Intel releases information about their 10nm+ Ice lake series of CPUs

Intel releases information on their 10nm+ Ice lake series of CPUs

Intel releases information about their 10nm+ Ice lake series of CPUs

While most of us are waiting for more information on Intel’s upcoming 8th generation of desktop CPUs, codenamed Coffee Lake, many of us are more excited about that is coming next from Intel. 
It has been a while since Intel has made the move to 14nm, first with Broadwell, then with Skylake and now with Kaby Lake. Soon 14nm will be getting another generation of CPUs, which is set to increase core counts on desktop platforms and move mainstream PCs away from the Quad-core Era. 
Intel’s 9th Generation of CPUs will be using the company’s new Ice Lake CPU architecture, which is expected to offer some IPC improvements compared to Intel’s current-generation offerings and use Intel’s new 10nm+ manufacturing process. At this time the release of Ice Lake is not expected until at least the second half of 2018, though at this time we cannot be certain. 


Intel releases information on their 10nm+ Ice lake series of CPUs

When Intel officially announced their 10nm manufacturing technology, they stating that they will be able to pack 2x as many transistors than other competing “10nm” manufacturing nodes. 
This will be Intel’s third Generation FinFET manufacturing node and it will be a true 10nm design, effectively calling TSMC’s upcoming 10nm process a “fake 10nm”. This is because the industry does not use a standardised unit of measurement for defining the “size” of a manufacturing node. 

These days, most fabs use the ??nm term for marketing, rather than a physical measurement or characteristic of characteristic of their processing node, this has lead to a situation where TSMC is now calling an enhanced version of their 16nm FinFET node 12nm, despite the fact that the size of the node has remained unchanged.    

Intel right now is the only major Fab that properly scales their node naming schemes with their process node sizes, giving them what could be defined as the “best” processing technology in terms of transistors per unit area. 


       The minimum gate pitch of Intel’s 10 nm process shrinks from 70 nm to 54 nm and the minimum metal pitch shrinks from 52 nm to 36 nm. These smaller dimensions enable a logic transistor density of 100.8 mega transistors per mm2, which is 2.7x higher than Intel’s previous 14 nm technology and is expected to be approximately 2x higher than other industry 10 nm technologies.


With Ice Lake Intel will be moving to their second generation 10nm node, which may make you wonder what happened to 1st generation 10nm, which was not used for any mass produced Intel products. 

Intel’s 10nm manufacturing process has so far not yielded the performance results that Intel wanted, with their 1st generation node offering either a 25% performance boost over 14nm or 45% lower power draw. Compare this to 14nm++, Intel’s 3rd generation 14nm node, which offers a 26% performance boost or a 26% reduction in power draw when compared to 14nm.  

This meant that Intel was better off using 14nm++ for now, as 10nm would not yield any performance gains over 14nm++, a node which likely has higher yields and manufacturing capacity. Intel’s baseline 10nm process would be superior for low power designs, but not for high-end products.

Intel has now created their 10nm+ node, which will offer a 15% performance boost over 10nm, with up to a 30% reduction in power, which is more than enough to make it a suitable replacement for 14nm++. 

Intel officially announces their 10nm manufacturing technology


Ice Lake is a long way from releasing, but it should be the first full process node advancement from Intel in several generations, which should yield them some noteworthy performance benefits. 


You can join the discussion on Intel’s Ice Lake series of CPUs on the OC3D Forums. 


Special Thanks to TheF34RChannel for providing us with this information.