AMD EPYC 9734 vs Intel Xeon w9-3595X Comparison
AMD EPYC 9734
Xeon w9-3595X
PERFORMANCE BENCHMARKS
Analysis: AMD EPYC 9734 vs Intel Xeon w9-3595X
The Verdict
The benchmark data is unambiguous: the AMD EPYC 9734 is the dominant processor in this comparison, winning 14 of 16 head-to-head tests. The Intel Xeon w9-3595X takes only two wins, both in single-threaded workloads. For anyone prioritizing raw multi-threaded throughput, data compression, encryption, or math-heavy integer work, the EPYC 9734 is the clear choice. The Xeon w9-3595X is the pick only for workloads that are strictly latency-sensitive and single-threaded, where its PassMark single-thread score of 3720 versus 2310 represents a 37.9% advantage. Everything else points to the AMD part.
The average benchmark scores reinforce this split. The EPYC 9734 posts an average benchmark score of 310619, while the Xeon w9-3595X sits at 209881. That is a substantial gap. The EPYC also holds a 99th percentile ranking among all CPUs, matching the Xeon's own 99th percentile placement, but the EPYC achieves that standing while winning nearly every category. The verdict is straightforward: choose the EPYC 9734 for nearly all server and workstation compute, and reserve the Xeon w9-3595X for edge cases where per-thread responsiveness beats aggregate throughput.
Where Each One Wins
The AMD EPYC 9734 wins every multi-threaded benchmark in the comparison. In Cinebench R15, R20, and R23 multicore tests, it leads by 3.1% in each case. The margin is consistent, roughly three points across all three Cinebench versions, which suggests a stable architectural advantage rather than a workload-specific fluke. The PassMark multithread test shows a narrower lead of only 2.7%, but that is still a win.
The EPYC's biggest victories come in data-heavy and math-heavy workloads. PassMark data compression shows a 58.3% lead, data encryption a 94.5% lead, integer math a 73.8% lead, and random string sorting an 87.5% lead. Floating point math is 44.9% ahead, extended instructions 44.2% ahead, and find prime numbers 42.9% ahead. Physics simulation is 15.5% ahead. These are not marginal differences; they are category-defining gaps.
The Intel Xeon w9-3595X wins only the two single-thread tests, both recorded as PassMark single-thread and singlethread with identical scores of 3720 versus 2310, a 37.9% margin. That is a genuine strength, but it is isolated. No other benchmark in the recorded data favors Intel.
Architecture Differences
The two processors take fundamentally different design paths. The AMD EPYC 9734 is built on Zen 4 architecture with the Bergamo codename, fabricated on a 5 nm process at TSMC. It packs 112 cores and 224 threads. The Intel Xeon w9-3595X uses Sapphire Rapids silicon on Intel's 10 nm process, with 60 cores and 120 threads. The EPYC has nearly double the core count, which explains its multi-threaded dominance.
Cache configurations differ sharply. The EPYC 9734 carries 64 KB of L1 per core, 1 MB of L2 per core, and a shared 256 MB L3. The Xeon w9-3595X has 80 KB of L1 per core, 2 MB of L2 per core, and 112.5 MB of L3. The Xeon gives each core more private cache, but the EPYC's total L3 is more than double. Transistor counts and die sizes reflect the different approaches: the EPYC uses 71,000 million transistors across 8x 73 mm² chiplets, while the Xeon is built from 4x 477 mm² dies with no transistor figure recorded.
Memory subsystems diverge as well. Both support DDR5 and ECC, but the EPYC 9734 uses a twelve-channel memory bus with 460.8 GB/s bandwidth, while the Xeon w9-3595X uses an eight-channel bus with 307.2 GB/s. PCIe lanes also differ: the EPYC provides Gen 5 with 128 lanes, the Xeon provides Gen 5 with 112 lanes. The Xeon has an unlocked multiplier and lists integrated graphics as N/A, while the EPYC lists no integrated graphics and has a locked multiplier.
FAQ
Q: Which processor is faster in multi-threaded workloads?
A: The AMD EPYC 9734 wins all multi-threaded benchmarks in the database, including Cinebench R23 multicore with 86943 points versus 84304 for the Intel Xeon w9-3595X, a 3.1% lead.
Q: Is the Intel Xeon w9-3595X better at anything?
A: Yes, it wins single-threaded tests. Its PassMark single-thread score is 3720 versus 2310 for the EPYC 9734, a 37.9% advantage.
Q: How do their core counts compare?
A: The AMD EPYC 9734 has 112 cores and 224 threads, while the Intel Xeon w9-3595X has 60 cores and 120 threads.
Q: Which processor has more cache?
A: The EPYC 9734 has 256 MB of shared L3 cache versus 112.5 MB on the Xeon. The Xeon has more per-core L1 and L2, with 80 KB and 2 MB per core respectively, while the EPYC has 64 KB and 1 MB per core.
Q: What are their memory bandwidth figures?
A: The EPYC 9734 offers 460.8 GB/s over a twelve-channel DDR5 bus, while the Xeon w9-3595X offers 307.2 GB/s over an eight-channel DDR5 bus.
Q: Which processor was released more recently?
A: The Intel Xeon w9-3595X was released on 2024-08-23, while the AMD EPYC 9734 was released on 2023-06-12.
Head-to-Head Benchmarks
The head-to-head data shows a lopsided contest. In Cinebench R15 multicore, the EPYC 9734 scores 8763 against the Xeon's 8497, a 3.1% win. Single-core R15 goes to the EPYC as well, 1237 versus 1199, also 3.1%. Cinebench R20 multicore repeats the pattern: 36516 versus 35407, another 3.1% margin. Single-core R20 is 5155 versus 4998, again 3.1%. Cinebench R23 multicore gives 86943 versus 84304, 3.1%. The consistency is striking, the EPYC holds a steady advantage across every Cinebench iteration.
The PassMark suite tells a different story, with far larger deltas. Data compression shows the EPYC at 2900008 versus 1831962, a 58.3% blowout. Data encryption is even more extreme: 179390 versus 92249, a 94.5% lead. Extended instructions come in at 205925 versus 142785, 44.2% ahead. Find prime numbers is 829 versus 580, a 42.9% margin. Floating point math scores 549045 against 379008, 44.9% ahead. Integer math is the second-largest gap at 823150 versus 473507, a 73.8% lead. Multithread is the closest PassMark result, 102286 versus 99576, just 2.7%. Physics goes 6747 versus 5842, 15.5% ahead. Random string sorting is the largest single margin at 357638 versus 190745, an 87.5% lead.
The only Intel wins are the two identical single-thread tests. PassMark single-thread and singlethread both record 3720 for the Xeon and 2310 for the EPYC, a 37.9% deficit for AMD. Those two tests are the entire Intel highlight reel. The database records 14 wins for AMD and 2 for Intel.
Specification Differences
The differences between these two processors are extensive. The AMD EPYC 9734 belongs to the EPYC 9004 series with Zen 4 architecture and the Bergamo codename, while the Intel Xeon w9-3595X is part of the Xeon W family with the Sapphire Rapids codename. The EPYC uses a 5 nm TSMC process, the Xeon a 10 nm Intel process. Core counts are 112 versus 60, threads 224 versus 120. Base clocks are 2.20 GHz for the EPYC and 2.00 GHz for the Xeon, but boost clocks reverse the order: 3.00 GHz for the EPYC, 4.80 GHz for the Xeon. TDP differs, with the EPYC at 340 and the Xeon at 385.
Sockets are incompatible: AMD Socket SP5 for the EPYC, Intel Socket 4677 for the Xeon. The EPYC uses 8x 73 mm² dies with 71,000 million transistors, while the Xeon uses 4x 477 mm² dies. L1 cache is 64 KB per core on the EPYC versus 80 KB per core on the Xeon. L2 is 1 MB per core versus 2 MB per core. L3 is 256 MB shared versus 112.5 MB. Memory channels are twelve for AMD, eight for Intel. Memory bandwidth is 460.8 GB/s versus 307.2 GB/s. PCIe lanes are 128 for the EPYC and 112 for the Xeon, both Gen 5. The Xeon has an unlocked multiplier, the EPYC does not. The EPYC lists a part number of 100-000001235, the Xeon uses SRN71. Release dates differ by over a year, with the EPYC arriving first.