The tighter lithography might contribute to instability problems in Intel chips.
The tighter lithography might contribute to instability problems in Intel chips.
Wendel is addressing the matter now. It seems the issue hasn't been discussed before, but on the impacted workstation chipset servers, he noted that 50% were functioning perfectly. Among those with issues, about half displayed a detected error, occurring once each week. It was mentioned that in less demanding desktop settings, such problems might go unnoticed. Around 10-15% were described as "extremely problematic."
they’re built to handle high power consumption, but I’m focusing on the energy usage here. It’s important they’re optimized for performance rather than cooling. Remember when the Ryzen 7000 first came out? There were many guides showing how to get the best performance from the 7950X with significantly less power than Intel at similar temps? That’s what matters more than heat—it’s about efficiency and scaling compared to Intel.
I get what you're saying now. What I was really pointing out was that AMD shouldn't raise the TDP of the 9700X from 65W to 120W since rumors suggest they're doing this to outpace the 7800X3D. It looks like they'll stick with the 65W and channel all that extra capacity into PBO.
The lithography shows almost no faults related to instability. Other Intel components like test PDKs and SRAMs don’t share these problems, as do Alder Lake and Tiger Lake, which use the same process. Issues might stem from power domain changes, a synthesis step before silicon fabrication, or other factors. It’s challenging to pinpoint without detailed testing. Relying on intuition about power or heat isn’t helpful here, as higher TDP actually improves yields.