Doom

# Doom Performance Analysis: Intel Core Ultra 5 135H + AMD Radeon RX 7900M

The measured frame rates for this mobile combo in Doom (2016) reveal a system that scales predictably across resolutions, with the RX 7900M's 16 GB frame buffer and 576.0 GB/s bandwidth proving ample for a game from this era. The Intel Core Ultra 5 135H, a 14-core Meteor Lake part with an 18 MB shared L3 cache, keeps pace at lower resolutions but shows signs of becoming a secondary factor as pixel count rises. The data spans 12 measured scenarios, from 258 FPS at 1080p Low to 54 FPS at 4K Ultra, giving a complete picture of how this laptop-class hardware handles one of the most optimized shooters of its generation.

Resolution Scaling

The frame rate drop from 1080p to 4K is steep but expected for a GPU-centric title. At High settings, the transition from 1920x1080 (190 FPS) to 2560x1440 (123 FPS) represents a 35% reduction, while moving from 1440p to 3840x2160 (67 FPS) cuts performance by another 46%. This compounding decline suggests the RX 7900M, despite its 38.52 TFLOPS FP32 compute and 4608 shading units, faces a growing pixel throughput challenge at higher resolutions. The 4K High result of 67 FPS is just above the 60 FPS threshold, indicating the GPU is the primary limiter at this resolution.

The scaling pattern becomes more revealing when examining Low settings. At 1080p Low, the combo hits 258 FPS, but at 4K Low it drops to 90 FPS — a 65% reduction. This steeper relative decline at lower quality presets points to the GPU's raw rasterization limits rather than any memory or bandwidth constraint. The 256-bit memory bus with 2250 MHz GDDR6 (18 Gbps effective) delivers 576.0 GB/s, which is more than sufficient for Doom's textures even at 4K. The data implies that resolution scaling is governed almost entirely by the RDNA 3.0 architecture's pixel and texture throughput — 401.3 GPixel/s and 601.9 GTexel/s respectively — rather than any system-level bottleneck.

Between Medium and High settings, the frame rate differences at each resolution are modest. At 1080p, Medium (225 FPS) beats High (190 FPS) by 18%, while at 1440p the gap is also 18% (145 vs 123 FPS). At 4K, Medium (79 FPS) outperforms High (67 FPS) by 18% as well. This consistent 18% delta across resolutions suggests the settings change affects the same GPU workload components uniformly. However, the jump from Medium to Ultra at 4K (54 FPS) represents a 32% drop from Medium, which is steeper than the 31% drop seen at 1080p (155 vs 225 FPS) and 31% at 1440p (100 vs 145 FPS). Ultra settings introduce effects that disproportionately tax the GPU at higher pixel counts.

The 1080p to 1440p scaling at Ultra (155 to 100 FPS, a 35% drop) closely mirrors the High preset's 35% drop, but the 1440p to 4K Ultra transition (100 to 54 FPS, a 46% drop) is more severe than High's 46% drop. This indicates that Ultra's additional visual features — likely including higher-quality shadows and post-processing — amplify the resolution penalty. The min FPS data for Medium settings shows the frame time consistency: at 1080p, the min of 191 FPS stays close to the 225 FPS average, while at 4K the min of 67 FPS matches the 79 FPS average reasonably well, suggesting no severe stuttering despite the heavy load.

GPU Role

The AMD Radeon RX 7900M sits in the 94th percentile among all GPUs, with an average benchmark score of 97487. Its nearest rivals — the AMD Radeon Pro VII (0.4% faster), NVIDIA Quadro RTX 6000 (4.3% slower), AMD Radeon Instinct MI60 (5.4% faster), and NVIDIA RTX A4500 (6.3% faster) — bracket this mobile part in workstation-class territory. In Doom specifically, the 16 GB GDDR6 memory never appears to be a constraint; even at 4K Ultra, the game's texture requirements fall well within this capacity. The 72 RT cores and RDNA 3.0 architecture provide DirectX 12 Ultimate (12_2) support, though Doom's 2016 release predates ray tracing and relies on traditional rasterization.

The GPU's clock behavior — 1825 MHz base and 2090 MHz boost — combined with 288 TMUs and 192 ROPs, explains the consistent performance scaling. The texture rate of 601.9 GTexel/s and pixel rate of 401.3 GPixel/s are the physical limits that manifest at 4K. At 1080p, the GPU is not fully saturated, which is why Low settings reach 258 FPS — likely limited by CPU draw-call submission or driver overhead rather than the GPU itself. The transition from 1080p to 1440p shows the GPU beginning to assert dominance, and by 4K it is firmly the bottleneck.

The fact that 4K High (67 FPS) outperforms 4K Ultra (54 FPS) by 24% while both use the same 16 GB memory buffer indicates that the performance gap comes from shader complexity and fill-rate demands, not capacity. The 57,700 million transistors on a 529 mm² die at 5 nm from TSMC give this GPU substantial compute headroom, but Doom's rendering pipeline at Ultra settings pushes the 192 ROPs harder as resolution increases. The 180 W TDP, while not directly measured in frame rates, suggests the mobile platform can sustain these clocks during extended sessions, as evidenced by the stable min FPS values at Medium settings.

CPU Role

The Intel Core Ultra 5 135H, with 14 cores and 18 threads, ranks in the 81st percentile among all CPUs. Its nearest rival, the AMD Ryzen 7 5800X, scores nearly identically (0.1% faster), while the Intel Core i5-13500HX is 0.6% slower. This places the mobile chip in desktop-class territory for multi-threaded workloads, though Doom's engine from 2016 is not known for heavy multi-threading. The 18 MB shared L3 cache and 89.6 GB/s memory bandwidth provide sufficient data feeding for the GPU at most settings.

At 1080p Low, the 258 FPS result likely approaches the CPU's frame submission limit. The base clock of 3.60 GHz and boost of 4.60 GHz across the 14 cores give the processor enough single-thread headroom for Doom's main game thread, but the 112 KB L1 cache per core and 2 MB L2 per core may create occasional bottlenecks at very high frame rates. The data shows that going from 1080p Low (258 FPS) to 1080p Medium (225 FPS) reduces the CPU load, allowing the GPU to become the limiting factor — the CPU is only truly constrained at the lowest resolution and quality settings.

The Cinebench R23 scores — 11875.5 multi-core and 1700 single-core — indicate a processor that can handle both parallel and single-threaded tasks competently. For Doom specifically, the single-core performance matters more, and the 1700 R23 single-core score aligns with the observed ability to push beyond 250 FPS at 1080p. The PassMark single-thread score of 3521 further confirms this. As resolution increases, the CPU's role diminishes; the 4K results show frame rates that are entirely GPU-bound, with the CPU waiting on the GPU's render output rather than the reverse.

The passmark physics score of 1482 and floating-point math score of 54867 suggest the CPU handles game physics calculations without issue, but Doom's Vulkan API support (the GPU lists Vulkan 1.4) can leverage multi-threading more effectively than older DirectX paths. The 14-core configuration with 18 threads (likely 6 performance cores plus 8 efficiency cores with hyperthreading on the performance cluster) provides enough parallelism for draw-call batching, though the data does not isolate CPU-specific frame times.

Settings Recommendations

The measured rows provide clear guidance for optimizing the experience. At 1080p, every preset delivers playable frame rates, with Ultra (155 FPS) still far above the 60 FPS standard. The jump from High (190 FPS) to Ultra (155 FPS) costs 18% performance while adding visual fidelity — for competitive play, High or Medium offers headroom above 200 FPS. At 1440p, Ultra (100 FPS) remains smooth, but High (123 FPS) provides a 23% boost with minimal visual compromise. The Medium preset at 1440p (145 FPS) is the sweet spot for high-refresh displays, as it maintains frame rates above 120 FPS.

At 4K, the recommendations shift decisively. Ultra (54 FPS) falls below the 60 FPS threshold and is not ideal for smooth gameplay. High (67 FPS) barely exceeds 60 FPS, making it the minimum viable preset for 4K. Medium (79 FPS) offers the best balance, delivering 47% more performance than Ultra while retaining most visual quality. Low (90 FPS) is the only preset that provides a comfortable margin above 60 FPS at 4K, though it sacrifices shadows and texture quality. For a 4K display, Medium is the recommended starting point, with High as an alternative if the user prioritizes visuals over frame rate stability.

The min FPS data for Medium settings reveals frame time consistency across resolutions: 191 min at 1080p, 124 min at 1440p, and 67 min at 4K. These minimums stay within 15-16% of their respective averages, indicating that the combo avoids severe frame drops even under load. This consistency suggests that Medium at 4K (79 avg, 67 min) is a safer choice than High at 4K, where no min FPS data is available but the lower average leaves less headroom. For users with variable refresh rate displays, Medium at 1440p or 4K will provide the most stable experience.

FAQ

Q: Can this combo run Doom at 4K 60 FPS?

A: Yes, but only at Medium settings (79 FPS average, 67 FPS minimum) or Low settings (90 FPS average). High settings drop to 67 FPS average, which is marginal, while Ultra falls to 54 FPS, below the 60 FPS threshold.

Q: What is the best preset for 1080p gaming?

A: High settings (190 FPS) provide a strong balance, but Ultra (155 FPS) is also viable if you want maximum visual quality. For competitive play, Medium (225 FPS) or Low (258 FPS) offer the highest frame rates.

Q: Is the CPU or GPU the bottleneck in this game?

A: At 1080p Low, the CPU (Intel Core Ultra 5 135H) is likely the limiting factor at 258 FPS. At 4K and higher quality settings, the GPU (AMD Radeon RX 7900M) becomes the clear bottleneck, as shown by the steep frame rate drops with resolution.

Q: How does the RX 7900M compare to its nearest GPU rivals?

A: The RX 7900M ranks in the 94th percentile. It is 0.4% slower than the AMD Radeon Pro VII, 4.3% faster than the NVIDIA Quadro RTX 6000, 5.4% slower than the AMD Radeon Instinct MI60, and 6.3% slower than the NVIDIA RTX A4500.

Q: Does the 16 GB VRAM ever limit performance in Doom?

A: No. The game from 2016 does not exceed 16 GB of VRAM at any resolution or settings level. The memory bandwidth of 576.0 GB/s is also sufficient, as frame rate scaling follows GPU compute limits rather than memory constraints.

Q: What frame rates can I expect at 1440p with this setup?

A: Ultra settings deliver 100 FPS, High reaches 123 FPS, Medium hits 145 FPS, and Low achieves 167 FPS. All presets are well above 60 FPS, making 1440p the most versatile resolution for this combo.

Measured FPS Breakdown

At 1920x1080, the combo delivers its highest frame rates. Low settings produce 258 FPS average, Medium reaches 225 FPS (with a minimum of 191 and maximum of 258), High achieves 190 FPS, and Ultra drops to 155 FPS. The spread from Low to Ultra is 40%, with each step from Low to Medium (13% drop), Medium to High (16% drop), and High to Ultra (18% drop) showing diminishing returns on visual quality versus performance cost.

At 2560x1440, the frame rates remain strong. Low settings hit 167 FPS, Medium averages 145 FPS (minimum 124, maximum 167), High reaches 123 FPS, and Ultra maintains 100 FPS. The scaling from 1080p shows a 35% reduction at High, 36% at Medium, 35% at Low, and 35% at Ultra — remarkably consistent, indicating the GPU's pixel throughput scales linearly with resolution until 4K.

At 3840x2160, performance becomes more constrained. Low settings yield 90 FPS, Medium averages 79 FPS (minimum 67, maximum 91), High drops to 67 FPS, and Ultra falls to 54 FPS. The 4K to 1440p ratios show a 46% reduction at High, 46% at Medium, 46% at Low, and 46% at Ultra — again consistent, but steeper than the 1080p to 1440p drop. This suggests the GPU hits a throughput ceiling around 4K that is independent of settings.

The Medium preset is the only one with complete min/max data across all three resolutions, providing insight into frame pacing. At 1080p, the 191-258 FPS range shows a 67 FPS spread. At 1440p, the 124-167 FPS range has a 43 FPS spread. At 4K, the 67-91 FPS range narrows to a 24 FPS spread. This tightening of frame time variance at higher resolutions indicates that the GPU becomes more consistent as it becomes the bottleneck, reducing CPU-induced stutter.

How This Combo Ranks

The Intel Core Ultra 5 135H + AMD Radeon RX 7900M combination ranks 1226 out of 2930 tested combos in Doom, placing it in the 58th percentile of all system configurations. This mid-pack ranking is somewhat surprising given the GPU's 94th percentile standing, but it reflects the CPU's more modest 81st percentile position. The ranking suggests that while this combo handles Doom well in absolute terms, many other tested systems — likely with stronger CPUs or desktop GPUs — achieve higher frame rates, particularly at lower resolutions where the CPU matters more.

The gap between the GPU's high percentile and the combo's mid-pack rank implies that the 135H may be holding back the RX 7900M in CPU-limited scenarios. At 1080p Low, the 258 FPS result is likely near the CPU's maximum frame delivery capability, whereas a higher-end CPU might push the GPU further. However, at 4K Ultra, the combo's 54 FPS is consistent with GPU-bound performance, where the CPU's influence is minimal. The ranking of 1226 out of 2930 means 1704 combos perform worse, but 1225 perform better — a competitive position that favors users targeting 1440p or 4K gaming over esports-level frame rates.

The CPU's nearest rival, the AMD Ryzen 7 5800X (0.1% faster in average benchmark score), represents a desktop-class comparison point. In Doom, the 135H's 14-core configuration with 18 threads may not provide the same single-thread performance as the 5800X at lower resolutions, which could explain why the combo ranks lower than systems pairing the 7900M with stronger CPUs. The GPU's nearest rivals — all workstation parts — suggest that the 7900M's raw compute is competitive, but Doom's optimization for consumer gaming hardware means the combination's ranking reflects the full system balance rather than any single component's strength.