News
 

Bookmark and Share

(17) 

Taiwan Semiconductor Manufacturing Co., the world’s largest contract maker of chips, said on Thursday that it had initiated mass production of chips using 20nm process technology, about a month ahead of its plan. It is unknown which chips are now in production, but it is known that the new process technology will allow to make very complex SoCs with increased performance and low power consumption.

“We have two fabs, fab 12 and fab 14 that complete the core of the 20nm-SoC. As a matter of fact, we have started production. We are in the [high]-volume [20nm] production as we speak right now,” said C. C. Wei, co-chief executive officer and co-president of TSMC, during a conference call with investors and financial analysts.

TSMC’s CLN20SOC technology is designed for various highly-integrated system-on-chips (SoC) that benefit from high transistor density. Thanks to use of high-k metal gate dielectrics, chips made using the new process will be able to achieve fairly high frequencies without significant increase of leakage currents. TSMC offers only one version of process technology at 20nm node.

Earlier TSMC indicated that the first five products to be made using 20nm fabrication process were designed for mobile computing, CPU and PLD [programmable logic device] industries.

At present TSMC is producing chips using 20nm fabrication process at select modules of fabs 12 and 14. Starting from May, the company will initiate 20nm production at the fab 15 modules 3 and 4.

The world’s largest contract maker of semiconductors predicts that wafers processed using 20nm process technologies will account for around 10% of its 2014 total wafer revenue. In Q4 the 20nm manufacturing technology is expected to account for 20% of TSMC revenue.

TSMC believes that it will be extremely competitive with 20nm process technology thanks high-volume ramp at multiple fabs and engagements with tens of clients.

Tags: TSMC, Semiconductor, 20nm, HKMG

Discussion

Comments currently: 17
Discussion started: 01/17/14 01:11:25 AM
Latest comment: 01/22/14 06:00:22 AM
Expand all threads | Collapse all threads

[1-6]

1. 
So it is the evolution of the current 28nm SHP process used by AMD Kaveri. It looks like Excavator HSA compute cores will be fabbed on time for release in 12 months.
5 3 [Posted by: linuxlowdown  | Date: 01/17/14 01:11:25 AM]
Reply
- collapse thread

 
I'm not sure if Excavator is still on schedule - not from GloFo but from AMD.
Also: is it known yet, if Excavator is a completely new Architecture (doubtfully) or Bulldozer-Family based. If the latter, I will not be looking forward to it. As long as AMD doesn't reach an IPC and efficiency CLOSE to Intel (I'd also be fine with reaching 2008 core i7 levels of IPC per clock).
Since 2008 I am waiting for a very performant tech from AMD CPUs. Kombining with HSA and the GPU, + lots of DDR4 this would be the CPU of choice for my Family Office/ HTPC Computer.
So I really hope for something great here. Or that the software will finally catch up, using HSA and OpenCL like I hope it to do for several years now. Also the games, using the iGPU of AMD and Intel CPUs for AI or physics.
But not really much playable techdemos around at this point of time, there will be years before we see HSA/OpenCL or our Intel/AMD CPUs with iGPU in action to really accelerate games again (I am waiting for this, since the slow performance improvements of x86 performance in desktop class CPUs is annoying, but if games start to use the iGPU for some extra work like AI, physics... this would finally make use of some unused CPU/GPU aerea
3 0 [Posted by: Rollora  | Date: 01/17/14 05:26:16 AM]
Reply
 
Kaveri's tricks are not fully known or understood. For example, AMD intends for when you combine it with a R9 graphics card, it will then act as a graphics co-processor. We are also yet to see Mantle in action. This is very significant tech. I reckon Intel "influenced" tech sites, such as Anandtech, have taken advantage of Kaveri's "stepped" launch to gun it down before it can be truly judged on its merits.
7 5 [Posted by: linuxlowdown  | Date: 01/17/14 07:07:02 AM]
Reply
 
show the post
2 8 [Posted by: Tristan  | Date: 01/17/14 09:54:03 AM]
Reply
 
***** Silenced by Fanboyslayer *****
4 2 [Posted by: fanboyslayer  | Date: 01/18/14 02:18:36 AM]
Reply
 
show the post
4 9 [Posted by: Tristan  | Date: 01/17/14 08:36:07 AM]
Reply

2. 
It's great to see that TSMC now enjoys the technological superiority over Intel's FABs and now AMD, Nvidia, Qualcomm along with others have access to this technology.

I don't remember the last time a company got ahead of Intel. Maybe it was AMD back in the 0.18 micron days, when they were the first to use copper interconnects along with the smaller transistors.

It's god to see that a large part of the semiconductor industry now can really compete against Intel.
6 4 [Posted by: East17  | Date: 01/17/14 06:46:25 AM]
Reply
- collapse thread

 
I thought Intel is going to release Broadwell, it's 14nm CPU later this year and Ivy Bridge is on 22nm since early 2012. So I don't know what you're talking about?
3 1 [Posted by: Memristor  | Date: 01/17/14 08:54:34 PM]
Reply
 
I'm talking about the fact that Intel doesn't have a single CPU that's currently being produced in volume, using a technology smaller than 22nm.

Therefore, if TSMC's clients manage to launch 20nm chips ahead of Intel's 14nm products, they will actually have a net technological manufacturing advantage.
0 0 [Posted by: East17  | Date: 01/22/14 06:00:22 AM]
Reply

3. 
Each smaller node from 32nm brings minute benefits - mainly lower power consumption, not increased frequency. That is why there is no great rush to >28nm. Don't expect a whole lot of performance gains from here on out as all of the big gains have been had.
7 3 [Posted by: beenthere  | Date: 01/17/14 09:07:19 AM]
Reply
- collapse thread

 
Partly true. Performance gains will be made by utilising mulitcore designs. In AMD's case, HSA is superior to other multicore designs because of the seamless integration of CPU/GPU and the fact GPU compute units pack way more parallelism than CPU. The only way Intel will be able to compete in the medium term is to come up with their own HSA-like architecture.... and that can't be done overnight.
3 1 [Posted by: linuxlowdown  | Date: 01/18/14 02:28:04 AM]
Reply
 
The easier way is forcing OEM's not to go AMD. Also bribing software company's not to adopt HSA,Mantle or any other AMD tech. I'm sure INTEL is already taken these easy route, as always.
1 0 [Posted by: nader_21007@yahoo.com  | Date: 01/20/14 04:28:01 PM]
Reply

4. 
So... Who, from AMD or NVidia, will be first to market with a 20nm videocard?
5 0 [Posted by: BernardP  | Date: 01/17/14 11:09:55 AM]
Reply
- collapse thread

 
Why does it matter? It is not always advantageous to be "first". It's more important to both of them to keep yield efficiencies high, regardless of the process technology.
0 0 [Posted by: DivideOverflow  | Date: 01/21/14 01:51:12 PM]
Reply

5. 
The continued scalability issues with respect to process node shrinks and their Thermal as well other, Laws of physics, issues will only continue to worsen, moving towards the current silicon based process technology coming to an end, and other much more expensive non silicon process technologies, if the economics of the new markets will allow, begin to take over! Even Intel will be hard pressed with its declining unit sales, and with its declining profit margins per unit sold, to be able to amortize the development costs associated with the new beyond the silicon process technologies! It is going to cost more money to utilize the currently few remaining die process shrinks of the available silicon process also!
5 1 [Posted by: BigChiefRunAmok  | Date: 01/17/14 01:09:59 PM]
Reply

6. 
I'm just waiting for their next announcement: "we're having unexpected yield issues".

This is also 20nm planar, so not very impressive when Intel is already ramping 14nm trigate.
0 0 [Posted by: AnonymousGuy  | Date: 01/21/14 01:39:17 PM]
Reply
- collapse thread

 
Diminishing returns will make the difference from 20nm to 14nm nowhere near as impressive as the earlier steps.
0 0 [Posted by: DivideOverflow  | Date: 01/21/14 01:54:24 PM]
Reply

[1-6]

Add your Comment




Related news

Latest News

Sunday, August 24, 2014

6:12 pm | Former X-Bit Labs Editor Aims to Wed Tabletop Games with Mobile Platforms. Game Master Wants to Become a New World of Warcraft

Thursday, August 21, 2014

10:59 pm | Khronos Group to Follow DirectX 12 with Cross-Platform Low-Level API. Khronos Unveils Next-Generation OpenGL Initiative

10:33 pm | Avexir Readies 3.40GHz DDR4 Memory Modules. DDR4 Could Hit 3.40GHz This Year

12:10 pm | AMD to Lower Prices of A-Series APUs for Back-to-School Season. New Prices of AMD A-Series APUs Revealed

Wednesday, August 20, 2014

10:53 am | AMD to Cut Prices on FX-9000, Other FX Processors: New Prices Revealed. AMD to Make FX Chips More Affordable, Discontinue Low-End Models

10:32 am | LG to Introduce World’s First Curved 21:9 Ultra-Wide Display. LG Brings Curved Displays to Gamers, Professionals

9:59 am | AMD Readies FX-8370, FX-8370E Microprocessors. AMD Preps Two New “Mainstream” FX Chips