AMD EPYC 9015 vs Intel Xeon Gold 6338T Comparison

Where Each One Wins

The benchmark data splits these two server processors into clearly different usage profiles. The AMD EPYC 9015 wins 4 of the 11 head-to-head tests, while the Intel Xeon Gold 6338T wins 7. That split is not random; it tracks the workloads each chip was built to accelerate.

The AMD EPYC 9015 dominates in single-threaded and latency-sensitive tasks. Its PassMark single-thread score of 3265 beats the Intel part's 2348 by 39.1%. That gap is decisive for database queries, web serving front-ends, and any workload where a single core must finish a task before the next one starts. The EPYC 9015 also crushes the Xeon in prime number finding (304 vs 118, a 157.6% advantage) and in the PassMark physics test (2893 vs 1160, a 149.4% edge). These are classic integer-heavy, branch-prediction-sensitive operations where the newer Zen 5 core design shows its strength.

The Intel Xeon Gold 6338T, by contrast, is a throughput monster. Its 24 cores and 48 threads give it a massive parallel processing advantage over the 8-core, 16-thread EPYC 9015. In multithreaded workloads, the Xeon leads by 14.3% (35801 vs 30689). That lead expands dramatically in specific parallel tasks: integer math goes 152275 vs 92524 (39.2% ahead), floating point math goes 90584 vs 61615 (32% ahead), and data compression goes 557816 vs 353014 (36.7% ahead). The Xeon also wins in data encryption (24677 vs 17790, 27.9% ahead), extended instructions (43292 vs 27970, 35.4% ahead), and random string sorting (66495 vs 39772, 40.2% ahead).

The practical interpretation is straightforward. For single-socket servers running mixed workloads with heavy per-core demands, the EPYC 9015's architectural efficiency per thread makes it the stronger pick. For batch processing, data analytics, virtualization hosts with many light VMs, or any environment where core count translates directly into parallel throughput, the Xeon Gold 6338T's 3x core advantage is the deciding factor.

Architecture Differences

The two processors come from different design generations and manufacturing approaches. The AMD EPYC 9015 uses the Zen 5 architecture, codenamed Turin, built on a 4 nm process at TSMC. It packs 16,630 million transistors across a die size of 2x 70.6 mm². The Intel Xeon Gold 6338T uses the older Ice Lake-SP architecture, built on Intel's 10 nm process, with no transistor or die size figures recorded in the database.

Core counts differ sharply: the EPYC 9015 has 8 cores and 16 threads, while the Xeon Gold 6338T has 24 cores and 48 threads. The EPYC makes up for its lower core count with much higher clock speeds: 3.60 GHz base and 4.10 GHz boost versus the Xeon's 2.10 GHz base and 3.40 GHz boost. The EPYC also has a lower TDP at 125 W compared to the Xeon's 165 W, despite being on a smaller process node.

Cache hierarchies reflect different design philosophies. The EPYC 9015 allocates 80 KB of L1 per core and 1 MB of L2 per core, with a large 64 MB shared L3. The Xeon Gold 6338T has 64 KB L1 per core, 1 MB L2 per core, but only 36 MB of shared L3. The EPYC's larger L3 cache gives it a significant advantage for workloads that repeatedly access the same data set.

Memory support is a major differentiator. The EPYC 9015 supports DDR5 memory over a twelve-channel bus, delivering 576.0 GB/s of memory bandwidth. The Xeon Gold 6338T uses DDR4 over eight channels, capping at 204.8 GB/s. That is nearly three times the memory bandwidth in favor of the AMD part, which matters for memory-bound server workloads.

PCIe capabilities also differ. The EPYC 9015 offers Gen 5 with 128 lanes (CPU only), while the Xeon Gold 6338T offers Gen 4 with 64 lanes (CPU only). The AMD platform doubles the lane count and moves to the newer standard, enabling faster I/O for NVMe storage and accelerators. Both processors support ECC memory, and both are in active production. The EPYC 9015 launched on 2024-10-09, while the Xeon Gold 6338T launched on 2021-04-05.

Head-to-Head Benchmarks

The single-thread test is the EPYC 9015's strongest statement. It scores 3265 against the Xeon's 2348, a 39.1% advantage. This is not a small margin; it reflects a fundamental per-core performance gap that no amount of core count can fix for latency-bound work.

The prime number finding test shows the most extreme delta. The EPYC 9015 scores 304 versus the Xeon's 118, a 157.6% advantage. This test is highly sensitive to branch prediction and integer execution efficiency, both strengths of the Zen 5 core. Similarly, the physics test shows the EPYC at 2893 versus 1160, a 149.4% edge. Both tests reward fast single-core execution over raw parallel throughput.

The Xeon Gold 6338T's biggest win comes in random string sorting, where it scores 66495 versus 39772, a 40.2% lead. This test benefits from the Xeon's higher thread count and memory parallelism. Integer math shows the Xeon at 152275 versus 92524, a 39.2% advantage, and data compression shows 557816 versus 353014, a 36.7% lead. Extended instructions favor the Xeon 43292 to 27970, a 35.4% gap, while floating point math goes 90584 to 61615, a 32% edge.

The multithread test is closer than the core count suggests. The Xeon wins 35801 to 30689, only 14.3% ahead, despite having three times the cores. This indicates that the EPYC 9015's per-thread performance is so strong that it partially compensates for the core deficit. Data encryption also favors the Xeon, 24677 to 17790, a 27.9% margin.

Overall, the EPYC 9015's average benchmark score is 57555, while the Xeon Gold 6338T averages 60572. Both sit at the 92nd percentile among all CPUs in the database. The Xeon's nearest rival is the Intel Core i9-13900 at 60676 (0.2% higher), while the EPYC's nearest rival is the Intel Core i9-14900 at 58115 (1% higher). The data shows two very different strategies to reach similar overall performance.

FAQ

Q: Which processor has higher single-thread performance?

A: The AMD EPYC 9015 wins single-thread tests decisively. Its PassMark single-thread score of 3265 beats the Intel Xeon Gold 6338T's 2348 by 39.1%.

Q: Which processor is better for heavily multithreaded workloads?

A: The Intel Xeon Gold 6338T leads in multithreaded performance. It scores 35801 in the PassMark multithread test versus 30689 for the EPYC 9015, a 14.3% advantage.

Q: How do the two compare in memory bandwidth?

A: The AMD EPYC 9015 supports DDR5 memory over a twelve-channel bus, delivering 576.0 GB/s. The Intel Xeon Gold 6338T uses DDR4 over eight channels, providing 204.8 GB/s.

Q: What are the core and thread counts for each processor?

A: The AMD EPYC 9015 has 8 cores and 16 threads. The Intel Xeon Gold 6338T has 24 cores and 48 threads.

Q: Which processor has more L3 cache?

A: The AMD EPYC 9015 has 64 MB of shared L3 cache. The Intel Xeon Gold 6338T has 36 MB of shared L3 cache.

Q: What are the PCIe capabilities of each processor?

A: The AMD EPYC 9015 supports PCIe Gen 5 with 128 lanes (CPU only). The Intel Xeon Gold 6338T supports PCIe Gen 4 with 64 lanes (CPU only).

Specification Differences

| Specification | AMD EPYC 9015 | Intel Xeon Gold 6338T |

|---|---|---|

| Cores | 8 | 24 |

| Threads | 16 | 48 |

| Base Clock | 3.60 GHz | 2.10 GHz |

| Boost Clock | 4.10 GHz | 3.40 GHz |

| TDP | 125 W | 165 W |

| Socket | AMD Socket SP5 | Intel Socket 4189 |

| Architecture | Zen 5 | Ice Lake |

| Codename | Turin | Ice Lake-SP |

| Process Node | 4 nm | 10 nm |

| Foundry | TSMC | Intel |

| Transistors | 16,630 million | Not recorded |

| Die Size | 2x 70.6 mm² | Not recorded |

| L1 Cache | 80 KB (per core) | 64 KB (per core) |

| L2 Cache | 1 MB (per core) | 1 MB (per core) |

| L3 Cache | 64 MB (shared) | 36 MB (shared) |

| Memory Support | DDR5 | DDR4 |

| Memory Bus | Twelve-channel | Eight-channel |

| Memory Bandwidth | 576.0 GB/s | 204.8 GB/s |

| PCIe | Gen 5, 128 Lanes (CPU only) | Gen 4, 64 Lanes (CPU only) |

| Release Date | 2024-10-09 | 2021-04-05 |

| Launch MSRP | $527 | Not recorded |

| Part Number | 100-000001553 | Not recorded |