Yes, ansy1968, hopefully it would happen within 1-2 yrs because China won't slow down the current momentum
Chinese semiconductor industry
- Thread starter Hendrik_2000
- Start date
- Status
Hi machupicu,Yes, ansy1968, hopefully it would happen within 1-2 yrs because China won't slow down the current momentum
I'm a greedy person, I want it to happen now
hehe... But I can picture it, 2022 is the target date for mass production of 7nm, by that time TSMC will be introducing their 2nm tech, so 2 generation behind. if we can able to bridge that gap by at least 1 generation then China can force the issue and impose an import substitute program. It will force TSMC to invest in China with their latest tech with Chinese made equipment. A win win.A more detail report on SMIC 7nm breakthrough from cnTechPost
SMIC achieves breakthrough in 7nm process
2020-10-12 19:46:37 GMT+8 | cnTechPost
Chinese IP and custom chip company Innosilicon has completed the world's first chip tape-out and testing based on SMIC's FinFET N+1 process, the Zhuhai government's Zhuhai Special Zone Newspaper on October 11, adding that all of its IP is made in China and its functionality passed the test in one go.
Regarding the N+1 process, SMIC co-CEO Mong-Song Liang had revealed earlier this year that the process is very similar to the 7nm process in terms of power and stability, and does not require a EUV lithography machine.
But the performance improvement is not enough, so the N+1 process is for low-power applications, Liang said.
And tape-out success means getting devices that meet performance targets in the lab. If you want to achieve real mass production, the reliability of the device, degradation mechanism, and some other characteristics need a lot of data support and repeated testing.
Zhuhai Special Zone News reported that Innosilicon's first chip tape-out and successful testing based on SMIC's advanced process not only means that the 'domestic core' with independent intellectual property rights has once again broken the foreign monopoly but also shows that China-made 7nm chip manufacturing. The technology has been broken.
In fact, tape-out is a necessary step before a chip can be mass-produced.
In order to test the success of the IC design, the chip needs to be pre-produced to verify that the circuit has the required performance and functionality.
If the tape-out is successful, the chip can be mass-produced; if not, the cause of the problem needs to be identified and the design needs to be optimized accordingly.
Guancha.cn quoted someone from the Semiconductor Institute of the Chinese Academy of Sciences as saying: "Before the technology is delivered, it will go through a user trial phase, in different cases, and maybe tape-out will be followed by months or years before mass production can actually take place."
On September 18, SMIC had said that FinFET N+1 has entered the customer introduction phase and is expected to go into small-batch pilot production by the end of 2020.
Although SMIC has never confirmed it, there has been speculation that "N+1" is the company's 7nm process.
In February, Liang disclosed for the first time at the earnings meeting that the N+1 and 7nm processes are very similar in terms of power and stability, with the only difference in performance being that the N+1 process is less efficient.
When it comes to specific data, he revealed that SMIC's N+1 process has a 20% performance increase and 57% power consumption reduction compared to 14nm.
The 7nm process market benchmark performance improvement should be 35%, so the company's N+1 process is for low-power applications.
Previously, SMIC had ordered a EUV lithography machine for next-generation process manufacturing, worth up to $150 million, from ASML, which was scheduled to be delivered in early 2019.
But the U.S. government has blocked the deal, which has yet to be completed.
In the current plan, N+1 and N+2 processes will not use EUV equipment, until the equipment is ready, N+2 will turn to use EUV equipment.
Previously, it was reported that SMIC's 7nm process development is similar to the TSMC route, 7nm node development of a total of three processes, respectively, low-power N7, high-performance N7P, the use of EUV process N7+.
It should be noted that at present, the world can mass produce 7nm chip foundries only TSMC and Samsung.
IDM manufacturer Intel 10nm products have just recently been launched, 7nm chips may be delayed until 2023. guancha.cn reported that this means that SMIC is expected to catch up with Intel in the process.
SMIC achieves breakthrough in 7nm process
2020-10-12 19:46:37 GMT+8 | cnTechPost
Chinese IP and custom chip company Innosilicon has completed the world's first chip tape-out and testing based on SMIC's FinFET N+1 process, the Zhuhai government's Zhuhai Special Zone Newspaper on October 11, adding that all of its IP is made in China and its functionality passed the test in one go.
Regarding the N+1 process, SMIC co-CEO Mong-Song Liang had revealed earlier this year that the process is very similar to the 7nm process in terms of power and stability, and does not require a EUV lithography machine.
But the performance improvement is not enough, so the N+1 process is for low-power applications, Liang said.
And tape-out success means getting devices that meet performance targets in the lab. If you want to achieve real mass production, the reliability of the device, degradation mechanism, and some other characteristics need a lot of data support and repeated testing.
Zhuhai Special Zone News reported that Innosilicon's first chip tape-out and successful testing based on SMIC's advanced process not only means that the 'domestic core' with independent intellectual property rights has once again broken the foreign monopoly but also shows that China-made 7nm chip manufacturing. The technology has been broken.
In fact, tape-out is a necessary step before a chip can be mass-produced.
In order to test the success of the IC design, the chip needs to be pre-produced to verify that the circuit has the required performance and functionality.
If the tape-out is successful, the chip can be mass-produced; if not, the cause of the problem needs to be identified and the design needs to be optimized accordingly.
Guancha.cn quoted someone from the Semiconductor Institute of the Chinese Academy of Sciences as saying: "Before the technology is delivered, it will go through a user trial phase, in different cases, and maybe tape-out will be followed by months or years before mass production can actually take place."
On September 18, SMIC had said that FinFET N+1 has entered the customer introduction phase and is expected to go into small-batch pilot production by the end of 2020.
Although SMIC has never confirmed it, there has been speculation that "N+1" is the company's 7nm process.
In February, Liang disclosed for the first time at the earnings meeting that the N+1 and 7nm processes are very similar in terms of power and stability, with the only difference in performance being that the N+1 process is less efficient.
When it comes to specific data, he revealed that SMIC's N+1 process has a 20% performance increase and 57% power consumption reduction compared to 14nm.
The 7nm process market benchmark performance improvement should be 35%, so the company's N+1 process is for low-power applications.
Previously, SMIC had ordered a EUV lithography machine for next-generation process manufacturing, worth up to $150 million, from ASML, which was scheduled to be delivered in early 2019.
But the U.S. government has blocked the deal, which has yet to be completed.
In the current plan, N+1 and N+2 processes will not use EUV equipment, until the equipment is ready, N+2 will turn to use EUV equipment.
Previously, it was reported that SMIC's 7nm process development is similar to the TSMC route, 7nm node development of a total of three processes, respectively, low-power N7, high-performance N7P, the use of EUV process N7+.
It should be noted that at present, the world can mass produce 7nm chip foundries only TSMC and Samsung.
IDM manufacturer Intel 10nm products have just recently been launched, 7nm chips may be delayed until 2023. guancha.cn reported that this means that SMIC is expected to catch up with Intel in the process.
machupicu, WTAN,
I'm excited, the picture I dream is becoming a reality and they beat Intel by 3years, handicap without an EUVL machine, Take that Trump...
Mong-Song Liang is a true patriot and he deserve all the reward financially, If ever I have the chance to meet him in person I will bow down and kiss his feet as a sign of deep respect.
I'm excited, the picture I dream is becoming a reality and they beat Intel by 3years, handicap without an EUVL machine, Take that Trump...
Mong-Song Liang is a true patriot and he deserve all the reward financially, If ever I have the chance to meet him in person I will bow down and kiss his feet as a sign of deep respect.
Yes......i am sure their American suppliers and even ASML would have forewarned SMIC of the coming sanctions. This would have given them time to place large orders of essential parts and also supplies.
This is very much like how Huawei stockpiled Chips enough to last for years.
Personally, i think that ASML and Equipment manufacturers like LAM, KLA Tencor etc would still supply parts and supplies to SMIC but it would be done through 3rd party vendors. The Chinese Government can easily force them to supply SMIC for Pre-existing equipment or face being put on the Entitiy List, Fined or be kicked out of China.
Thats why i think SMIC will go ahead and produce its N+1 Chips and it will be busineas as usual.
SHANGHAI, Oct. 12, 2020 /PRNewswire/ --
Please join SMIC's earnings conference call, with:
Please join SMIC's earnings conference call, with:
- Dr. Zhao Haijun, Co-Chief Executive Officer and Executive Director
- Dr. Liang Mong Song, Co-Chief Executive Officer and Executive Director
- Dr. Gao Yonggang, Chief Financial Officer, Executive Director, and Joint Company Secretary
- Guo Guangli, Domestic Disclosure Representative and Senior Finance Director
as they announce the company's third quarter 2020 results and take questions from investors on Thursday, November 12, 2020.
SMIC can produce the N+1 and future N+2 with their current machine tools. But they want to expand production at these nodes and have a lot of imports basically frozen. This means they can't build new production lines.
They also import a lot of consumable materials like masks, chemicals, and machine parts. These are necessary to produce chips.
Without an EUV lithograph SMIC's N+1 and N+2 processes will take more processing steps and have less production output making their process less economically competitive vs TSMC or Samsung.
But SMIC has done a great achievement. They are basically #3 in terms of process technology in the fab business even with all these restrictions. GlobalFoundries has given up on leading edge processes altogether because they can't raise capital to develop them anymore.
SMEE DUV 28nm immersion lithography tool should be able to support these production processes as well and allow them to expand their production lines even if the sanctions on machine tools continue. However China needs EUV tools too.
They also import a lot of consumable materials like masks, chemicals, and machine parts. These are necessary to produce chips.
Without an EUV lithograph SMIC's N+1 and N+2 processes will take more processing steps and have less production output making their process less economically competitive vs TSMC or Samsung.
But SMIC has done a great achievement. They are basically #3 in terms of process technology in the fab business even with all these restrictions. GlobalFoundries has given up on leading edge processes altogether because they can't raise capital to develop them anymore.
SMEE DUV 28nm immersion lithography tool should be able to support these production processes as well and allow them to expand their production lines even if the sanctions on machine tools continue. However China needs EUV tools too.
D
Deleted member 15887
Guest
But hey, at least Huawei can still produce 5G base stations, just albeit at a lower rate.
FriedRiceNSpice
Captain
That's a very last-decade analysis of Java. More and more high performance and ultra-low latency applications are written in Java, the only times where Java would not be a good choice would be on extremely memory constrained platforms.
- Status