Intel Bartlett Lake vs i9-13900K: More Cores, Fewer Frames

Intel's Bartlett Lake Flagship Loses to a Four-Year-Old CPU in Gaming Benchmarks

Intel's Core 9 273PQE is Bartlett Lake-S's flagship — a processor engineered with substantially more P-cores than the 2022 Core i9-13900K. On paper, the newer chip should be in a different performance tier. In practice, public benchmark analysis compiled by outlets including Tom's Hardware and TechPowerUp consistently shows the 273PQE unable to surpass the older Raptor Lake champion in typical gaming workloads. The result is counterintuitive but reproducible, and the architectural reasons illuminate what gamers actually need from a CPU in 2025.

The Core i9-13900K's documented specification profile is useful framing: eight Raptor Cove P-cores paired with sixteen Gracemont E-cores, a 36 MB L3 cache per Intel's ARK database, and a single-core boost frequency of 5.8 GHz. That architecture was tuned for bursty, latency-sensitive workloads — precisely the access pattern that most commercial game engines impose on a processor. Bartlett Lake-S was designed with a different buyer in mind, and its gaming performance reflects that priority.

---

Why More P-Cores Don't Automatically Mean More Frames

The intuitive assumption — that additional performance cores yield proportionally better gaming results — breaks down when examining how modern game engines schedule work. Per Tom's Hardware's CPU benchmark hierarchy, the majority of commercially released titles in 2024–2025 remain bounded by a single heavy thread or a small cluster of tightly coupled threads, rarely saturating more than six to eight logical threads simultaneously.

The Core i9-13900K's Raptor Cove P-cores deliver high instructions-per-clock (IPC) throughput at frequencies that routinely reach 5.6–5.8 GHz under single-core boost conditions. Bartlett Lake-S adds more P-cores to its die but does not introduce a substantially revised microarchitecture that determines how many instructions each core retires per cycle. Community benchmark data compiled by outlets such as TechPowerUp suggests the per-core performance delta — rather than total core count — is the primary driver of gaming frame-rate differences between the two platforms.

In practice, a game that fully saturates four to six threads sees diminishing returns from a processor carrying two or three times as many cores, particularly when those additional cores operate at reduced single-thread frequencies under sustained load. Thermal constraints amplify this: as Bartlett Lake-S's larger die builds heat during all-core workloads, per-core boost headroom compresses and the frequency ceiling narrows further.

---

Cache Hierarchy and Latency: The Underappreciated Battleground

Beyond raw core count and clock speed, cache design significantly shapes gaming performance — especially in open-world titles and games with large streaming asset sets. The Core i9-13900K's 36 MB L3 cache, per Intel's ARK specification, serves its 24 threads with low-latency access to frequently touched data. Gaming workloads issue short, cache-sensitive reads across texture buffers, physics state tables, and AI decision data; a cache topology optimised for sustained throughput rather than low-latency random access can leave per-frame times higher than raw specifications suggest.

This dynamic was documented extensively during the AMD 3D V-Cache era: the Ryzen 7 5800X3D demonstrated that cache topology — not just capacity — can override IPC and frequency advantages in game-bound scenarios. Per benchmark analysis from GamersNexus and TechPowerUp, the stacked-cache approach produced frame-rate gains that a straightforward frequency or IPC improvement of equivalent die cost could not have matched.

Bartlett Lake-S does not employ an equivalent stacked-cache approach. Its cache configuration is competitive for server and workstation workloads that prefer high bandwidth and large working sets, but the bursty, low-latency access pattern of real-time game rendering is a different problem.

For a forward-looking look at how Intel is addressing the cache problem for gaming, the upcoming Nova Lake-S platform reportedly introduces dedicated game-cache configurations. See Intel Nova Lake-S's 22-core game-cache SKUs for the latest details.

---

Where Bartlett Lake Does Win: Productivity and Parallel Workloads

The 273PQE's additional P-cores are not wasted — they serve a different user profile. In sustained multi-threaded workloads such as video encoding (x264, x265, AV1), 3D rendering (Blender, V-Ray), large software compilation jobs, and data-science pipelines, higher core counts translate directly into throughput gains. Per public multi-threaded benchmark data compiled by Tom's Hardware and Puget Systems, Bartlett Lake-S leads Raptor Lake in sustained parallel compute scenarios.

The tradeoff is clear: a professional workstation benefits from the 273PQE's architecture in ways a gaming PC does not. For a content creator who also games, the single-CPU choice depends on which workload dominates. Where gaming — particularly competitive titles at 1080p where frame rate and 1% lows determine the experience — takes priority, the legacy benchmark data favours the 13900K.

AMD's Ryzen 9 7950X occupies an interesting middle position. Per community benchmark compilations from multiple outlets, Zen 4's IPC improvements and efficient memory subsystem allow it to bridge multi-threaded productivity and gaming more gracefully than Bartlett Lake-S, delivering competitive throughput while sustaining stronger gaming frame rates than the 273PQE across a range of tested titles.

---

The IPC Gap: Architectural Context

Intel's Raptor Cove microarchitecture, which powers the Core i9-13900K's P-cores, represented the company's highest-IPC mainstream desktop design at launch. IPC advances require significant microarchitectural investment: wider decode widths, deeper out-of-order execution windows, improved branch predictors, and revised execution port configurations. Per public architectural analysis from outlets including AnandTech's historical coverage, Bartlett Lake-S does not introduce the same order of per-core capability revision that differentiated, say, Golden Cove from Cypress Cove.

The situation is not unique to Intel. AMD's Zen 5 generation delivered more modest-than-expected single-threaded IPC gains over Zen 4 in some workloads, illustrating that sequential IPC scaling has become harder as high-performance CPU microarchitectures mature. For competitive gaming at high framerates, architectural efficiency per cycle matters more than adding more cores.

For a look at where Intel's per-core trajectory is heading, the Intel Core Ultra X7 Panther Lake Linux 7.1 benchmarks provide early signals on the next-generation P-core design, and the Panther Lake NPU vs RTX 3060 analysis shows how Intel is addressing non-gaming AI compute workloads separately.

---

Gaming Workload Profiling: What the Benchmarks Actually Measure

Benchmark methodology matters for interpreting these results. When hardware reviewers report CPU gaming performance, they typically measure average frame rates and 1% low frame times across a curated title suite at a GPU-unconstrained resolution — usually 1080p with a high-end discrete GPU to ensure the CPU is the binding constraint. Under this methodology, the characteristics that matter most are:

1. Single-core boost frequency — how fast the CPU can sustain when one thread carries the primary workload
2. P-core IPC — how many useful instructions complete per clock cycle
3. Cache hit rate and latency — how often the CPU retrieves needed data without stalling on system memory
4. Thread scheduling efficiency — relevant when the game engine spawns multiple interdependent short threads

The Core i9-13900K scores well across all four dimensions for mainstream gaming. Bartlett Lake-S offers more capacity on dimension four — additional threads reduce inter-thread queuing — but makes architectural tradeoffs against dimensions one through three in the process. Since most commercial titles stress dimensions one through three more than four, the benchmark outcome reflects the architecture's priorities rather than a design error.

For further historical context on how Intel's platform succession works in practice, the Intel Core i7 6700K and Intel Core i5 6500 illustrate how platform longevity and architectural efficiency determine real-world relevance years after launch.

---

Is Bartlett Lake Worth Buying for a Gaming PC?

For a pure gaming build, the Core i9-13900K remains — per current pricing aggregators — a compelling secondhand platform choice. Z790 and Z690 motherboards, paired DDR5 memory, and the CPU itself have all declined substantially in secondhand price since launch, and public benchmark analysis consistently places the 13900K at or near the top of the mainstream gaming hierarchy.

The 273PQE makes more sense for a workstation used heavily for multi-threaded professional applications alongside occasional gaming sessions, or for a platform where the buyer expects game engines to better exploit higher core counts over a long ownership window.

The Bartlett Lake situation also reinforces a practical buying-guide principle: review the benchmark workload mix — not the spec sheet — before purchasing. A processor with more cores and a higher TDP does not automatically deliver more frames per second, and in gaming-specific evaluations, it sometimes delivers fewer.

---

What's Next: Nova Lake-S and the IPC Correction

Intel's roadmap points to Nova Lake-S as the next major desktop architecture. Per leaked specifications and industry reporting cited by Tom's Hardware and TechPowerUp, Nova Lake-S is expected to introduce substantially revised P-cores with a new microarchitectural design — the first significant IPC-focused change since Raptor Cove. Two 22-core SKUs with dedicated game-cache configurations are already in pre-production, as covered in the Intel Nova Lake-S 22-core game-cache SKU analysis and the Nova Lake-S game-cache news brief.

If Nova Lake-S delivers on IPC commitments while retaining the high core counts Bartlett Lake-S introduced, Intel would field a platform capable of leading both gaming and productivity benchmarks simultaneously — the combination the 273PQE cannot currently offer. For a direct comparison of how that architectural approach stacks up against AMD's current generation, see Intel Nova Lake-S 22-Core vs AMD Ryzen 7 5800X.

For AI inference workloads — an increasingly relevant consideration as gaming PCs double as local LLM hosts — the picture shifts again. Dedicated NPU silicon changes the calculus outside frame-rate-focused gaming, as examined in the Panther Lake NPU vs RTX 3060 local LLM analysis.

Until Nova Lake-S arrives with verified benchmark data, the Core i9-13900K's continued relevance in gaming comparisons reflects not nostalgia but the enduring importance of per-core efficiency — a lesson Bartlett Lake-S makes concrete.

---

Citations and sources

• https://ark.intel.com — Intel Core i9-13900K official specifications (P-core count, L3 cache, boost frequency)
• https://www.tomshardware.com/reviews/cpu-benchmark-charts-hierarchy,4393.html — Tom's Hardware CPU benchmark hierarchy and gaming workload analysis
• https://www.techpowerup.com — TechPowerUp CPU review archive, benchmark database, and architectural commentary
• https://www.gamersnexus.net — GamersNexus CPU benchmark methodology, 1% low analysis, and per-core performance reviews
• https://www.pugetsystems.com/go/cpu — Puget Systems multi-threaded workstation CPU benchmarks (rendering, encoding)

This piece is editorial synthesis based on publicly available information. No independent first-party benchmarking is reported.