TSMC's 4nm process coming ahead of schedule

02 June 2021
The 3nm node's development is going smoothly too.

Sort by:

  • m
  • madsam
  • tet
  • 04 Jun 2021

TSMC is really turning it on. Where are those other big players like Samasung, Intel and Global foundry?

    • S
    • Say What
    • 7$r
    • 03 Jun 2021

    It’s all good for the consumer. Technological progress, companies pushing forward whether they are innovating or introducing tech to new consumers.
    You can tell 786 is a fanboy kid with psychological issues. The silly talk comparison to motor engines shows you that and then cut and paste from who knows where is plagiarism because they don’t disclose their sources and references

      Whats with all the copypasta

        • 7
        • 786
        • mHu
        • 03 Jun 2021

        Anonymous, 03 Jun 2021This is the dumbest thing I've read, lol. The smaller ... moreThere are several types of gate-all-around technologies, including nanosheet FETs and nanowire FETs. Gate-all-around itself is an evolutionary step from the finFET. In gate-all-around, a finFET is placed on its side and is then divided into separate horizontal pieces. Each separate piece makes up the channels. A gate material wraps around each sheet.

        Compared to the nanowire FET, the nanosheet FET has a wider channel, which translates into more performance and drive current. “The nanosheet has a larger effective width,” “The nanowire is very good for the electrostatics. But the cross section is rather small. That will not bring an advantage for the effective channel width.”

          • 7
          • 786
          • mHu
          • 03 Jun 2021

          YUKI93, 03 Jun 2021I wonder which one between Qualcomm, Mediatek, and Apple wi... moreSamsung is the only company that has announced its 3nm plans so far. For that node, the foundry will move to a new gate-all-around technology called the nanosheet. TSMC has yet to disclose its plans, leaving some to believe that it’s behind the curve. “At 3nm, Samsung has a high probability of initial high-volume production in 2021,” Jones said. “TSMC is accelerating the development to try to close the gap with Samsung.”

            • 7
            • 786
            • mHu
            • 03 Jun 2021

            OhNom, 03 Jun 2021Wow wow friend, I really have to congratulate you, I did no... moreThe issue with making larger chips isn't what we call "impurities". The main problems are physical defects caused by lithography/etching issues and particulates (which I guess you could call "impurities", but no one does) and variation in transistor parameters across the chip. These can cause parts to have failing circuits in them, to where they need to be either downgraded or rejected entirely. Both defects and variation get worse as the chip gets larger, in an exponential fashion.

            And the timing issues of larger vs smaller chips is mostly due to the aforementioned transistor variation (and interconnect variation) as you put more stuff on one chip, not due to the larger distances that signals need to propagate. That is a design concern, but not really a manufacturing yield issue.

            I would also take issue with your last paragraph. Having CPUs and GPUs that can run at 80C+ is quite common; you can safely run them much hotter. The problems aren't so much safety issues as reliability and design constraints.

              • 7
              • 786
              • mHu
              • 03 Jun 2021

              TheLastOracle, 03 Jun 2021Didn't realise Shakespeare had been resurrected. You E... moreLots of reasons.

              As was pointed out, the bigger they are the more power they consume.

              But there's more to it than that. Cooling is hard - really hard. There are low power CPU's and high power CPUs, typically the top range part you can cool with fans and air is about 250 watts (those are GPU's or new AMD 9950 or something along those lines). The top range desktop intel CPU's have been 130-140 Watts for at least a decade.

              Much more power than that any you get into all sorts of problems - not the least of which is the risk if fire or injury to someone trying to work on your equipment. Remember you've basically got a 150W light bulb in an area the size of a quarter. Trying to move the heat away from that is really quite difficult while not also having your computer sound like a jet engine.

              There are also manufacturing difficulties. CPU's are made out of wafers of silicon - big wafers (different companies use different sizes etc.). The larger the CPU the more likely it is to have an impurity in it, making all or part of it not work - parts with impurities that don't work now are usually lower tier parts. People have figured out a good size to make the CPU's out of big wafers to fit a lot of CPU's (or GPU's) out of one wafer, while still making sure most of them work at close to full potential. Small chips also have less timing issues (it does take time for electricity to travel the length of a CPU) eventually we're going to have to deal with this problem as CPU's become fast enough that that timing becomes a problem.

              Making pure silicon wafers is EXTREMELY difficult. Air contaminates them. The glass in the room contaminates them. The floor contaminates them. The clothes people wear when they go into handle them contaminates them.

                • D
                • AnonD-762416
                • m2A
                • 03 Jun 2021

                mixedfish, 03 Jun 2021Yes and you can tell that's what's happening. Nvi... moreYeah, the RTX x070 series used to cost around 5-600 bucks, now they're closing in on 1,500 bucks. Over one generation. It's lies and scams, and we the consumers are the ones picking up the tab, as always.

                  • 7
                  • 786
                  • mHu
                  • 03 Jun 2021

                  YUKI93, 03 Jun 2021Lol dude, electronic engineering doesn't work the same... moreI'd take a 11nm at much higher and more power torque with more 12mb L2cache, with 96mb bandwidth than you're maggot 5nm v6 2.0 potsy vs my 5.0 litre V12 would trash and destroy yours.
                  1x3.84 GHz Kryo 680 & 3x3.42 GHz Kryo 680 & 4x2.80 GHz Kryo 680) -

                  As the process size gets smaller, power usage decreases.

                  Smaller transistor processes allow the use of lower voltages combined with the improvements in construction technique mean that a ~45nm processor can use less than half the power that a 90nm processor uses with similar transistor counts.

                  The reason for this is that as the transistor gate gets smaller, threshold voltage and gate capacitance (required drive current) gets lower.

                  It should be noted that as Olin pointed out this level of improvement doesn't continue to smaller process sizes as leakage current becomes very important.

                  One of your other points, the speed at which signals can travel around the chip:

                  At 3ghz the wavelength is 10cm, however the 1/10th wavelength is 1cm which is where you need to start considering transmission line effects for digital signals. Additionally remember that in the case of Intel processors some parts of the chip runs at twice the clock speed so 0.5cm becomes the important distance for transmission line effects. NOTE: they may be operating on both clock edges in this case, meaning the clock doesn't run at 6Ghz but some processes going on are moving data that fast and have to consider the effects.

                  Outside transmission line effects, you also have to consider clock synchronization. I don't actually know what the propagation velocity is inside a microprocessor, for unshielded copper wire its like 95% of the speed of light but for coax is like 60% the speed of light.

                  At 6Ghz the clock period is only 167 picoseconds the so high/low time is ~ 84 picoseconds. In vacuum, light can travel 1cm in 33.3 picosends. If the propagation velocity was 50% the speed of light then its more like 66.6 picoseconds to travel 1 cm. This combined with the propagation delays of the transistors and possibly other components means that the time the signal takes to move around even a small die at 3-6Ghz is significant for maintaining proper clock synchronization.

                    • 7
                    • 786
                    • mHu
                    • 03 Jun 2021

                    YUKI93, 03 Jun 2021Lol dude, electronic engineering doesn't work the same... moreTell that too samsung and every other OEMS, too start using liquid cool nitrogen copper cooling underneath and over the Cpu. They are expensive, no wonder the Note20Ultra didn't use any.
                    Only Sony has good above average cooling copper materials.
                    Instead off blaming OEMS, you claim that smaller die cores are better, when you can't fit V8 in a bonnet but have too justify the size too use V6 too cut costs. Yea really, lmfao. It's using 10W off power then use 6000mah battery to compensate. People won't complain if the phone size reaches too 9mm thick. Look at camera bump, it's getting bigger , look at Lumia 900 by Nokia, Nokia 808, no one complain about that huge carlZeiss optics with Xenon flash.
                    You're thinking and ideas are average, you don't like premium or prestige products, since Asia is poor, they think poor vs European USA UK middleeast.

                      786, 02 Jun 2021I'd take big 11nm die core, which is V12 5.0 litre ove... moreLol dude, electronic engineering doesn't work the same as mechanical engineering. The lower the nm value, the more powerful the performance. It's a starking contrast to mechanical engineering where bigger is better.

                        I wonder which one between Qualcomm, Mediatek, and Apple will be the first to get the new 4nm chipset.

                          • m
                          • mixedfish
                          • YUB
                          • 03 Jun 2021

                          AnonD-762416, 03 Jun 2021That is not what's been happening though. It's ma... moreYes and you can tell that's what's happening. Nvidia out of nowhere is releasing brand new stock of obsolete graphics cards. If there was a supply issue then how would they be able to retool machines to have the capacity to make larger die cards. Clearly they're interested on being a cartel and pushing for higher margins as they bludgeon production of low yield chips.

                            • D
                            • AnonD-762416
                            • m2A
                            • 03 Jun 2021

                            Soo, 03 Jun 2021So, im guessing the most 4nm will go again first for apple ... moreThat is not what's been happening though. It's maybe a convenient excuse to hide fabrication issues, but it's not connected to reality.
                            AMD, Nvidia, Qualcomm and all the rest have "always" relied on that fab, now they can't. It's clearly a terrible yield problem that they're trying to get past by rushing the next one out.

                              • S
                              • Soo
                              • nDy
                              • 03 Jun 2021

                              So, im guessing the most 4nm will go again first for apple devices and the rest of us must wait for second wave of 4nm soc

                                • A
                                • Aha
                                • nDy
                                • 03 Jun 2021

                                786, 02 Jun 2021I'd take big 11nm die core, which is V12 5.0 litre ove... moreThats why an 2 l charged has more performance and more torque that a v12 ?!
                                And lets not forget about the electric hypercars
                                The R&D for the v12/w12 are way down and on its way out
                                So, no, scaling when you have 4nm is more important than keeping the same perf because of the heat on a 14nm
                                I guess im talking with the wrong guy since i bet you cant understand
                                Bye and stop misleading people

                                  786, 02 Jun 2021I'd take big 11nm die core, which is V12 5.0 litre ove... moreDidn't realise Shakespeare had been resurrected. You English is worthy of Knighthood.

                                    • ?
                                    • Anonymous
                                    • tV4
                                    • 03 Jun 2021

                                    4nm for sd8955 next year flagship?

                                      • ?
                                      • Anonymous
                                      • 8pc
                                      • 03 Jun 2021

                                      Anonymous, 03 Jun 2021Latest Chinese technolgy leads the world once again. bravoTSMC is taiwanese not chinese. So this is taiwanese innovation.

                                        786, 02 Jun 2021I'd take big 11nm die core, which is V12 5.0 litre ove... moreWow wow friend, I really have to congratulate you, I did not think that someone who entered GSMARENA, could be so ignorant, with you I take a surprise that cannot be overcome in a long time