Samsung is reportedly trialing the production of 3nm foundry processing for chips enabling bitcoin miners to operate faster and with less power consumption.
Samsung is reportedly in a trial production stage for 3nm foundry processing which could be used for mining bitcoin.The first customer will be PanSemi, a Chinese ASIC manufacturer and Qualcomm may also take part in the future.3nm foundry processing can enable chips to have reduced power consumption, increased speed, and higher quantities of transistors.
Samsung will reportedly start trial production this week of three nanometer (3nm) chips for application-specific integrated circuits (ASICs) – the most efficient machines for mining bitcoin.
Samsung’s first customer, according to the report, is a Chinese ASIC company known as PanSemi, which designs ASICs used for mining bitcoin. Similarly, Qualcomm – Samsung’s largest customer – has also made reservations to take advantage of the new manufacturing process with sources reportedly stating Qualcomm can opt in at any point, but is not committed.
Previously, Qualcomm had placed orders for 4nm chips, but were canceled this past February due to a surprising lack of production from Samsung. This caused Qualcomm to lean on another company – Taiwan Semiconductor Manufacturing Company (TSMC).
The latest offering from Samsung is being referred to as gate-all-around (GAA) which, as the name implies, will have gates on all four surfaces. Until now, the most commercially successful process is FinFET, which has only utilized three surfaces instead of four. This upgrade reportedly allows the gates to be narrower and allows more precise control of currents. Reports suggest this could cause a 45% area reduction and 30% increase in efficiency if the trial production is successful.
In addition, Tech Monitor reported that TSMC’s 3nm process would shrink the size of the semiconductor which, in turn, would allow up to 30% reduced power consumption, up to 15% speed increases, while also allowing 33% more transistor density – which makes the hardware more powerful.
Even though that report was from last year, it is still valuable to understand the possible impacts this advancement will have on the technology.
Samsung is reportedly trialing the production of 3nm foundry processing for chips enabling bitcoin miners to operate faster and with less power consumption.
Samsung will reportedly start trial production this week of three nanometer (3nm) chips for application-specific integrated circuits (ASICs) – the most efficient machines for mining bitcoin.
Samsung’s first customer, according to the report, is a Chinese ASIC company known as PanSemi, which designs ASICs used for mining bitcoin. Similarly, Qualcomm – Samsung’s largest customer – has also made reservations to take advantage of the new manufacturing process with sources reportedly stating Qualcomm can opt in at any point, but is not committed.
Previously, Qualcomm had placed orders for 4nm chips, but were canceled this past February due to a surprising lack of production from Samsung. This caused Qualcomm to lean on another company – Taiwan Semiconductor Manufacturing Company (TSMC).
The latest offering from Samsung is being referred to as gate-all-around (GAA) which, as the name implies, will have gates on all four surfaces. Until now, the most commercially successful process is FinFET, which has only utilized three surfaces instead of four. This upgrade reportedly allows the gates to be narrower and allows more precise control of currents. Reports suggest this could cause a 45% area reduction and 30% increase in efficiency if the trial production is successful.
In addition, Tech Monitor reported that TSMC’s 3nm process would shrink the size of the semiconductor which, in turn, would allow up to 30% reduced power consumption, up to 15% speed increases, while also allowing 33% more transistor density – which makes the hardware more powerful.
Even though that report was from last year, it is still valuable to understand the possible impacts this advancement will have on the technology.
Feedzy
