SYSTEM ANALYZER

Rate My PC: AMD Ryzen AI Max PRO 390 + Intel Arc B370

Get a comprehensive performance analysis of your gaming rig with detailed benchmarks, bottleneck detection, and upgrade recommendations

86 / 100
HIGH-END

Power Build

Top 14% of systems. Excellent for 1440p Ultra or 4K High gaming.

1440p Ultra4K High

System Balance Analysis

CPU vs GPU performance ratio
Well Balanced
CPU
96%
VS
GPU
76%
PROCESSOR

AMD Ryzen AI Max PRO 390

63,765 Benchmark Score
Top 4% Market Ranking
View Full Specs →
GRAPHICS CARD

Intel Arc B370

1,184 Benchmark Score
Top 24% Market Ranking
View Full Specs →

Market Position

How your build compares to others
Budget
0-30
Mid-Range
30-60
High-End
60-85
Enthusiast
85-100
Your Build

Game Performance Benchmarks

Real-world 4K FPS in popular titles
View All Games →

Performance Insights

Tips to maximize your system

Strong Performance

Excellent for 1440p gaming. Most games will run at high/ultra settings smoothly.

Compatible Games See what you can play Compare CPUs Find upgrades Compare GPUs Find upgrades

Performance Tiers Explained

90-100

Ultimate

4K Ultra gaming, VR ready, ray tracing enabled, professional workloads

4K 60+ FPS VR Ready
70-89

High-End

1440p Ultra or 4K High settings, excellent for modern AAA titles

1440p Ultra 4K High
50-69

Mid-Range

1080p Ultra or 1440p Medium, great value for most gamers

1080p Ultra 1440p Med
30-49

Entry Level

1080p Medium settings, suitable for eSports and older titles

1080p Med eSports
0-29

Legacy

Basic gaming, older titles, consider upgrading for modern games

720p-1080p Low Older Games

The AMD Ryzen AI Max PRO 390 paired with the Intel Arc B370 is an unusual combination: a 93rd-percentile mobile CPU bolted to a GPU sitting in just the 5th percentile of all graphics processors. That asymmetry defines everything about this build. The processor is a 12-core, 24-thread Zen 5 flagship with 64 MB of shared L3 cache and quad-channel LPDDR5X delivering 256.0 GB/s of memory bandwidth — compute hardware that trades blows with the Intel Core i9-13900KS. The graphics side, an integrated Intel Arc B370 based on the Xe3-LPG architecture, is a lightweight iGPU with 1280 shading units and a 3DMark Steel Nomad score of 1184. The result is a build whose combined percentile of 49 says "average" while hiding an extreme internal imbalance: workstation-class processing power with entry-level rendering capability. One important caveat applies throughout: no measured FPS data exists for this exact CPU+GPU pairing in the database, so all frame-rate discussion below is estimated from benchmark scores rather than directly measured.

Balance and Bottleneck

This build is bottlenecked by the GPU in essentially every graphics-limited workload, and the gap is not subtle. The CPU sits at the 93rd percentile against all processors; the GPU sits at the 5th percentile against all GPUs. That is an 88-point spread, and it means the Ryzen AI Max PRO 390 will spend most of its time in games waiting on the Arc B370 to finish drawing frames.

The evidence is in the rival comparisons. On the CPU side, the AMD chip is statistically indistinguishable from the Intel Core i9-13900KS (deltaPct of -0.4), the AMD EPYC 7343 (-0.7), the AMD Ryzen AI 7 450G (+0.7), and the Intel Core Ultra 7 265HX (+0.9). In other words, its average benchmark score of 63765 lands within a fraction of a percent of desktop-flagship and server-class silicon. On the GPU side, the Arc B370's average score of 1184 puts it alongside the ATI Radeon HD 5770 (-0.5), ATI Mobility Radeon HD 5570 (-0.2), AMD Radeon HD 7650M (-0.7), and AMD FirePro M2000 (+1.4) — a rival set drawn largely from mobile and entry-level hardware of a much earlier era.

For CPU-bound workloads — compilation, compression, physics simulation, multi-threaded number crunching — the balance flips entirely and the Arc B370 is irrelevant. PassMark multithread at 42912 and Cinebench R23 multi-core at 24828 confirm the processor can saturate 24 threads effectively. The practical conclusion: any workload that renders frames is GPU-limited here, and any workload that computes numbers is CPU-limited only in the sense that the CPU itself is the ceiling — a high one.

There is one structural detail that softens the GPU bottleneck slightly. The Arc B370 shares system memory, and the system memory here is unusually fast: quad-channel LPDDR5X at 256.0 GB/s. Integrated graphics are typically starved for bandwidth, so this platform's memory subsystem is the one place where the build actively helps its weakest component.

Benchmark Performance

The CPU's benchmark sheet is deep and consistently strong. Cinebench R23 returns 24828 multi-core and 1976 single-core; Cinebench R15 returns 3918 multi-core and 303 single-core. The single-core figures place the Zen 5 cores among fast contemporary designs, while the multi-core results show the 12-core layout scaling well — the R23 multi-to-single ratio is over 12x on a 12-core part, indicating near-linear thread utilization in rendering workloads.

PassMark results reinforce the picture: 42912 multithread, 3967 single-thread, 148508 integer math, 94666 floating point math, 506170 data compression, 26027 data encryption, 55248 string sorting, 2773 physics, 40888 extended instructions, and 323 find-prime-numbers. The compression and encryption numbers in particular show the quad-channel memory and modern core design paying off in memory- and integer-sensitive tasks.

The GPU has a single recorded benchmark: 3DMark Steel Nomad (DirectX 12) at 1184. That is the whole story, and it is a modest one. Against its nearest rivals the Arc B370 is effectively tied with hardware scoring between 1168 and 1192 — a spread of barely two percent across the entire comparison set. There is no benchmark in which this GPU distinguishes itself, and no workload class in the data where it climbs out of the 5th percentile.

The combined percentile of 49 is the arithmetic middle ground between a top-tier CPU and a bottom-tier GPU, and it should not be read as "balanced all-rounder." It is better read as two specialists averaged together: a compute powerhouse and a basic display adapter sharing one motherboard.

CPU Analysis

The Ryzen AI Max PRO 390 is a Strix Halo design on TSMC's 4 nm process, part of AMD's Ryzen AI Max PRO generation built on the Zen 5 architecture. It runs a 3.20 GHz base clock with a 5.00 GHz boost across 12 cores and 24 threads, inside a 55 W TDP. Note that this is a mobile-segment part on Socket FP11 — it is soldered-class laptop silicon, not a socketed desktop upgrade, and the multiplier is locked.

The cache hierarchy is generous: 80 KB of L1 and 1 MB of L2 per core, plus a large 64 MB of shared L3. That L3 pool is a workstation-sized resource for a mobile chip and explains part of the strong showing in latency-sensitive tests like the PassMark single-thread run of 3967. The quad-channel LPDDR5X bus delivering 256.0 GB/s is the other half of the explanation — few mobile platforms feed their cores this well, and the data compression score of 506170 is a direct beneficiary.

For real workloads, the numbers translate cleanly. The Cinebench R23 multi-core score of 24828 means sustained rendering and encoding threads will complete quickly; the 1976 single-core score means lightly threaded applications — most office software, many game engines' main loops — run responsively. ECC memory support adds a reliability dimension unusual outside professional lines, and the integrated Radeon 8050S graphics (separate from the Arc B370 in this pairing's data) exist on-die, though the database treats the Arc part as the GPU of record here.

Against its rivals, the CPU is remarkably well-placed: within one percent of the i9-13900KS, EPYC 7343, Ryzen AI 7 450G, and Core Ultra 7 265HX in average score. The practical read is that this processor behaves like a desktop flagship in compute, despite being a 55 W mobile part. PCIe is limited to Gen 4 with 16 CPU lanes, which is adequate but worth noting for expansion-heavy use.

Who Should Build It

Because this is a laptop-class platform (buildClass: laptop), the audience is buyers of a specific machine, not component pickers — but the performance profile still maps cleanly onto user types.

The strongest fit is the professional compute user on the move: developers, data analysts, and engineers who need the 24828 Cinebench R23 multi-core score for compilation, simulation, or batch processing, and who value the 256.0 GB/s memory bandwidth and ECC support for data integrity. Small business workstations fit here too — the EPYC-adjacent benchmark performance and locked, efficient 55 W envelope suit always-on office machines running databases or VMs.

Content creators who work in CPU-bound stages of the pipeline — video editors doing timeline scrubbing and proxy work, 3D artists doing scene assembly and simulation — get real value from the processor. But creators whose bottleneck is final render or GPU export should look elsewhere: the 5th-percentile GPU and its lone 1184 Steel Nomad score rule out serious GPU rendering.

Gamers are a poor match. High-refresh gaming at any resolution is off the table given the Arc B370's rival set (Radeon HD 5770-class scoring), and even modest 1080p settings will be constrained. Students and general office users, by contrast, will find the machine excellent for everything they actually do — documents, browsing, video calls, coding assignments — because those tasks lean on the 3967 PassMark single-thread score, not on the GPU.

GPU Analysis

The Intel Arc B370 is an integrated graphics part (IGP bus interface, no power connectors, 25 W TDP) built on Intel's own 3 nm process under the Xe3-LPG architecture, part of the Arc Graphics-M generation on the Panther Lake chip. Its specifications: 1280 shading units, 40 TMUs, 20 ROPs, and 10 RT cores. Clocks run from a 300 MHz base to a 2400 MHz boost, producing a 48.00 GPixel/s pixel rate, a 96.00 GTexel/s texture rate, 6.144 TFLOPS of FP32, and 12.29 TFLOPS of FP16 at a 2:1 ratio. Memory is system-shared in every respect — size, type, bus width, and bandwidth are all listed as system-dependent.

What do those numbers mean for rendering? The 6.144 TFLOPS FP32 figure is modest, and the single recorded benchmark — 3DMark Steel Nomad at 1184 — confirms the modesty. The ray tracing story is thin: 10 RT cores exist, and DirectX 12 Ultimate (12_2), Vulkan 1.4, and OpenGL 4.6 support mean the API features are present, but the 5th-percentile overall standing indicates that ray-traced rendering will be limited to minimal workloads or upscaled, low-resolution scenarios at best. Tensor core counts are not specified in the data.

The one genuine advantage is bandwidth context. As a system-shared design, the B370 draws on that 256.0 GB/s quad-channel pool rather than a conventional single-channel laptop memory configuration — the best possible feeding arrangement for an iGPU of this class. It is not enough to change the tier, but it is why the part avoids being even slower than its 1184 score suggests. For rendering verdicts: this GPU handles display output, media playback, and light 3D. It is not a rendering engine for professional output, and the data contains nothing to suggest otherwise.

Usage Scenarios

High-refresh gaming: Not viable. The Arc B370's 5th-percentile standing and its tie with Radeon HD 5770-class rivals put high frame rates out of reach; no CPU, however strong, can compensate for a GPU this far down the chart. The 93rd-percentile CPU would be idle most of the time in any modern title.

Streaming: CPU-side encoding and scene compositing are well covered by the 42912 PassMark multithread score, but the GPU cannot carry the game-side rendering load alongside it. Single-PC streaming of demanding games is a mismatch; streaming of lightweight or non-graphics content (audio, screen-sharing of desktop apps) works fine on the CPU alone.

Video editing: Timeline work, proxies, and CPU-accelerated stages benefit from the 24828 Cinebench R23 multi-core result and the 506170 data compression score. GPU-accelerated effects and final GPU export are the weak link given the 6.144 TFLOPS FP32 figure. Editors whose workflow is CPU-weighted will be productive; those relying on GPU rendering will not.

3D rendering: CPU rendering is genuinely strong — the near-parity with the EPYC 7343 (deltaPct -0.7) is the tell, as that is server-class silicon. GPU path tracing and viewport-heavy work are constrained by the 10 RT cores and the 1184 Steel Nomad score. Model with the CPU, render on the CPU.

Software development: An excellent fit. Compilation scales with cores and memory, and this platform offers 24 threads, 64 MB of shared L3, and 256.0 GB/s of bandwidth. The 3967 single-thread PassMark score keeps IDEs and interpreters snappy. Docker and VM workloads benefit from the same multi-core headroom and ECC support.

Student and office work: Effectively overpowered in the best way. Documents, browsing, conferencing, and coursework all run on the strong single-core results. Battery-relevant characteristics like the 55 W CPU TDP and 25 W GPU TDP keep the total platform envelope modest for a laptop of this compute class.

Gaming Performance

No measured FPS rows exist for this exact CPU+GPU combination in the database — the FACT PACK contains no per-game frame-rate data. Everything in this section is therefore an estimate derived from the benchmark scores, not a measurement.

The estimating basis is stark. A 3DMark Steel Nomad score of 1184, a 5th-percentile GPU ranking, and a rival set of the ATI Radeon HD 5770, ATI Mobility Radeon HD 5570, AMD Radeon HD 7650M, and AMD FirePro M2000 collectively indicate that modern AAA titles at 1080p and ultra settings are beyond this configuration's realistic reach. Esports-class and older titles at reduced settings and modest resolutions are the realistic envelope, and even there, expectations should be conservative — the deltaPct values against its rivals are all under two percent, meaning there is no hidden reserve of performance in the data.

The CPU will never be the limiter in games: 303 Cinebench R15 single-core and 2773 PassMark physics are more than adequate to feed any frame rate this GPU could produce. Every dropped frame, every settings compromise, traces back to the Arc B370 and its system-shared memory arrangement. Players should treat this as a machine that can game in a pinch, not a gaming machine.

Build Overview

This is a laptop-class build: a mobile AMD Ryzen AI Max PRO 390 (Socket FP11, Zen 5 / Strix Halo, 4 nm TSMC, released January 5, 2025) paired with Intel Arc B370 integrated graphics (Panther Lake, Xe3-LPG, 3 nm Intel, released January 26, 2026). The CPU targets the mobile segment and remains in active production, as does the GPU.

In tier terms, the build is a tale of two percentiles. The processor at the 93rd percentile is flagship-tier compute hardware for a mobile platform, rivaling the Intel Core i9-13900KS and AMD EPYC 7343 in aggregate score. The graphics solution at the 5th percentile is entry-level in absolute terms. The combined percentile of 49 lands the pairing mid-pack overall, but the composition is lopsided: this is a portable compute workstation with basic graphics, and it should be evaluated as such. Its strengths are cores, bandwidth, cache, and ECC; its weakness is anything that requires the GPU to render.

FAQ

Q: Is this build good for gaming?

A: No. The Arc B370 sits in the 5th percentile of all GPUs with a 3DMark Steel Nomad score of 1184, statistically tied with ATI Radeon HD 5770-class hardware. The 93rd-percentile CPU cannot compensate for a GPU bottleneck this severe.

Q: What CPU does it compare to in performance?

A: Its average benchmark score of 63765 is within one percent of the Intel Core i9-13900KS (-0.4%), AMD EPYC 7343 (-0.7%), AMD Ryzen AI 7 450G (+0.7%), and Intel Core Ultra 7 265HX (+0.9%) — desktop-flagship and server-class company.

Q: Are the frame rates in this page measured?

A: No. No measured FPS data exists for this exact CPU+GPU pairing; all gaming performance discussion is estimated from the benchmark scores.

Q: How much memory bandwidth does the platform have?

A: 256.0 GB/s via quad-channel LPDDR5X — an unusually high figure for a laptop that notably benefits the system-shared Arc B370 graphics.

Q: Is this a desktop or laptop build?

A: It is a laptop build. The CPU is a mobile-segment part on AMD Socket FP11 with a 55 W TDP, and the GPU is an integrated 25 W design with no power connectors.

Q: Does the CPU support ECC memory?

A: Yes, ECC memory support is present, which suits professional and small-business workloads where data integrity matters.

Q: Can it handle content creation?

A: CPU-bound creation tasks, yes — Cinebench R23 multi-core is 24828 and PassMark multithread is 42912. GPU-bound rendering is weak: 6.144 TFLOPS FP32 and 10 RT cores limit viewport and final GPU render performance.