01 英特尔:不再按晶体管尺寸命名先进芯片工艺
在出口强劲和消费反弹的推动下,今年第二季度韩国经济增速创10年来新高,但这一乐观势头被一波新冠疫情蒙上阴影。
Photo credit:FTChinese
全文共511个词,by Richard Waters in San Francisco
Intel has come up with a new weapon in its attempt to regain the edge in advanced chipmaking lost to Asian rivals: change the way it names its most advanced technology.
英特尔(Intel)在先进芯片制造领域已被亚洲竞争对手反超,试图重新夺回这一领域优势的该公司如今祭出了一件新武器:改变其最先进技术的命名方式。
The US chipmaker said on Monday that it would no longer refer to its latest generation of manufacturing technology based on the size of the transistors on its semiconductors. For decades this has been considered the gold standard in advanced chipmaking, as companies raced to pack ever-larger numbers of transistors on their chips.
这家美国芯片制造商周一表示,将不再以半导体晶体管尺寸命名其最新一代芯片制造技术。几十年来,这一直被认为是先进芯片制造的黄金标准,各公司争相在芯片上封装更多晶体管。
Photo credit:Getty Images
The shift follows Intel’s loss of global leadership in manufacturing to TSMC and Samsung, ending its decades-long advantage of being the first to reach each new process step, or“node”.
英特尔在芯片制造领域的全球领先地位已被台积电(TSMC)和三星(Samsung)夺走,结束了其数十年来总是最先实现每一个新工艺“节点”的优势。
The new naming convention drops any reference to size. It means that Intel will no longer refer to the delayed “7 nanometre” chips that have caused its stumble, which will not appear until 2023.
新的命名规则不再按照晶体管尺寸。这意味着英特尔将不再提及因延迟推出导致其陷入困境的“7纳米”芯片。英特尔要到2023年才能推出该芯片。
Instead, it will use the new brand Intel7 to refer to an intermediate generation of chips that are set to hit the market a year earlier.
相反,英特尔将使用新品牌“Intel 7”指代将在2022年上市的中间一代芯片。
Photo credit: Getty Images
Sanjay Natarajan, an Intel senior vice-president, denied that the change was an attempt to obscure Intel ’ s manufacturing issues. The company was enacting the shift to ease comparisons with rivals for customers as it moves into the foundry market, where it makes chips for other companies, he said.
英特尔高级副总裁桑贾伊·纳塔拉詹(Sanjay Natarajan)否认这一改变是为了掩盖英特尔的制造问题。他表示,随着该公司进入代工市场,为其他公司生产芯片,采取这一举措是为了方便客户将其与竞争对手进行比较。
Some analysts also rejected suggestions that Intel was trying to hide its problems, and said the move was a belated recognition that the old way of describing advances in chipmaking was no longer relevant.
一些分析师也驳斥了英特尔意在掩盖问题的说法,并表示此举是一种迟来的承认,即用来描述芯片制造进步的旧方式已不再适用。
Although rivals still refer to their products in terms of size — TSMC calls its most advanced chips 5nm — the term does not have much meaning as other aspects of chipmaking have come to the fore, said Patrick Moorhead, a US chip analyst.
美国芯片分析师帕特里克·穆尔黑德(Patrick Moorhead)表示,尽管其竞争对手仍然用晶体管尺寸命名自己的产品——台积电将其最先进的芯片称为5纳米——但就芯片制造技术而言,随着其他方面的重要性日益突显,这个术语已没有多少意义。
The change came as Intel on Monday laid out a set of technologies and manufacturing targets for the next three years that it hopes will enable it to catch up with TSMC and Samsung. They included new ways to package different parts of a chip into a single semiconductor—a technology that Pat Gelsinger, Intel’s new chief executive, is counting on to put the company back in front.
周一,英特尔公布了未来三年的一系列技术和制造目标,希望通过达到这些目标来追赶上台积电和三星。其中包括采用新方法将芯片的不同部分封装到单个半导体上——英特尔新任首席执行官帕特·盖尔辛格(Pat Gelsinger)正指望凭借这种新技术让公司重回领先地位。
Photo credit:Getty Images
Moorhead said that Intel’s most advanced chips already matched TSMC in performance per watt of power consumed, one of the main measures, but were lagging behind in density,or the number of transistors it can cram on to a chip — a critical measure for uses such as smartphones.
穆尔黑德表示,英特尔最先进的芯片在每瓦功耗性能方面——主要衡量指标之一——已经与台积电不相上下,但在密度方面——即一块芯片上可容纳的晶体管数量,智能手机等设备所使用的芯片的关键指标——落后于台积电。
Intel’s old naming system made the company look two years behind TSMC when in fact it is only about six months behind, said Dan Hutcheson, head of VLSI Research.
VLSI Research主管丹·哈奇森(Dan Hutcheson)表示,英特尔的旧命名规则让该公司看起来落后台积电两年,而实际上只落后大约6个月。
“If they hit all their dates, they have the potential to be ahead in 2024—but that’s a big‘if’,” Moorhead said.
穆尔黑德说:“如果他们如期实现所有目标,他们就有可能于2024年重回领先地位——但这是一个很大的‘如果’。”