AMD EPYC 7513 vs Intel Xeon 654 Comparison

The AMD EPYC 7513 and Intel Xeon 654 occupy distinctly different positions in the server and workstation landscape, despite sharing a 200 W TDP. The recorded data shows a clear split: the Intel Xeon 654 wins 13 of the 17 head-to-head benchmarks, while the AMD EPYC 7513 takes 4, but the AMD part’s victories are often massive, suggesting workload-specific dominance rather than overall superiority. The database places the AMD EPYC 7513 at the 97th percentile among all CPUs, with the Intel Xeon 654 at the 96th percentile, indicating both are near the top of the performance stack.

Where Each One Wins

The Intel Xeon 654 is the consistent winner across rendering and general compute workloads. In every Cinebench test, from R15 to R23, the Intel part leads by a narrow margin of roughly 3.3% in both single-core and multi-core runs. The multi-core scores are telling: Cinebench R23 shows 52,150 for Intel versus 50,431 for AMD, and the Intel part also takes PassMark’s multithread test at 61,353 versus 59,331. Single-thread performance is where Intel pulls ahead decisively, with a 34.4% advantage in PassMark single-thread (3,778 versus 2,479). This pattern extends to floating-point math, where Intel leads by 7% (163,093 versus 151,713), and to physics simulations, where the margin is 8.5% (5,596 versus 5,118). Extended instruction workloads also favor Intel by 11.2% (63,539 versus 56,451).

The AMD EPYC 7513 wins where thread count and cache capacity translate into throughput for data-centric tasks. The largest margin is in data encryption, where AMD scores 63,628 against Intel’s 40,675, a 56.4% advantage. Integer math shows a 31% lead for AMD (272,145 versus 207,745), and random string sorting goes to AMD by 26.4% (104,660 versus 82,828). Data compression is another AMD win, with 932,240 versus 818,902, a 13.8% difference. These are not marginal edges; they indicate that the EPYC 7513’s 32 cores and 64 threads, combined with 128 MB of shared L3 cache, excel at workloads that involve large datasets and parallelizable integer operations. The data suggests AMD wins where memory bandwidth and cache capacity matter more than raw clock speed.

Architecture Differences

The two processors are built on fundamentally different designs. The AMD EPYC 7513 uses the Zen 3 architecture, codenamed Milan, on a 7 nm process from TSMC. It features 32 cores and 64 threads, with a base clock of 2.60 GHz and a boost clock of 3.65 GHz. The Intel Xeon 654 uses the Granite Rapids architecture on Intel’s 5 nm process, with 18 cores and 36 threads, but clocks much higher: 3.10 GHz base and 4.80 GHz boost. This explains the single-thread dominance of the Intel part, since it has nearly a 1.15 GHz boost advantage.

Cache configurations differ sharply. AMD provides 64 KB of L1 and 512 KB of L2 per core, with a shared 128 MB L3 cache. Intel gives each core 112 KB of L1 and 2 MB of L2, but the shared L3 is only 72 MB. The larger L3 on the AMD part is likely a factor in its data compression and encryption wins, as more data can reside on-die. Intel compensates with a much larger die: the Xeon 654 uses two dies of 598 mm² each, while the EPYC 7513 uses eight dies of 81 mm² each. The AMD chip has 33,200 million transistors, while Intel does not report a transistor count in the database.

Memory support also diverges. AMD uses DDR4 with an eight-channel bus and 204.8 GB/s of bandwidth. Intel uses DDR5, also eight-channel, but doubles the bandwidth to 409.6 GB/s. This is a major architectural difference, and it explains why the Intel part, despite fewer cores, stays competitive in multi-core rendering. PCIe connectivity is comparable, with both offering 128 lanes, but AMD is Gen 4 while Intel is Gen 5. Both support ECC memory, and neither has integrated graphics. The Intel part has an unlocked multiplier, while the AMD does not. Release dates are far apart: the EPYC 7513 launched in March 2021, while the Xeon 654 is dated February 2026.

The Verdict

The data points to a clear recommendation based on workload type. For single-threaded performance, the Intel Xeon 654 is the unambiguous choice. Its 34.4% lead in PassMark single-thread and consistent wins across all Cinebench single-core tests make it superior for lightly threaded applications, database queries that rely on one core, or any software that does not scale across many threads. The higher boost clock of 4.80 GHz versus 3.65 GHz is the obvious driver.

For multi-threaded integer and data-heavy workloads, the AMD EPYC 7513 wins despite the Intel part’s higher memory bandwidth. The 56.4% encryption advantage and 31% integer math lead are too large to ignore. If the workload involves compression, encryption, sorting large strings, or integer-heavy parallel processing, the AMD part’s 32 cores and 128 MB L3 cache deliver results that the Intel part cannot match. The Cinebench multi-core scores are close, with Intel ahead by about 3.3%, but that margin is small relative to the massive AMD wins in data-centric tasks.

The overall average benchmark score favors AMD slightly. The EPYC 7513 has an average score of 102,244 across the database, while the Xeon 654 sits at 90,717. This is interesting because the Intel part wins more individual tests, but the AMD part’s huge wins in encryption and integer math pull its average up. The nearest rivals for the AMD part include the Intel Xeon 6521P, which scores 105,845 on average, 3.4% higher. For the Intel Xeon 654, the closest rival is the AMD Ryzen AI Max+ 392 at 90,541, just 0.2% lower. This suggests that in the broader market, both parts are positioned near their respective performance tiers.

FAQ

Q: Which processor has more cores and threads?

A: The AMD EPYC 7513 has 32 cores and 64 threads, while the Intel Xeon 654 has 18 cores and 36 threads.

Q: Which CPU is faster in single-threaded workloads?

A: The Intel Xeon 654 wins all single-core Cinebench tests and leads PassMark single-thread by 34.4%, with a score of 3,778 versus 2,479 for AMD.

Q: What is the biggest performance gap between the two?

A: The largest delta is in PassMark data encryption, where the AMD EPYC 7513 scores 63,628 against Intel’s 40,675, giving AMD a 56.4% advantage.

Q: How do they compare in memory bandwidth?

A: The Intel Xeon 654 supports DDR5 with 409.6 GB/s, while the AMD EPYC 7513 uses DDR4 with 204.8 GB/s, giving Intel twice the memory bandwidth.

Q: Which CPU has a higher boost clock?

A: The Intel Xeon 654 boosts to 4.80 GHz, while the AMD EPYC 7513 boosts to 3.65 GHz.

Q: Are both processors still in production?

A: Yes, the database lists both the AMD EPYC 7513 and Intel Xeon 654 as Active in production status.

Head-to-Head Benchmarks

The Cinebench suite shows a consistent pattern. In Cinebench R15 multi-core, Intel scores 5,256 versus 5,083 for AMD, a 3.3% lead. Single-core R15 is 742 versus 717, a 3.4% margin. R20 multi-core is 21,903 against 21,181, again 3.3%. R20 single-core is 3,092 versus 2,989, also 3.3%. R23 multi-core is 52,150 versus 50,431, and R23 single-core is 7,362 versus 7,119, both at 3.3%. These narrow, consistent margins indicate that Intel’s higher clocks give it a steady edge in rendering, but not a dominant one.

PassMark tests reveal the true divergence. Data compression goes to AMD at 932,240 versus 818,902, a 13.8% win. Data encryption is the standout: AMD 63,628, Intel 40,675, a 56.4% blowout. Extended instructions favor Intel, 63,539 versus 56,451, an 11.2% margin. Find prime numbers is close, 390 versus 380, a 2.6% Intel win. Floating-point math goes to Intel at 163,093 versus 151,713, a 7% edge. Integer math is a major AMD win: 272,145 versus 207,745, a 31% lead. PassMark multithread gives Intel 61,353 versus 59,331, a 3.3% win. Physics favors Intel, 5,596 versus 5,118, an 8.5% margin. Random string sorting is another large AMD win: 104,660 versus 82,828, a 26.4% edge. The single-thread tests are the most lopsided in Intel’s favor, with 3,778 versus 2,479, a 34.4% difference.

The overall win count is 13 for Intel and 4 for AMD, but the magnitude of AMD’s wins in encryption and integer math suggests that for specific enterprise workloads, the EPYC 7513 is the stronger part. The Intel Xeon 654 is the better all-rounder, especially for mixed workloads that include single-threaded tasks and floating-point operations. The data does not support a universal winner; it supports a workload-specific choice.