Intel Comet Lake-S processors have only been out for a couple of months at this point, but the next-generation Rocket Lake leaks and rumors just keep piling up.
The latest of which comes courtesy of known hardware leaker @_rogame, who reiterates the 5GHz boost clocks we reported on earlier, but the bigger story is the IPC performance. According to @_rogame, we’re going to see a 10% IPC (instructions per clock) increase over Skylake, which when combined with the high clock speeds, may be enough to hold off AMD.
Some Intel RocketLake info > Cypress cove core (confirming @mooreslawisdead info)> IPC over SKL is around 10% +-> 5GHz+ final clocks> TDP is highJuly 25, 2020
However, while we have heard rumors that the 11th-generation Intel Rocket Lake processors will be out in the near future, it’s important to note that AMD Zen 3 chips will be out this year, as Team Red has reiterated again and again. The shrink down from 12nm Zen+ to 7nm Zen 2 brought an IPC uplift somewhere in the ballpark of 15% in 2019. And, it’d be foolish to expect that AMD has been sitting around not improving IPC performance for Zen 3 – as even with the new 3000XT chips, AMD is still behind on the single-core battle.
If Intel wants to maintain its hold on the “best gaming processor” title that it loves to parade around whenever the company releases a new desktop lineup, then this rumored IPC increase is necessary. We somehow doubt that the AMD Ryzen 4000 processors are going to boost up to 5GHz, but if AMD has stronger IPC, then Team Red could do more with less clock speed, which would be a nightmare for Intel.
Only time will tell either way. According to that @_rogame tweet, the next generation desktop lineup will be based on a new Cypress Cove architecture, still based on the 14nm manufacturing node. This basically has to be the case if these IPC improvements are ever going to be true. Either way, if Intel is getting ready to follow up Comet Lake-S with Rocket Lake this soon, the CPU battleground is going to heat up immensely in the near future.