You are currently viewing SemiWiki as a guest which gives you limited access to the site. To view blog comments and experience other SemiWiki features you must be a registered member. Registration is fast, simple, and absolutely free so please, join our community today!
News about Intel 18AP from VLSI_2026 Paper T1.2
- Intel 18AP will offer 9% perf or 18% power over 18A
- New VT pairs, low and high power devices
- Numbers based on Arm core sub-block 30% skew corner tightening
- Matched SRAM Vmin.
Intel 18A-P looks impressive - 9% more speed even at a lower voltage (0.76V --> 0.75V).
"Improved Thermal Conductivity by 50% sounds pretty good, given BSPD challenges.
..
I wonder how TSMC's variant of BSPD compares to Intel in terms of thermal conductivity, (and related ability to clock high without thermal issues. ). A16 should be denser than 18A/-P.
Intel 18A-P looks impressive - 9% more speed even at a lower voltage (0.76V --> 0.75V).
"Improved Thermal Conductivity by 50% sounds pretty good, given BSPD challenges.
..
I wonder how TSMC's variant of BSPD compares to Intel in terms of thermal conductivity, (and related ability to clock high without thermal issues. ). A16 should be denser than 18A/-P.
I also wonder about where the yields will be for these new processes at Intel with respect to AMD. Rumor is that AMD's desktop N2 chip will clock to around 6.2-6.4 (with some crazy people saying 7). I know I have heard that one issue with BSPDN is hot spots in the die that limit clock speed.
Also, the equipment to make these chips has gotten so horribly expensive, and the throughput per wafer keeps going down due to more passes. It's certainly not like the old days when Moore's Law was alive and well .
I have actually been more excited about the new packaging and stacking techniques than the lithography improvements.
.