AMD Ryzen 9 PRO 9945 vs Intel Xeon 654 Comparison

The AMD Ryzen 9 PRO 9945 and the Intel Xeon 654 occupy the same performance percentile tier, both sitting at the 96th percentile against all CPUs, but they achieve that status through starkly different design philosophies. The AMD part, a 12-core/24-thread Granite Ridge chip on a 4nm TSMC process, posts an average benchmark score of 96083, while the Intel Xeon 654, an 18-core/36-thread Granite Rapids-WS part on Intel's 5nm process, averages 90717. Despite the Intel chip's higher core count, the AMD processor holds a 5.6% advantage in average score, a lead driven entirely by its overwhelming single-thread dominance, as the head-to-head data reveals a 22.3% gap in the PassMark single-thread test (4619 vs 3778).

Head-to-Head Benchmarks

The most lopsided result in the entire comparison is the PassMark physics test, where the Intel Xeon 654 scores 5596 against the AMD Ryzen 9 PRO 9945's 2902, a delta of -48.1% for the AMD part. This is the single largest margin of victory for either processor, and it is more than double the next-largest gap. The floating-point math test also heavily favors Intel, with the Xeon 654 scoring 163093 versus 107340 for the AMD, a 34.2% advantage. These two results indicate that the Intel chip's additional six cores and twelve threads are being effectively utilized in workloads that scale with parallel execution and heavy mathematical throughput.

Intel extends its winning streak across the remaining multithreaded and data-centric benchmarks, though the margins vary considerably. The extended instructions test shows a 30.2% lead for Intel (63539 vs 44374), while data compression comes in at a 29.2% advantage (818902 vs 579972). Data encryption shows a 25.1% gap (40675 vs 30450), and random string sorting is 24.6% in Intel's favor (82828 vs 62458). The multithread test itself, a broad measure of parallel performance, gives Intel a 20.7% lead (61353 vs 48664). Integer math is closer, with Intel ahead by 17.6% (207745 vs 171175), and even the find prime numbers test, which often favors AMD's architecture, goes to Intel by 12.3% (390 vs 342).

The AMD Ryzen 9 PRO 9945 has exactly one winning category, but it is a decisive one. The PassMark single-thread test, reported identically as both "single_thread" and "singlethread," gives AMD a 22.3% victory with a score of 4619 against Intel's 3778. This is the only benchmark in the head-to-head set where AMD wins, and it is a critical one for responsiveness and lightly threaded applications. The overall win tally is 9 for Intel and 2 for AMD, but the single-thread result is the one that explains the average score discrepancy, as AMD's superior per-core performance compensates for its lower core count in the averaging formula.

Where Each One Wins

The Intel Xeon 654 is the clear choice for throughput-oriented workloads that can utilize all 18 cores and 36 threads. The data shows its biggest wins in physics simulation (48.1% ahead), floating-point math (34.2% ahead), and extended instruction sets (30.2% ahead). These are the workloads typical of scientific computing, financial modeling, and video rendering, where parallel execution across many cores directly translates to higher scores. The data compression and encryption results, with Intel leading by 29.2% and 25.1% respectively, further cement its position for database operations, file servers, and any task involving large data sets that need to be processed or secured.

The AMD Ryzen 9 PRO 9945 wins exclusively in single-thread performance, with a 22.3% advantage over Intel. This makes it the better option for legacy applications, single-threaded scripts, and any workload that cannot be easily parallelized. The absolute single-thread score of 4619 is notable because it places AMD in a performance tier that Intel cannot reach with this particular chip, despite Intel's higher core count. For users running a mix of lightly threaded and heavily threaded tasks, the AMD part's average score of 96083 versus Intel's 90717 suggests that the single-thread advantage more than compensates for the multithread deficit in the aggregate benchmark. However, for purely parallel workloads, Intel's multithread score of 61353 against AMD's 48664 is the definitive metric.

Architecture Differences

The two processors are built on fundamentally different platforms. The AMD Ryzen 9 PRO 9945 uses the Granite Ridge architecture, which is part of the Zen 5 family, and is manufactured on a 4nm process at TSMC. It features 12 cores and 24 threads, with a base clock of 3.40 GHz and a boost clock of 5.40 GHz. The Intel Xeon 654 uses the Granite Rapids architecture, part of the Xeon 600 series, and is manufactured on a 5nm process at Intel. It has 18 cores and 36 threads, with a base clock of 3.10 GHz and a boost clock of 4.80 GHz. The AMD chip has a 65W TDP, while the Intel chip has a 200W TDP, a significant difference that reflects the Intel part's higher core count but also its higher power envelope.

Cache configurations also differ substantially. The AMD part has 80 KB of L1 cache per core, 1 MB of L2 cache per core, and 64 MB of L3 cache. The Intel part has 112 KB of L1 per core, 2 MB of L2 per core, and 72 MB of shared L3 cache. The larger L2 and L3 caches on the Intel chip likely contribute to its strong performance in data compression and encryption tasks, which benefit from larger working sets being held closer to the cores. The AMD chip's cache is smaller but its memory bandwidth is lower as well, at 89.6 GB/s over dual-channel DDR5, versus the Intel chip's 409.6 GB/s over eight-channel DDR5. This is a massive 4.6x difference in memory bandwidth and is likely a major factor in the Intel chip's physics and floating-point math victories, which often saturate memory bandwidth.

The platforms diverge further in connectivity and features. The AMD part uses the AMD Socket AM5 and supports PCIe Gen 5 with 24 lanes from the CPU, and it includes integrated Radeon Graphics. The Intel part uses the Intel Socket 4710 and supports PCIe Gen 5 with 128 lanes from the CPU, which is more than five times the lane count, and has no integrated graphics. Both support ECC memory, and both are active production parts. The AMD part has a locked multiplier, while the Intel part has an unlocked multiplier, meaning the Intel chip can be overclocked. The Intel part has a launch MSRP of $1199, while the AMD part has no listed launch MSRP. The release dates are also different, with the AMD part released on 2025-09-15 and the Intel part on 2026-02-01.

FAQ

Q: Which processor has the higher single-thread performance?

A: The AMD Ryzen 9 PRO 9945, with a PassMark single-thread score of 4619, which is 22.3% higher than the Intel Xeon 654's score of 3778.

Q: What is the biggest performance gap between the two in any benchmark?

A: The PassMark physics test, where the Intel Xeon 654 scores 5596 versus the AMD's 2902, a 48.1% difference.

Q: How do the core and thread counts compare?

A: The Intel Xeon 654 has 18 cores and 36 threads, while the AMD Ryzen 9 PRO 9945 has 12 cores and 24 threads. The Intel part also has a 200W TDP versus 65W for the AMD part.

Q: Which processor offers more PCIe lanes?

A: The Intel Xeon 654 provides 128 PCIe Gen 5 lanes from the CPU, while the AMD Ryzen 9 PRO 9945 provides 24 lanes.

Q: Do both processors support ECC memory?

A: Yes, both the AMD Ryzen 9 PRO 9945 and the Intel Xeon 654 support ECC memory.

Q: What is the average benchmark score for each, and how do they rank against their nearest rivals?

A: The AMD Ryzen 9 PRO 9945 has an average score of 96083, which is 0.3% ahead of the AMD EPYC 4565P and 0.5% ahead of the AMD EPYC 9175F. The Intel Xeon 654 has an average score of 90717, which is 0.2% ahead of the AMD Ryzen AI Max+ 392 and 2.9% ahead of the Intel Xeon w7-2575X.

The Verdict

The data presents a clear split: the Intel Xeon 654 is the superior processor for heavily multithreaded, data-intensive workloads, winning 9 of the 11 head-to-head benchmarks. Its 20.7% lead in multithread performance, combined with massive margins in physics (48.1%) and floating-point math (34.2%), makes it the appropriate choice for server and workstation tasks that require maximum parallel throughput. The 18-core/36-thread configuration, along with eight-channel memory support delivering 409.6 GB/s, is a decisive advantage in environments where memory bandwidth and core count are the primary constraints.

The AMD Ryzen 9 PRO 9945 is the better option for single-threaded and lightly threaded applications, where its 22.3% single-thread advantage over Intel gives it a clear edge. Its higher average benchmark score of 96083 versus Intel's 90717 indicates that, in a mixed workload environment, the AMD part's per-core strength can offset its lower core count. The 65W TDP also makes it a more power-efficient choice for those not requiring the Intel chip's parallel performance. For users who prioritize raw multithreaded compute, the Intel Xeon 654 is the data-backed winner; for those who need maximum single-thread responsiveness and overall average performance, the AMD Ryzen 9 PRO 9945 is the superior pick.