 2007
Messages
41,799 (8.19/day)
Location
Hyderabad, India
Processor AMD Ryzen 7 2700X
Motherboard ASUS ROG Strix B450-E Gaming
Cooling AMD Wraith Prism
Memory 2x 16GB Corsair Vengeance LPX DDR4-3000
Video Card(s) Palit GeForce RTX 2080 SUPER GameRock
Storage Western Digital Black NVMe 512GB
Display(s) BenQ 1440p 60 Hz 27-inch
Case Corsair Carbide 100R
Audio Device(s) Creative Sound Blaster Recon3D PCIe
Power Supply Cooler Master MWE Gold 650W
Mouse ASUS ROG Strix Impact
Keyboard Microsoft Sidewinder X4
Software Windows 10 Pro
3D NAND flash revolutionized flash storage as it used the third dimension (height) to stack multiple NAND flash layers, resulting in infinitesimally smaller footprint and reduced costs. SK Hynix believes that a "4-dimensional" NAND flash package is possible. Don't worry, such a stack doesn't look like a tesseract. Conventional 3D NAND flash relies on stacks of charge-trap flash (CTF) cells spatially located alongside its periphery block (which is responsible for wiring out each of the layers of the CTF stack). On a 2-D plane you'd be spending substrate real-estate on both the CTF and periphery block.

SK Hynix believes that the periphery block can be stacked along with the CTF stack, with microscopic vias wiring up the stack along the periphery, reducing the footprint of each cell stack. 4D stacking will also allow for greater number of CTF stacks per cell. Just to be clear, we're talking about stacks of cell and not stacks of NAND flash dies. The V5 cell-stack in SK Hynix's design entails 4 cells and periphery blocks sandwiched. The first implementation of this technology is a 96-layer 4D NAND flash chip with 512 Gb of capacity and TLC (3 bits per cell) density, although the technology is ready for QLC cells. This 512 Gb chip will begin sampling by the end of 2018, and the company is already working on a 1 Tb chip for 2019.



View at TechPowerUp Main Site
 

bug

Joined
May 22, 2015
Messages
9,234 (3.98/day)
Processor Intel i5-6600k (AMD Ryzen5 3600 in a box, waiting for a mobo)
Motherboard ASRock Z170 Extreme7+
Cooling Arctic Cooling Freezer i11
Memory 2x16GB DDR4 3600 G.Skill Ripjaws V (@3200)
Video Card(s) EVGA GTX 1060 SC
Storage 500GB Samsung 970 EVO, 500GB Samsung 850 EVO, 1TB Crucial MX300 and 3TB Seagate
Display(s) HP ZR24w
Case Raijintek Thetis
Audio Device(s) Audioquest Dragonfly Red :D
Power Supply Seasonic 620W M12
Mouse Logitech G502 Proteus Core
Keyboard G.Skill KM780R
Software Arch Linux + Win10
This seems to be strictly a manufacturing improvement, it won't improve endurance or speed.
 
Joined
Nov 11, 2004
Messages
9,029 (1.46/day)
Location
Formosa
System Name Overlord Mk MXVI
Processor AMD Ryzen 7 3800X
Motherboard Gigabyte X570 Aorus Master
Cooling Corsair H115i Pro
Memory 32GB Viper Steel 3600 DDR4 @ 3800MHz 16-19-16-19-36
Video Card(s) Gigabyte RTX 2080 Gaming OC 8G
Storage 1TB WD Black NVMe (2018), 2TB Viper VPN100, 1TB WD Blue 3D NAND
Display(s) Asus PG27AQ
Case Corsair Carbide 275Q
Audio Device(s) Corsair Virtuoso SE
Power Supply Corsair RM750
Mouse Logitech G502 Lightspeed
Keyboard Wooting Two
Software Windows 10 Pro
Benchmark Scores https://valid.x86.fr/33u9si
This seems to be strictly a manufacturing improvement, it won't improve endurance or speed.

Did you even bother reading the slides? SKHynix claims 25-30% performance improvement, but apparently to you, that's nothing. It's not likely to improve endurance, as it doesn't change anything with regards to the memory cells, but it's possible that the improved power efficiency might do something to improve the endurance a smidgen.
 

bug

Joined
May 22, 2015
Messages
9,234 (3.98/day)
Processor Intel i5-6600k (AMD Ryzen5 3600 in a box, waiting for a mobo)
Motherboard ASRock Z170 Extreme7+
Cooling Arctic Cooling Freezer i11
Memory 2x16GB DDR4 3600 G.Skill Ripjaws V (@3200)
Video Card(s) EVGA GTX 1060 SC
Storage 500GB Samsung 970 EVO, 500GB Samsung 850 EVO, 1TB Crucial MX300 and 3TB Seagate
Display(s) HP ZR24w
Case Raijintek Thetis
Audio Device(s) Audioquest Dragonfly Red :D
Power Supply Seasonic 620W M12
Mouse Logitech G502 Proteus Core
Keyboard G.Skill KM780R
Software Arch Linux + Win10
Did you even bother reading the slides? SKHynix claims 25-30% performance improvement, but apparently to you, that's nothing. It's not likely to improve endurance, as it doesn't change anything with regards to the memory cells, but it's possible that the improved power efficiency might do something to improve the endurance a smidgen.
It uses the same components only arranged differently. I don't see how that would improve performance. Like, at all.
 
Joined
Mar 29, 2018
Messages
590 (0.46/day)
It uses the same components only arranged differently. I don't see how that would improve performance. Like, at all.



lol... so does a model t ford a new corvette 4 tires engion block pistons gas , steering wheel , seats ,windshield, transmission , a driver , ect....... so the new corvett is no better that that 100 year old model t ?

oh, and the model t was only about 500 bucks not 50+ thousand$
 

bug

Joined
May 22, 2015
Messages
9,234 (3.98/day)
Processor Intel i5-6600k (AMD Ryzen5 3600 in a box, waiting for a mobo)
Motherboard ASRock Z170 Extreme7+
Cooling Arctic Cooling Freezer i11
Memory 2x16GB DDR4 3600 G.Skill Ripjaws V (@3200)
Video Card(s) EVGA GTX 1060 SC
Storage 500GB Samsung 970 EVO, 500GB Samsung 850 EVO, 1TB Crucial MX300 and 3TB Seagate
Display(s) HP ZR24w
Case Raijintek Thetis
Audio Device(s) Audioquest Dragonfly Red :D
Power Supply Seasonic 620W M12
Mouse Logitech G502 Proteus Core
Keyboard G.Skill KM780R
Software Arch Linux + Win10
lol... so does a model t ford a new corvette 4 tires engion block pistons gas , steering wheel , seats ,windshield, transmission , a driver , ect....... so the new corvett is no better that that 100 year old model t ?
My apologies, the striking similarities between the workings of flash and a car must have escaped me when I posted that. :wtf:
 
Joined
Mar 29, 2018
Messages
590 (0.46/day)
My apologies, the striking similarities between the workings of flash and a car must have escaped me when I posted that. :wtf:


its fine I was joshing some

thing is they promote this stuff as a sales pitch [opinion] what it comes down to is it all that at your house in your computer ? cause funny how in the end at your house it don't quite do as claimed in the promotions and reviews .
 

bug

Joined
May 22, 2015
Messages
9,234 (3.98/day)
Processor Intel i5-6600k (AMD Ryzen5 3600 in a box, waiting for a mobo)
Motherboard ASRock Z170 Extreme7+
Cooling Arctic Cooling Freezer i11
Memory 2x16GB DDR4 3600 G.Skill Ripjaws V (@3200)
Video Card(s) EVGA GTX 1060 SC
Storage 500GB Samsung 970 EVO, 500GB Samsung 850 EVO, 1TB Crucial MX300 and 3TB Seagate
Display(s) HP ZR24w
Case Raijintek Thetis
Audio Device(s) Audioquest Dragonfly Red :D
Power Supply Seasonic 620W M12
Mouse Logitech G502 Proteus Core
Keyboard G.Skill KM780R
Software Arch Linux + Win10
its fine I was joshing some

thing is they promote this stuff as a sales pitch [opinion] what it comes down to is it all that at your house in your computer ? cause funny how in the end at your house it don't quite do as claimed in the promotions and reviews .
I'm not concerned with marketing.
But the engineer in me wants to know how moving the wiring from besides the cell, under the cell can improve the performance. Because, going by what was released here, that doesn't compute.
Back to the car analogy, I can move the wiring inside a car as much as I want. It won't make it go any faster.
 
Joined
Mar 29, 2018
Messages
590 (0.46/day)
I'm not concerned with marketing.
But the engineer in me wants to know how moving the wiring from besides the cell, under the cell can improve the performance. Because, going by what was released here, that doesn't compute.
Back to the car analogy, I can move the wiring inside a car as much as I want. It won't make it go any faster.


i'll imagine it will be they manipulated the controller when it all said and done . like you say if the exact same parts you'd think its got to be in the software ? its still a tlc in the end
 

bug

Joined
May 22, 2015
Messages
9,234 (3.98/day)
Processor Intel i5-6600k (AMD Ryzen5 3600 in a box, waiting for a mobo)
Motherboard ASRock Z170 Extreme7+
Cooling Arctic Cooling Freezer i11
Memory 2x16GB DDR4 3600 G.Skill Ripjaws V (@3200)
Video Card(s) EVGA GTX 1060 SC
Storage 500GB Samsung 970 EVO, 500GB Samsung 850 EVO, 1TB Crucial MX300 and 3TB Seagate
Display(s) HP ZR24w
Case Raijintek Thetis
Audio Device(s) Audioquest Dragonfly Red :D
Power Supply Seasonic 620W M12
Mouse Logitech G502 Proteus Core
Keyboard G.Skill KM780R
Software Arch Linux + Win10
i'll imagine it will be they manipulated the controller when it all said and done . like you say if the exact same parts you'd think its got to be in the software ? its still a tlc in the end
It's probably other improvements that haven't been announced. I would know whey are.
 
Joined
Nov 11, 2004
Messages
9,029 (1.46/day)
Location
Formosa
System Name Overlord Mk MXVI
Processor AMD Ryzen 7 3800X
Motherboard Gigabyte X570 Aorus Master
Cooling Corsair H115i Pro
Memory 32GB Viper Steel 3600 DDR4 @ 3800MHz 16-19-16-19-36
Video Card(s) Gigabyte RTX 2080 Gaming OC 8G
Storage 1TB WD Black NVMe (2018), 2TB Viper VPN100, 1TB WD Blue 3D NAND
Display(s) Asus PG27AQ
Case Corsair Carbide 275Q
Audio Device(s) Corsair Virtuoso SE
Power Supply Corsair RM750
Mouse Logitech G502 Lightspeed
Keyboard Wooting Two
Software Windows 10 Pro
Benchmark Scores https://valid.x86.fr/33u9si
So because you don't understand something, it's then impossible?
It might be as simple that the data path is improved somehow by doing this.
I guess we'll have to wait for a proper explanation once these products are available, but it's not like a company of this size would go out with something like if they hadn't tested it and seen a performance improvement.
 
Joined
Apr 12, 2013
Messages
4,349 (1.41/day)
So because you don't understand something, it's then impossible?
It might be as simple that the data path is improved somehow by doing this.
I guess we'll have to wait for a proper explanation once these products are available, but it's not like a company of this size would go out with something like if they hadn't tested it and seen a performance improvement.
First of all these are just claims atm, remember Octane & those crazy numbers by Intel? SKH will have to go above & beyond what most SSD manufacturers claim to get this large an improvement over traditional 3D NAND, if such gains were this easy we'd have SSD's mainstream by now & Seagate would be bankrupt.
 
Joined
Nov 11, 2004
Messages
9,029 (1.46/day)
Location
Formosa
System Name Overlord Mk MXVI
Processor AMD Ryzen 7 3800X
Motherboard Gigabyte X570 Aorus Master
Cooling Corsair H115i Pro
Memory 32GB Viper Steel 3600 DDR4 @ 3800MHz 16-19-16-19-36
Video Card(s) Gigabyte RTX 2080 Gaming OC 8G
Storage 1TB WD Black NVMe (2018), 2TB Viper VPN100, 1TB WD Blue 3D NAND
Display(s) Asus PG27AQ
Case Corsair Carbide 275Q
Audio Device(s) Corsair Virtuoso SE
Power Supply Corsair RM750
Mouse Logitech G502 Lightspeed
Keyboard Wooting Two
Software Windows 10 Pro
Benchmark Scores https://valid.x86.fr/33u9si
First of all these are just claims atm, remember Octane & those crazy numbers by Intel? SKH will have to go above & beyond what most SSD manufacturers claim to get this large an improvement over traditional 3D NAND, if such gains were this easy we'd have SSD's mainstream by now & Seagate would be bankrupt.

Are you talking performance, or GB per stack? I think you might've misunderstood the previous comments here. And obviously SSDs are mainstream already, but clearly not at the sizes of HDD and most likely never will be as long a we use NAND flash.
 
Joined
Sep 15, 2016
Messages
434 (0.24/day)
I'm not concerned with marketing.
But the engineer in me wants to know how moving the wiring from besides the cell, under the cell can improve the performance. Because, going by what was released here, that doesn't compute.
Back to the car analogy, I can move the wiring inside a car as much as I want. It won't make it go any faster.

Smaller distance between cells = faster performance.
 

BreakfastSnacks

New Member
Joined
Feb 3, 2017
Messages
1 (0.00/day)
It uses the same components only arranged differently. I don't see how that would improve performance. Like, at all.

Major performance gains throughout computing history have been achieved by arranging components differently. Look at changes in CPU architecture over the past 20 years.

Smaller distance between cells = faster performance.

This!
 

bug

Joined
May 22, 2015
Messages
9,234 (3.98/day)
Processor Intel i5-6600k (AMD Ryzen5 3600 in a box, waiting for a mobo)
Motherboard ASRock Z170 Extreme7+
Cooling Arctic Cooling Freezer i11
Memory 2x16GB DDR4 3600 G.Skill Ripjaws V (@3200)
Video Card(s) EVGA GTX 1060 SC
Storage 500GB Samsung 970 EVO, 500GB Samsung 850 EVO, 1TB Crucial MX300 and 3TB Seagate
Display(s) HP ZR24w
Case Raijintek Thetis
Audio Device(s) Audioquest Dragonfly Red :D
Power Supply Seasonic 620W M12
Mouse Logitech G502 Proteus Core
Keyboard G.Skill KM780R
Software Arch Linux + Win10
Major performance gains throughout computing history have been achieved by arranging components differently. Look at changes in CPU architecture over the past 20 years.
That was my exact question: how does moving wiring around result in better performance? It's still left unanswered (i.e. I haven't found any explanation so far).
 
Top