GlobalFoundries’ 14nm-finfet-proced lwa provides no reduction

GlobalFoundries’ new 14nm-proced lwa with finfet transistors will be that size no advantages compared to the 20nm-bulk-cmos-proced lwa. It is the first time that a node-reduction no cost reduction implies.

GlobalFoundries announced on 21 september, in its new 14nm-14XM manufacturing technologies. The chip maker stressed in the introduction the 40 to 60 percent lower power consumption of its new 14nm-finfet-chipprocedé compared to 20nm transistors. GlobalFoundries, however, was surprisingly silent on possible reductions in die size. That is, there appear to be according to EETimes also not to be.

The new process is, according to Mojy Chian, head of design enablement at GlobalFoundries, namely focused on the transition of the 20nm cmos process to the more energy-efficient 14XM-production technology as simple as possible. The re-use of chipontwerpen in practice little reduction in size. Would it be possible to have parts of the chips smaller, but the effectiveness for the reduction is by design different, according to the chip maker.

Although the 14XM process does not make it possible to have more dies on a wafer and so achieve cost savings, GlobalFoundries’ new manufacturing technology, with the advantage that partial use can be made of existing techniques on the basis of double patterning lithography. Hence, the 14nm-processornode already a year after the launch of GlobalFoundries’ 20nm cmos node is available for customers.

That way developers can get faster, more energy-efficient processorontwerpen on the market without having to wait on new lithografietechologie. Competitor TSMC is currently engaged in the development of a 16nm process using finfet transistors.