Imho, I think what he's trying to say is that LNC didn't bring any great new ideas. Mostly minor upgrades rather than sweeping changes.
For comparison, Zen 5 didn't bring in any major performance uplift this gen, but the excellent ground work they've laid should help them immensely moving forward.
Intel has been spending way too much time and energy porting their existing designs to newer tool and design methodologies. But now they're on solid foundation and should be able to slice and dice easily to bring in sweeping changes next gen onwards.
He clearly wrote that LionCove is not contributing anything new, just like GoldenCove was not contributing anything to SunnyCove etc. And that is a lie.
The OoO engine in LionCove has been radically redesigned with future generations in mind, as has modular decoding in Zen 5.
These are fundamental changes for Intel and AMD that, to the average user who has no idea, will be nothing more than an evolution in the final performance of the LionCove and Zen5 generations.
"Unlike AMD Zen 5’s clustered decoder, all eight decode slots on Lion Cove can serve a single thread. Lion Cove can therefore sustain eight instructions per cycle as long as code fits within the 64 KB instruction cache."
Test Comment Lion Cove IPC Redwood Cove IPC Zen 5 IPC
XOR r,r Commonly used to zero registers. The exclusive-or of two identical values is always zero 7.31 5.7 5.01
XOR xmm, xmm Same as above but for a vector/FP register 7.31 5.71 4.99
Dependent MOV r,r >1 indicates move elimination 7.02 5.56 6.65
Independent MOV r,r Easy 7.25 5.71 5.01
Dependent increment Actual math, normally would create a dependency chain limiting the test to 1 IPC 5.6 5.53 1
Dependent add immediate As above but adding small numbers up to 20 instead of just 1 4.36 5.47 1
"Core Summary
All those caches help feed Lion Cove’s core, which has huge upgrades over Redwood Cove. The pipeline is wider, structures are larger, and a reorganized out-of-order engine helps Intel achieve higher scheduling capacity.
Much like Redwood Cove, Lion Cove is a wide and high clocked out-of-order design. But it’s easily the biggest change to Intel’s performance oriented architecture since Golden Cove. After Redwood Cove’s minor changes over Raptor Cove, and Raptor Cove barely doing anything over Golden Cove, it’s great to see Lion Cove’s sweeping changes.
Intel must have put a lot of effort into Lion Cove’s design. Compared to Redwood Cove, Lion Cove posts 23.2% and 15.8% gains in SPEC CPU2017’s integer and floating point suites, respectively. Against AMD’s Strix Point, single threaded performance in SPEC is well within margin of error. It’s an notable achievement for Intel’s newest P-Core architecture because Lunar Lake feeds its P-Cores with less L3 cache than either Meteor Lake or Strix Point. A desktop CPU like the Ryzen 9 7950X3D only stays 12% and 10.8% ahead in the integer and floating point suites respectively. Getting that close to a desktop core, even a last generation one, is also a good showing."
"Final Words
P-Cores have been Intel’s bread and butter long before the company started calling them P-Cores. Progress with Intel’s performance oriented cores hasn’t always been fast. Redwood Cove was only a slight tweak over Golden Cove. Skylake filled out five generations of Intel designs the same architecture. Going back further, Intel used the P6 architecture on the Pentium Pro, Pentium II, and Pentium III with just minor tweaks and clock speed increases in between.
Lion Cove is a much improved architecture compared to Redwood Cove, and shows Intel still has potent engineering muscle despite recent setbacks. Traditionally Intel delivered significant architecture changes during a “tock” in a tick-tock cycle. That reduces risk by separately handling process node and architecture changes. Lunar Lake not only combines a new architecture with a move to a new node, but also drops system level changes on top. At a time when Intel’s facing increased pressure from all sides, a move like Lunar Lake is a sign that Intel can adapt and survive.
Intel’s upcoming Arrow Lake desktop CPU will let Lion Cove stretch its legs with more cache and a larger power budget. Lower latency DDR5 should improve performance even further. After seeing Lion Cove perform well in a mobile form factor, I’m optimistic about what the same architecture can do on desktop. Recently Intel has been sitting on a rather unstable foundation with Raptor Lake, and Arrow Lake’s release will be a great time to put the company’s high performance chips back on stable footing."