AMD Ryzen 9 PRO 9945 vs Intel Xeon 658X Comparison

Where Each One Wins

The benchmark data splits this comparison into two very distinct profiles. The Intel Xeon 658X dominates across every multithreaded and throughput-oriented workload in the recorded results, while the AMD Ryzen 9 PRO 9945 takes the single-thread crown. Out of the 11 head-to-head benchmark entries, Intel wins 9 and AMD wins 2, with both AMD victories being the identical single-thread test recorded under slightly different naming conventions.

The Xeon 658X is clearly built for heavy parallel workloads. Its wins span data compression, encryption, extended instruction sets, prime number calculations, floating-point math, integer math, multithreaded throughput, physics simulation, and random string sorting. The margins are substantial across the board, ranging from 51% in multithread to a massive 122.9% in physics. This is a processor that scales with thread count and memory bandwidth, which makes sense given its 24 cores and 48 threads versus the Ryzen's 12 cores and 24 threads.

The Ryzen 9 PRO 9945, by contrast, wins only in single-thread performance. Its score of 4619 beats the Xeon's 3728 by 19.3%. This suggests the Ryzen's higher boost clock of 5.40 GHz against the Xeon's 4.90 GHz gives it a meaningful edge in latency-sensitive, lightly threaded tasks. The Ryzen also benefits from a higher base clock of 3.40 GHz versus 3.00 GHz on the Xeon.

Architecture Differences

The two processors come from fundamentally different design philosophies. The Intel Xeon 658X uses Granite Rapids architecture on a 5 nm process, fabricated by Intel. It features 24 cores and 48 threads with an L1 cache of 112 KB per core, L2 of 2 MB per core, and a massive 144 MB shared L3 cache. The die is large at 2x 598 mm², and it supports eight-channel DDR5 memory with a bandwidth of 409.6 GB/s. It also provides 128 PCIe Gen 5 lanes from the CPU alone. This is a server/workstation part on Intel Socket 4710, with a launch MSRP of $1699.

The AMD Ryzen 9 PRO 9945 uses the Granite Ridge codename, part of the Zen 5 generation, on a 4 nm process from TSMC. It has 12 cores and 24 threads, with L1 at 80 KB per core, L2 at 1 MB per core, and 64 MB of L3. The die is much smaller at 2x 70.6 mm², and the chip contains 16,630 million transistors. Memory support is dual-channel DDR5 with 89.6 GB/s bandwidth, and it provides 24 PCIe Gen 5 lanes. It includes Radeon Graphics as integrated graphics, which the Xeon lacks entirely. The Ryzen sits on AMD Socket AM5 and has a locked multiplier.

The cache hierarchy differences are notable. The Xeon's 144 MB shared L3 dwarfs the Ryzen's 64 MB, and its per-core L2 is double the Ryzen's. This helps explain the Xeon's dominance in data-heavy parallel workloads. The memory bandwidth gap is even more pronounced: 409.6 GB/s versus 89.6 GB/s, a factor of roughly 4.6. For workloads that stream large datasets, this difference alone can determine the winner.

Head-to-Head Benchmarks

The largest single margin in the entire comparison comes in physics simulation, where the Xeon scores 6470 against the Ryzen's 2902, a 122.9% advantage. This workload appears to be extremely sensitive to both core count and memory throughput, and the Xeon's architecture clearly excels there.

Floating-point math shows the Xeon at 210480 versus 107340, a 96.1% lead. Extended instructions follow closely with 84626 against 44374, a 90.7% gap. Prime number finding sees 649 versus 342, an 89.8% difference. These results suggest the Xeon's wider execution resources and larger cache hierarchy translate directly into higher throughput for compute-intensive operations.

Data compression shows 1062062 versus 579972, an 83.1% advantage for Intel. Encryption follows at 52357 versus 30450, a 71.9% gap. Random string sorting yields 103028 versus 62458, a 65% margin. Even integer math, often considered more balanced across architectures, shows a 54.2% lead for the Xeon with 263995 versus 171175.

The multithreaded score of 73490 versus 48664 represents a 51% advantage. This is the most general-purpose measure of parallel performance, and it confirms the Xeon's overall throughput superiority. The Ryzen's only victory, single-thread at 4619 versus 3728, is a 19.3% edge. Notably, the Xeon's average benchmark score of 116060 versus the Ryzen's 96083, combined with the Xeon sitting at the 97th percentile versus the Ryzen's 96th, places the Intel part slightly higher in overall standing among all CPUs in the database.

The Verdict

The data presents a clear choice based on workload type. For any multithreaded, throughput-oriented task, the Intel Xeon 658X is the definitive winner. Its advantages range from 51% to 122.9% across the parallel benchmarks, and it offers 144 MB of shared L3, eight-channel memory at 409.6 GB/s, and 24 cores with 48 threads. Anyone running render farms, scientific simulations, data compression pipelines, or heavy encryption workloads should look at the Xeon first.

The AMD Ryzen 9 PRO 9945 wins only in single-thread performance, with a 19.3% edge. Its higher boost clock of 5.40 GHz and 4 nm process from TSMC give it an advantage in latency-sensitive applications that cannot leverage multiple cores. For code compilation with poor parallelism, certain legacy applications, or interactive workloads where a single thread is the bottleneck, the Ryzen is the better choice.

The Ryzen also offers integrated Radeon Graphics, which the Xeon does not, and a much lower TDP of 65 watts versus 250 watts. However, the data shows that the Xeon's performance advantages come at the cost of significantly higher power draw and a much larger physical package. The Xeon's launch MSRP is $1699. The Ryzen's market segment is also server/workstation, but its AM5 socket and dual-channel memory suggest a more modest platform footprint.

For buyers whose primary concern is raw parallel throughput, the Xeon is the clear pick. For those who need maximum single-thread responsiveness and can accept lower parallel performance, the Ryzen is the data-backed choice. The 9-to-2 win ratio in benchmarks strongly favors Intel for general compute, but the single-thread gap is real and should not be ignored.

FAQ

Q: Which processor wins in multithreaded workloads?

A: The Intel Xeon 658X wins in all parallel benchmarks, with its largest margin being 122.9% in physics and its smallest still being 51% in multithread score.

Q: Does the AMD Ryzen 9 PRO 9945 win any benchmark?

A: Yes, it wins the single-thread test with a score of 4619 versus the Xeon's 3728, a 19.3% advantage.

Q: How do the memory systems compare?

A: The Xeon supports eight-channel DDR5 with 409.6 GB/s bandwidth, while the Ryzen supports dual-channel DDR5 with 89.6 GB/s bandwidth.

Q: What are the core counts?

A: The Intel Xeon 658X has 24 cores and 48 threads. The AMD Ryzen 9 PRO 9945 has 12 cores and 24 threads.

Q: Which processor has more L3 cache?

A: The Xeon has 144 MB of shared L3 cache, while the Ryzen has 64 MB of L3.

Q: What is the difference in process node?

A: The Xeon uses a 5 nm process from Intel, while the Ryzen uses a 4 nm process from TSMC.

Specification Differences

| Specification | Intel Xeon 658X | AMD Ryzen 9 PRO 9945 |

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

| Cores | 24 | 12 |

| Threads | 48 | 24 |

| Base Clock | 3.00 GHz | 3.40 GHz |

| Boost Clock | 4.90 GHz | 5.40 GHz |

| TDP | 250 W | 65 W |

| Socket | Intel Socket 4710 | AMD Socket AM5 |

| Process Node | 5 nm | 4 nm |

| Foundry | Intel | TSMC |

| Die Size | 2x 598 mm² | 2x 70.6 mm² |

| L1 Cache | 112 KB (per core) | 80 KB (per core) |

| L2 Cache | 2 MB (per core) | 1 MB (per core) |

| L3 Cache | 144 MB (shared) | 64 MB |

| Memory Bus | Eight-channel | Dual-channel |

| Memory Bandwidth | 409.6 GB/s | 89.6 GB/s |

| PCIe Lanes | Gen 5, 128 Lanes (CPU only) | Gen 5, 24 Lanes (CPU only) |

| Integrated Graphics | N/A | Radeon Graphics |

| Multiplier Unlocked | Yes | No |

| Transistors | Not specified | 16,630 million |

| Release Date | 2026-02-01 | 2025-09-15 |

| Launch MSRP | $1699 | Not specified |