It’s earnings call season, and we have heard of a slowing DRAM market and NAND flash price declines from Micron, SK hynix, Intel, and now Samsung. DRAM prices have stopped increasing, and that can be viewed as a precursor to a price decline.
Samsung’s 31 October, 2018 3Q18 earnings call vindicated Objective Analysis‘ forecast for a 2H18 downturn in memories that will take the rest of the semiconductor market with it.
Those familiar with our forecast know that for a few years we have been predicting a downturn in the second half of this year as NAND flash prices fall, followed by a DRAM price collapse. After the DRAM collapse the rest of the semiconductor market will undergo a downturn.
We’ve been calling for this downturn for some time. Dan Hutcheson at VLSI Research has been videotaping our forecast every December for the past Continue reading “Memory Market Falling, as Predicted”
Many readers have probably wondered why NAND flash fabs are so enormous. Although DRAM fabs used to be the largest, running around 60,000 wafers per month, NAND flash fabs now put that number to shame, running anywhere from 100,000-300,000 wafers per month. Why are they so huge?
The reason is that you need to run that many wafers to reach the optimum equipment balance. The equipment must be balanced or some of it will be sitting idle, and with some tools costing $50 million (immersion scanners) you want to minimize their idle time to the smallest possible number. I am sure that this is a tough problem, although I have never had to solve it myself.
The most important reason that so much attention is focused on this is that the cost of the wafer depends on the efficiency of the fab. If you built a $13 billion NAND flash fab that produced 90,000 wafers per month instead of 100,000 wafers per month, then the amount of investment per wafer would be 10% higher. That can make a significant difference to Continue reading “Why are NAND Flash Fabs so Huge?”
Conventional wisdom holds that SSDs will someday displace all HDDs, but in reality SSDs are proving to be more of a challenge to the DRAM market than to the HDD market.
Right now you are probably reviewing the date of this post to make sure it’s not dated April 1. I assure you that this is the truth. To understand it, though, you must look at a computer as a computer architect would, or, in other words, the way that an application program sees the memory/storage hierarchy.
To the application program there is no HDD and memory, there is only memory. The Virtual Memory system, a part of the operating system, hides the difference between the two by moving code and data into DRAM as it is needed and back onto the HDD when it is no longer important, without telling the application program that it is moving anything around. I like to tell people that the DRAM makes the HDD look fast, and the HDD makes the DRAM look big.
If you think of the DRAM as something that makes the HDD look fast, then additional DRAM should help to make the Continue reading “Why DRAM is Threatened by SSDs”
In an interesting twist to today’s ongoing DRAM shortage, the Fuzhou Intermediate People’s Court, Fujian Province, China today granted a preliminary injunction to prevent Micron’s Chinese subsidiaries from manufacturing, selling, or importing certain DRAM modules and solid state drives in China.
This injunction, according to a Micron press release, was filed without allowing Micron to present its defense, a process which Micron finds to be: “inconsistent with providing a fair hearing through appropriate legal processes and procedures.”
Micron’s customers in China will find that the DRAM shortage has just become even worse than it already was. Before today China’s government was concerned enough about the shortage’s rising DRAM prices to have launched a price fixing investigation only one month ago. One result of today’s decision will be that there will be less DRAM in China, and that will probably cause prices to rise even more.
What will be the impact to Micron? I find it unlikely that this injunction is likely to change any DRAM maker’s business much during a shortage. Any lack of Micron DRAM in China is likely to be serviced by Samsung and SK hynix, but since there’s a shortage, these companies will need to reduce their shipments outside of China to satisfy Continue reading “How to Worsen a DRAM Shortage”
The Memory Guy has found that some people get confused about the terminology surrounding flash “Layers” and “Levels,” Sometimes confusing the two, and often misunderstanding what each one means. This post is meant to be a low-level primer to address that confusion.
There are actually three places where such terminology is used: The number of chips in a package, the number of conductor/insulator pairs in 3D NAND, and the number of voltage levels stored on any single bit cell within the chip. I will address them in that order.
CHIP STACKING: Since the 1990s Both NAND and NOR flash chip makers have been stacking chips within a single plastic package. Originally this approach was used to reduce the size of thin flip phones like the Motorola Razr by stacking an SRAM chip on top of NOR flash, but soon afterwards NAND chips began to use the same approach to get incredible storage capacities into a single IC package or eMMC, or into a microSD card format. What began as 2-die stacks became 4, then 8, and now 16 high. This post’s photo illustrates an 8-high stack.
Since the height of a standard plastic package for a chip is smaller than a stack of 16 full-thickness dice the wafers had to be Continue reading “NAND Flash’s Layers of Layers of Layers”
3D NAND presents an interesting conundrum. To improve bit costs and continue along the path of Moore’s Law the layer count must increase. Unfortunately 3D NAND can’t benefit from lithographic scaling; it’s pretty much stuck at 40nm design rules forever. The natural way to reduce costs and increase chip density is by adding layers.
But adding layers increases the size of the staircase structure used to access the wordline layers.
With today’s structures, the addition of layers means adding stairs to the staircase – if you double the number of layers then the amount of die area required by the staircase doubles. At some point the staircase becomes so large that the die has fewer GB/mm² than a die with half as many layers.
An example of a staircase structure can be seen in the Continue reading “Solving 3D NAND’s Staircase Problem”
Recently I have been hearing concerns that an impending wafer shortage might drive today’s DRAM and NAND flash shortages to epic proportions.
The Memory Guy doesn’t pretend to have any understanding of the raw wafer business, so I decided to consult Mark Thirsk, managing partner of Linx Consulting. Mark has been in this industry for quite a while and has a very good understanding of the ongoing status of the semiconductor materials supply chain.
Mark and I were on a panel together at SEMICON Korea in February, and he presented an interesting chart to compare the costs of different technologies. I asked him about this chart as well.
Here’s what Mark had to say:
“Our information is that major Continue reading “Wafer Shortages and DRAM/NAND”
Chip reverse-engineering consultant Dick James pointed The Memory Guy to an absolutely amazing 25-second video of a 3D NAND chip. The video’s made by the Carl Zeiss company. It’s the second one from the top on this page: https://www.zeiss.com/semiconductor-manufacturing-technology/products-solutions/process-control-solutions/crossbeam-fib-sem.html
The video zooms around a portion of a 3D NAND die as layers are etched away and then restored. Only the tungsten parts of the chip are shown, with the rest appearing to be empty space. This serves to clarify it a good bit. Dick James says that this makes it the equivalent of a 3D x-ray tomograph.
It’s a promotional piece for a Zeiss tool called the “Crossbeam FIB-SEM” that can both image and mill a chip.
Now I doubt that most Memory Guy readers would have a need for this tool, nor be able to afford something which is doubtlessly very expensive, but I am sure that anyone would admire what it can do. I certainly find it to be impressive!
Naturally, Dick James was able to identify the chip just by looking at it. He says that it’s Samsung’s 32-layer part.
It came as a surprise to the Memory Guy on Monday to receive a press release from Micron indicating that Intel and Micron had decided to end their NAND flash partnership.
This agreement, which was begun in 2006, helped the two companies to aggressively ramp into the NAND flash market by combining their resources. NAND flash makers (as well as DRAM makers) need to make very substantial capital investments to participate in the market, and that’s not easy for a new entrant. Micron at that time was a very small NAND flash maker, and Intel wasn’t involved in the NAND flash market at all, so neither was in a position to succeed. By combining their resources the companies were able to become important contributors to the market.
The agreement initially appeared to be modeled after the very successful joint venture that Toshiba and SanDisk enjoyed. Each company would contribute half of the JV’s capital investment, and the same designs would be used to make both companies’ chips.
Over time Intel found itself in a familiar Continue reading “Micron and Intel to End NAND Flash JV”
Error Correction Codes, ECC, are not only important to today’s NAND flash market, but they have been a cause of concern to NAND users for a number of years. The Memory Guy has been intending for some time to write a low-level primer on ECC, and I am finally getting it done!
Why is ECC necessary on NAND flash, yet it’s not used for other memory technologies? The simple answer is that NAND’s purpose is to be the absolute cheapest memory on the market, and one way to achieve the lowest-possible cost is to relax the standards for data integrity — to allow bit errors every so often. This technique has been used for a long time in both communications channels and in hard disk drives. Data communication systems can transfer more data using less bandwidth and a weaker signal over longer distances if they use error correction to restore distorted data. Hard disk drives can pack more bits onto a platter if the bits don’t all have to work right. These markets (and probably certain others) have invested a lot of money in ECC research and development, and as a result ECC today is a very well-developed science.
Denali Software published a nice Continue reading “How 3D NAND Shrinks ECC Requirements”