TSMC's 4nm process coming ahead of schedule
- ?
- Anonymous
- 3Rs
- 03 Jun 2021
Latest Chinese technolgy leads the world once again. bravo
- ?
- Anonymous
- PZa
- 03 Jun 2021
786, 02 Jun 2021I'd take big 11nm die core, which is V12 5.0 litre ove... moreThis is the dumbest thing I've read, lol. The smaller the nanometre process means you can fit more transistors into the same surface area. You can still make the same size chip as before but it will contain more features than before.
Taking your analogy, the smaller nanometre process allows you to shrink the V10 so you add two extra cylinders and turn it into a V12.
- ?
- Anonymous
- Y2T
- 03 Jun 2021
What happens after 3 nm 2nm 1nm what would they say next lol I don’t know how it works
- K
- KrazyDave
- XBA
- 03 Jun 2021
786, 03 Jun 2021It's like candy bars. They keep making them smaller at... moreNgl reading this gave me a headache but I understood nonetheless.
Though my knowledge on semiconductors is little, what I can gather is lesser fab process means higher chances of getting imperfect chips, economical as the imperfects can be used as binned chips no??
- 7
- 786
- mHu
- 03 Jun 2021
beep bop boop, 02 Jun 2021ok, i see ur point but that is a server chip, is there a wa... moreIt's like candy bars. They keep making them smaller at the same price to increase profit.
Seriosly though, there are good reasons for smaller chips. The first and foremost is that more chips can be fit onto a wafer. For large chips, the cost is all about what fraction of a wafer it uses. The cost to process a wafer is pretty much fixed, regardless of how many chips result from it.
Using less of the expensive wafer is only one part though. Yield is the other. All wafers have imperfections. Think of them as being small but randomly scattered about the wafer, and any IC that hits one of these imperfections is trash. When the wafer is covered by lots of small ICs, only a small fraction of the total are trash. As the IC size goes up the fraction of them that hit a imperfection goes up. As a unreal example that nonetheless points out the issue, consider the case where every wafer has one imperfection and is covered by one IC. The yield would be 0. If it were covered by 100 ICs, the yield would be 99%.
There's a lot more to yield than this, and this is greatly oversimplifying the issue, but these two effects do push towards smaller chips being more economical.
For really simple ICs, the packaging and testing cost dominates. In those cases, the features size is not so much a driving issue. This is also one reason we have seen a explosion of smaller and cheaper packages lately. Note that extreme small features size is being pushed by very large ICs, like main processors and GPUs.
- ?
- Anonymous
- nn8
- 03 Jun 2021
786, 02 Jun 2021It's the truth, truth hurts.Dude, cut the crap and take a chill pill!
- N
- Nikojas
- 32x
- 03 Jun 2021
What's the limit? how small van they get eventually? 1nm or less?
- ?
- Anonymous
- 6wN
- 03 Jun 2021
786, 02 Jun 2021I'd take big 11nm die core, which is V12 5.0 litre ove... moreForget the size and wait for speed test. It is the only thing that matters.
- b
- beep bop boop
- 3EM
- 02 Jun 2021
786, 02 Jun 2021The highest number of cores in a production CPU on a single... moreok, i see ur point but that is a server chip, is there a way u can explain it in context of a phone chip?
- 7
- 786
- mHu
- 02 Jun 2021
Raresmalinschi21, 02 Jun 2021Was that really necessary?It's the truth, truth hurts.
- 7
- 786
- mHu
- 02 Jun 2021
beep bop boop, 02 Jun 2021pls explain the science behind it rather than calling it a ... moreBigger die cost disproportionally more. Little processors cost $10’s, desktop CPU’s costs $100’s and server CPU’s cost $1,000’s and you can be certain the die size difference between these classes is much less than 10x.
The other reason is there is a physical limit to the printable area known as the reticle field. This is a limitation of the lithography equipment used to make the chips. The largest die sizes that can be printed are roughly 24mm x 34mm. That’s a bit over 800mm^2.
Some of the largest die commercially produced are/were IBM Power8, Nvidia GPUs, and Intel Itanium. I think Itanium wins at 700mm^2.
- R
- Raresmalinschi21
- Bke
- 02 Jun 2021
786, 02 Jun 2021I'd take big 11nm die core, which is V12 5.0 litre ove... moreWas that really necessary?
- 7
- 786
- mHu
- 02 Jun 2021
beep bop boop, 02 Jun 2021pls explain the science behind it rather than calling it a ... moreThe highest number of cores in a production CPU on a single die is 24 as far as I know. The processor in question is the Intel Xeon E7–8890 V4 (Intel® Xeon® Processor E7-8890 v4):
It’s hyperthreaded so that means the processor has 48 threads.
It has 60MB cache and supports up to 3.07 TB RAM.
Since it’s a server processor, you can have multiple processors connected to the same motherboard (in this case, the magic number is 8 which gives us a total of 192 cores and 384 threads).
The processor has a 165W TDP so it will need some specialized cooling equipment for proper functioning.
The die size is 456.12 mm².
The processor is priced at ~$7200.
Because the bigger the die NM, the more expensive it is.
- b
- beep bop boop
- 3EM
- 02 Jun 2021
786, 02 Jun 2021I'd take big 11nm die core, which is V12 5.0 litre ove... morepls explain the science behind it rather than calling it a supercar
- 7
- 786
- mHu
- 02 Jun 2021
I'd take big 11nm die core, which is V12 5.0 litre over a maggot 4nm 2.0litre V6 .
It's all about power torque pulling power, not too save petrol or be efficiency..
Ferrari Lamborghini porsche Porsche owners don't give rats about efficiency or too save.
4nm is for the poors.