1

We've seen benchmarks for Intel's Core Ultra X9 388H with its Arc B390 integrated graphics processor, codenamed Panther Lake. We've seen a comprehensive dismantling of AMD's competing Strix Point and Gorgon Point processors - and even some fascinating comparisons against Team Red's monstrous Strix Halo. But how does Panther Lake correlate to some established gameplay experiences? Well, based on my first tests with a Lenovo Idea Pad Pro 5, Panther Lake at 30W looks even more impressive than I expected: it can play Alan Wake 2 with smoother performance than Xbox Series S… and at better settings.

First of all, when Panther Lake can draw on much more power, why limit it to 30W? The answer here is simple: the 25W-35W range has become a "standard wattage" of sorts for gaming handhelds - and going into these tests, I really wanted to know just how well the new chip ran at more constrained power levels.

Next up, why test Alan Wake 2 in particular? Remarkably, in the era of aggressive dynamic resolution scaling, it's actually quite rare to find a cutting-edge console game where you can reliably drop performance beneath the target frame-rate. Alan Wake 2 is an exception. In its performance mode on PS5 and Xbox Series X, it runs with a 1440p output in FSR 2 balanced mode - a native 835p. We can also match settings very, very closely. Typically, you dial in the low setting, ramp up textures to medium or high (the game seems to sit somewhere between the two) and adjust the scattered object detail down to medium settings. Then you're good to go.

Series S comparisons? That's where things get tricky. Remedy's port is excellent bearing in mind the available memory, memory bandwidth and compute power. It caps at 30fps, it stays at 1440p, but it upscales using FSR2 from an internal 720p and it reduces fidelity in areas like foliage density and shadow quality. Matching those settings on PC doesn't look especially doable, so I tested by matching the resolution and FSR2 setting, but retained the "tweaked low" performance mode PC equivalent settings. I also added the 30fps cap from Series S.

The result from Panther Lake was impressive, especially bearing in mind that 1440p output resolution isn't typically associated with any PC SoC barring Strix Halo. I couldn't quite achieve a locked 30fps, but I did get quite close: a 29.54fps average - seven percent higher than Series S with its more compromised feature set. The improvement might not sound tremendous, but it's actually the difference between a reasonably steady 30fps with mostly even frame-pacing and a far less stable, more jerky experience on the junior Xbox.

To get some idea of how much more Panther Lake could offer, I also turned off v-sync, removed the frame-rate cap and tested again: a 32.77fps average showed that beyond 30W, my particular Panther Lake laptop (a Lenovo IdeaPad Pro 5 16IPH11) didn't offer that much more performance. CPU Package metrics suggest that this laptop's average TDP in its "geek performance" mode is around 45W and it seems to be thermally constrained there, something I've enquired with Intel about.

Next up, I went back to the standard PS5-equivalent settings: what I had already, but running FSR2 balanced mode instead of performance mode. Here, we're reminded quite rudely that Panther Lake has its limits, with the console offering a 92 percent performance lead - 55.25fps vs 28.91fps (the console's result hampered just a touch by the fact it can hit its 60fps v-sync cap.)

It's impressive stuff for an integrated graphics solution though, and while clearing the Series S standard may not get you up to PS5 performance levels, it's still worlds apart from the frame-rates delivered by the AMD Z2 Extreme within the ROG Xbox Ally and other handhelds. In fact, running the matched PS5-equivalence test still saw Panther Lake marginally exceed Series S performance by 1.32fps.

But let's finish by deploying the inevitable caveats. It's important not to read too much into just one result. We're just talking about one game here, the first of many I intend to test. A key reason I always kick off with the Alan Wake 2 "bench" is because the game seems to have weird scaling that really seems to hammer mobile integrated GPUs hard, particularly in the forest areas. Reducing FSR2 quality levels doesn't quite have the impact you may expect, for example. Scaling back settings? You'll claw back some performance but even at the minimum, it's still a tough ask for mobile processors. I'll report back soon with a more balanced range of results.