AMD EPYC 7261 vs Intel Xeon E5-2628L v3 Comparison

FAQ

Q: How does the AMD EPYC 7261 compare to the Intel Xeon E5-2628L v3 in overall average benchmark score?

A: The AMD EPYC 7261 has an average benchmark score of 2740, while the Intel Xeon E5-2628L v3 scores 2724. This places the AMD part 0.6% ahead of the Intel Xeon in the nearestRivals data, where the EPYC 7261 is listed as a rival with a deltaPct of 0.6 against the Xeon.

Q: Which processor wins in single-core Cinebench R23 performance?

A: The AMD EPYC 7261 wins with a score of 1337, compared to the Intel Xeon E5-2628L v3’s 1329. The delta is 0.6%, a narrow margin that suggests near-parity in single-threaded workloads.

Q: What is the core and thread configuration difference between these two CPUs?

A: The AMD EPYC 7261 has 8 cores and 16 threads, while the Intel Xeon E5-2628L v3 has 10 cores and 20 threads. Despite having fewer cores, the AMD part wins every multicore benchmark in the head-to-head results, indicating higher per-core efficiency.

Q: Which processor consumes more power according to the TDP figures?

A: The AMD EPYC 7261 has a TDP of 170 watts, while the Intel Xeon E5-2628L v3 has a TDP of 75 watts. This is a substantial difference, with the Intel part drawing less than half the thermal design power of the AMD chip.

Q: Are both processors still in production?

A: No. The AMD EPYC 7261 is listed as "Active" in production status, while the Intel Xeon E5-2628L v3 is marked as "End-of-life." The Intel part was released earlier, on 2014-09-07, versus the AMD part’s release on 2018-06-13.

Q: What memory channels does each processor support?

A: The AMD EPYC 7261 supports eight-channel memory with a bandwidth of 170.6 GB/s. The Intel Xeon E5-2628L v3 supports quad-channel memory with a bandwidth of 59.7 GB/s. This gives the AMD part nearly three times the memory bandwidth.

Architecture Differences

The AMD EPYC 7261 is built on the Zen architecture with the codename Naples, part of the EPYC 7001 series. It uses a 14 nm process node manufactured by GlobalFoundries, with 4,800 million transistors on a 213 mm² die. In contrast, the Intel Xeon E5-2628L v3 uses the Haswell architecture with the codename Haswell-EP, built on a 22 nm process node at Intel, with 2,600 million transistors on a 356 mm² die. The AMD chip is smaller in die size but packs nearly double the transistor count, reflecting the denser 14 nm process.

Cache configurations differ significantly. The EPYC 7261 has 96 KB of L1 cache per core, 512 KB of L2 cache per core, and 32 MB of shared L3 cache. The Xeon E5-2628L v3 has 64 KB of L1 per core, 256 KB of L2 per core, and 25 MB of shared L3 cache. The AMD part offers larger per-core caches and more total L3, which likely contributes to its benchmark edge despite fewer cores.

Memory architecture is another major divergence. The EPYC 7261 supports DDR4 with an eight-channel memory bus and 170.6 GB/s bandwidth. The Xeon E5-2628L v3 also supports DDR4 but only with a quad-channel bus, yielding 59.7 GB/s. Both support ECC memory, but the AMD part’s memory subsystem is far more capable. PCIe support is Gen 3 for both, though the Intel part specifies 40 lanes (CPU only), while the AMD part does not list a lane count.

The EPYC 7261 has an unlocked multiplier, while the Xeon E5-2628L v3 is locked. The Intel part also has a launch MSRP of $1364, which appears in the data as a one-time mention. The production statuses differ, with the AMD part active and the Intel part end-of-life. The sockets are incompatible: AMD Socket SP3 for the EPYC, Intel Socket 2011-3 for the Xeon.

The Verdict

The benchmark data consistently favors the AMD EPYC 7261 across all six head-to-head Cinebench tests, with win margins ranging from 0.5% to 0.8%. In Cinebench R15 multicore, the EPYC scores 954 versus 949 for the Xeon; in R20 multicore, it is 3979 versus 3955; in R23 multicore, it is 9476 versus 9417. Single-core results follow the same pattern: 134 vs 133 in R15, 561 vs 558 in R20, and 1337 vs 1329 in R23. The AMD part wins every benchmark, but the margins are slim — never exceeding 0.8%.

The Intel Xeon E5-2628L v3 offers more cores (10 vs 8) and threads (20 vs 16), but this does not translate into a performance advantage in the measured workloads. The EPYC 7261 also has a much higher TDP (170W vs 75W) and a newer process node (14 nm vs 22 nm). If the priority is absolute performance in Cinebench-style rendering tasks, the data points to the AMD EPYC 7261. If power efficiency is paramount, the Intel part’s 75W TDP is a clear advantage, though it comes with lower memory bandwidth and an older architecture.

For buyers who need an active production part with eight-channel memory and higher bandwidth, the EPYC 7261 is the choice. For those constrained by power budgets or thermal envelopes, the Xeon E5-2628L v3 offers a compelling low-TDP option, despite being end-of-life and losing every benchmark here. The data does not show a scenario where the Intel part wins a performance test, so any selection of the Xeon must be driven by power or platform considerations rather than raw score.

Specification Differences

• Cores: AMD EPYC 7261 has 8 cores; Intel Xeon E5-2628L v3 has 10 cores.
• Threads: AMD has 16 threads; Intel has 20 threads.
• Base Clock: AMD is 2.50 GHz; Intel is 2000.00 MHz (which is 2.00 GHz).
• Boost Clock: AMD is 2.90 GHz; Intel is 2.50 GHz.
• TDP: AMD is 170 watts; Intel is 75 watts.
• Socket: AMD uses Socket SP3; Intel uses Socket 2011-3.
• Architecture: AMD is Zen (Naples); Intel is Haswell (Haswell-EP).
• Process Node: AMD is 14 nm; Intel is 22 nm.
• Foundry: AMD uses GlobalFoundries; Intel uses Intel.
• Transistors: AMD has 4,800 million; Intel has 2,600 million.
• Die Size: AMD is 213 mm²; Intel is 356 mm².
• L1 Cache: AMD is 96 KB per core; Intel is 64 KB per core.
• L2 Cache: AMD is 512 KB per core; Intel is 256 KB per core.
• L3 Cache: AMD is 32 MB shared; Intel is 25 MB shared.
• Memory Bus: AMD is eight-channel; Intel is quad-channel.
• Memory Bandwidth: AMD is 170.6 GB/s; Intel is 59.7 GB/s.
• Release Date: AMD is 2018-06-13; Intel is 2014-09-07.
• Launch MSRP: Intel is $1364 (stated once); AMD has none listed.
• Multiplier Unlocked: AMD is true; Intel is false.
• Production Status: AMD is Active; Intel is End-of-life.
• Part Number: AMD is PS7261BEV8RAF; Intel is SR1XZ.

Head-to-Head Benchmarks

The EPYC 7261 wins all six head-to-head Cinebench tests. The largest margin is in Cinebench R15 single-core, where the AMD part scores 134 against the Intel part’s 133, a delta of 0.8%. The smallest margins are 0.5% in Cinebench R15 multicore (954 vs 949) and Cinebench R20 single-core (561 vs 558). The R20 multicore and R23 multicore tests show identical deltas of 0.6%, with scores of 3979 vs 3955 and 9476 vs 9417 respectively. The R23 single-core test also shows a 0.6% delta, with 1337 vs 1329.

The consistency of these wins is notable. Even though the Intel Xeon has 25% more cores and 25% more threads, it cannot overcome the AMD part’s per-core advantage. The EPYC’s higher boost clock (2.90 GHz vs 2.50 GHz) and larger cache hierarchy likely explain the edge. The deltaPct values are small, meaning the two chips are closely matched in raw Cinebench throughput, but the AMD part is never behind in any test. The average benchmark score also reflects this: 2740 for AMD versus 2724 for Intel, a 0.6% difference that matches the nearestRivals deltaPct for the Xeon E5-2628L v3 against the EPYC 7261.

Where Each One Wins

The AMD EPYC 7261 wins every benchmark in this comparison, so its strengths are clear: all six Cinebench tests, across single-core and multicore, from R15 to R23. The data shows it is 0.5% to 0.8% faster in each workload. This suggests that for rendering or compute tasks that scale with Cinebench scores, the EPYC 7261 is the better performer. Additionally, its eight-channel memory bus with 170.6 GB/s bandwidth is a major advantage for memory-intensive workloads, though no memory-specific benchmarks are provided in the data.

The Intel Xeon E5-2628L v3 does not win any benchmark, but it wins in other measurable dimensions. Its TDP of 75 watts is less than half the EPYC’s 170 watts, making it a far more power-efficient choice for dense server deployments where thermal limits are tight. It also has more cores and threads (10/20 vs 8/16), which could matter in workloads that are heavily parallel but do not exhibit the same performance scaling as Cinebench. The Intel part is end-of-life, while the AMD part is active, which affects long-term availability. The Intel chip also has a launch MSRP of $1364, but no price comparison is made here.

For a use-case split: choose the EPYC 7261 for peak Cinebench performance, higher memory bandwidth, and a newer, active platform. Choose the Xeon E5-2628L v3 for low power draw (75W TDP) and higher core counts, accepting that it loses every measured benchmark and is no longer in production. The data offers no scenario where the Intel part outperforms the AMD part in the provided tests, so any Intel selection must rely on power efficiency or platform compatibility rather than benchmark wins.