CPU Comparison
Intel Xeon 698X vs Intel Xeon w9-3595X
A side-by-side comparison of specs, performance and value. The Intel Xeon 698X is an 86-core, 172-thread workstation processor based on the Granite Rapids-WS architecture, built on Intel 3 and designed for single-socket workstations that need massive core counts, eight-channel DDR5/MRDIMM memory, and 128 PCIe 5.0 lanes.
The Bottom Line
Overview & Launch
Specifications Compared
Performance Compared
Productivity
Gaming
Virtualization
Efficiency
Specialized Performance
AI / ML
- AMX supports BF16, INT8, and native FP16, important for PyTorch/TensorFlow inference.
- No integrated GPU or dedicated NPU; AI acceleration is CPU-only via AMX and AVX-512.
- Best suited for CPU-based inference, small-to-medium model training, and data preprocessing rather than large-scale GPU training.
- Intel AMX accelerates matrix operations for AI inference and training on CPU.
- Intel Deep Learning Boost (VNNI) supported.
- Lacks integrated NPU; relies on CPU and GPU acceleration.
Content Creation
Gaming
- Not targeted at gaming; no official gaming benchmarks.
- High core count does not benefit most games, and many games won’t use more than a fraction of the available threads.
- Single-threaded performance is competitive, but gaming-focused CPUs will provide better value and often higher effective FPS per dollar.
- Single-core boost is competitive but many mainstream desktop CPUs match or exceed it at far lower power.
- No integrated graphics means a discrete GPU is mandatory.
- Not designed or optimized for gaming; professional workloads are the target.
Industry Impact
Best CPU by Use Case
Target Audience
Strengths & Weaknesses
Pros
- 86 cores and 172 threads for highly parallel workloads.
- 336 MB L3 cache improves performance on large data sets.
- Eight-channel DDR5/MRDIMM memory with up to 4 TB capacity.
- 128 PCIe 5.0 lanes for multi-GPU and storage-heavy configurations.
- Intel 3 process and Redwood Cove+ cores improve performance and efficiency over Sapphire Rapids.
- AMX with native FP16 acceleration for AI inference.
- Unlocked multiplier for overclocking, supported by Intel and partners.
Cons
- Very high power consumption (350 W base, up to 420 W turbo) requiring robust cooling and power supply.
- Expensive, with street prices around $8,300–$8,500 for the CPU alone.
- New platform (W890 chipset, LGA4710) with early-adoer considerations and limited long-term platform history.
- No integrated graphics, requiring a discrete GPU for display output.
- Gaming and lightly threaded workloads see little benefit relative to cheaper, lower-core-count CPUs.
Pros
- 60 Performance-cores and 120 threads for massive parallelism.
- 112 PCIe 5.0 lanes for extensive expansion.
- Eight-channel DDR5-4800 ECC with up to 4 TB capacity.
- Unlocked multiplier for performance tuning.
- Intel AMX and DL Boost for AI acceleration.
- Intel vPro Enterprise and remote management features.
- Turbo Boost Max 3.0 up to 4.8 GHz on favored cores.
- VT-x/VT-d virtualization support.
Cons
- High power draw: 385 W base and 462 W max turbo require serious cooling.【turn4fetch0】
- No integrated graphics.
- Single-threaded performance lower than many desktop CPUs.
- W790/LGA4677 platform has limited long-term upgrade path.
- Strong competition from AMD’s Threadripper PRO line in many creator workloads.
Competitors & Alternatives
Intel Xeon 698X
- AMD Ryzen Threadripper PRO 9995WXRival
Workstation
- AMD Ryzen Threadripper PRO 7995WXRival
Workstation
- AMD Ryzen Threadripper 9970XRival
HEDT/Workstation
- Compare head-to-headIntel Xeon w9-3595XRival
Workstation (previous gen)
- Intel Xeon 696XRival
Workstation (same gen, lower core count)
Intel Xeon w9-3595X
- AMD Ryzen Threadripper PRO 7995WXRival
Workstation
- AMD Ryzen Threadripper PRO 5995WXRival
Workstation
- AMD Ryzen Threadripper PRO 7985WXRival
Workstation
- Intel Xeon w9-3495XRival
Workstation
- Compare head-to-headIntel Core i9-14900KRival
High-End Desktop
- AMD Ryzen Threadripper 7960XAlt
Strong multi-threaded performance on TRX50 with lower cost if you can forgo WRX90 enterprise features.
Our Verdict on Each
An extremely powerful workstation CPU with best-in-class core count, memory capacity, and I/O for the Xeon 600 platform, best suited for professional workflows that can saturate its 86 cores and 128 PCIe lanes.
Best for: Professional workstations for rendering, simulation, AI development, or data processing that can leverage 86 cores, eight-channel memory, and 128 PCIe 5.0 lanes in a single socket.
Read the full reviewA top-end workstation processor with massive core count and I/O expansion, ideal for well-threaded pro workloads, but it demands serious power and cooling and faces strong competition from AMD’s Threadripper PRO line.
Best for: Professional workstations for rendering, simulation, AI development, or multi-GPU setups where Intel’s platform features and software ecosystem are preferred.
Read the full reviewFrequently Asked Questions
Which is better, Intel Xeon 698X or Intel Xeon w9-3595X?
Based on our editorial ratings, the Intel Xeon 698X comes out ahead with a score of 9/10. That said, the best choice depends on your workload — check the spec and performance breakdown above for gaming, productivity and efficiency differences.
Which is faster for gaming, Intel Xeon 698X or Intel Xeon w9-3595X?
For gaming, the Intel Xeon 698X leads with a gaming performance score of 0/100 among Intel Xeon 698X and Intel Xeon w9-3595X.
Which uses less power?
The Intel Xeon 698X has the lowest rated TDP. Power draw across these chips: Intel Xeon 698X (350 W), Intel Xeon w9-3595X (385 W).
Do Intel Xeon 698X and Intel Xeon w9-3595X use the same socket?
No. They use different sockets (Intel Xeon 698X: FCLGA4710, Intel Xeon w9-3595X: FCLGA4677), so each needs a compatible motherboard.
Which has more cores?
The Intel Xeon 698X has the most cores. Core counts: Intel Xeon 698X (86 cores), Intel Xeon w9-3595X (60 cores).
Which is faster in multi-core benchmarks?
The Intel Xeon 698X posts the highest multi-core benchmark score. Multi-core results: Intel Xeon 698X (0). Benchmark figures are approximate and workload-dependent.