Intel Xeon E3-1545M v5 vs Intel Xeon E5-2620 v3 Comparison

The Verdict

The benchmark database presents a surprisingly close contest between two Xeon processors from different Intel eras. The Intel Xeon E5-2620 v3, a 6-core Haswell-EP server chip from 2014, edges out the Intel Xeon E3-1545M v5, a 4-core Skylake-H mobile workstation chip from 2016, across every single recorded Cinebench test. The data shows the E5-2620 v3 winning all six head-to-head comparisons, though the margins are remarkably narrow, ranging from 1.1% to 1.6%. Neither processor escapes the 43rd percentile among all CPUs, placing both in the lower-middle tier of the database's performance distribution.

The E5-2620 v3's aggregate average benchmark score of 1907 slightly exceeds the E3-1545M v5's 1879, a gap of roughly 1.5%. For buyers, the choice hinges on platform context rather than raw compute dominance. The E5-2620 v3 offers more cores and threads plus quad-channel DDR4 memory, while the E3-1545M v5 delivers higher clock speeds, a newer architecture, integrated graphics, and dramatically lower power consumption. The E3-1545M v5 suits compact workstation builds where the 45 watt TDP and BGA 1440 socket matter; the E5-2620 v3 fits dual-socket servers or large workstations needing 40 PCIe lanes and up to 59.7 GB/s of memory bandwidth. Given that the E5-2620 v3 wins every benchmark, performance seekers should choose it, while those prioritizing efficiency and integrated graphics should select the E3-1545M v5.

Architecture Differences

The two processors represent different Intel design philosophies. The E5-2620 v3 uses the Haswell-EP microarchitecture on a 22 nm process node, while the E3-1545M v5 employs the Skylake-H architecture on a 14 nm node. This generational gap explains several specification divergences. The E5-2620 v3's die size measures 356 mm² and contains 2,600 million transistors; the E3-1545M v5's die is far smaller at 171 mm² with 2,300 million transistors. The smaller process node allows the E3-1545M v5 to pack nearly as many transistors into roughly half the silicon area.

Core configuration differs substantially. The E5-2620 v3 provides 6 cores and 12 threads, while the E3-1545M v5 offers 4 cores and 8 threads. The E5-2620 v3 compensates for its older architecture with two additional physical cores. Cache allocation follows the same pattern: both share 64 KB of L1 and 256 KB of L2 per core, but the E5-2620 v3 carries 15 MB of shared L3 cache versus 8 MB on the E3-1545M v5. The E5-2620 v3's larger L3 cache likely supports its extra cores and threads.

Memory architecture diverges sharply. The E5-2620 v3 supports DDR4 over a quad-channel memory bus, delivering 59.7 GB/s of bandwidth. The E3-1545M v5 supports both DDR3 and DDR4 over a dual-channel bus, capping at 34.1 GB/s. Both processors support ECC memory, a requirement for server and workstation reliability. PCIe connectivity also differs: the E5-2620 v3 provides 40 Gen 3 lanes, while the E3-1545M v5 provides 16 Gen 3 lanes. The E5-2620 v3 targets systems with many expansion cards, storage controllers, or GPUs, whereas the E3-1545M v5 suits more modest configurations.

The E3-1545M v5 uniquely integrates Iris Pro Graphics P580, a feature absent from the E5-2620 v3. This makes the E3-1545M v5 viable for systems without a discrete GPU, a capability the E5-2620 v3 cannot offer. The E3-1545M v5's socket, Intel BGA 1440, indicates a soldered mobile design, while the E5-2620 v3 uses Intel Socket 2011-3, a socketed server platform. Both processors carry the "Server/Workstation" market segment label and are end-of-life products.

Head-to-Head Benchmarks

The recorded Cinebench results reveal a pattern of consistent, narrow victories for the E5-2620 v3. In Cinebench R15 multicore, the E5-2620 v3 scores 664 against the E3-1545M v5's 654, a 1.5% advantage. The single-core R15 test shows 93 versus 92, a 1.1% edge. These margins are small enough to fall within typical run-to-run variance, yet they repeat across every test iteration.

Cinebench R20 multicore results show the E5-2620 v3 at 2769 versus 2728, again a 1.5% delta. The R20 single-core test records 391 against 385, a 1.6% gap, the largest single margin in the dataset. Cinebench R23 multicore produces scores of 6595 and 6497, maintaining the 1.5% pattern. The R23 single-core test yields 931 versus 917, a 1.5% difference.

The data implies the two processors deliver nearly identical compute performance despite their architectural differences. The E5-2620 v3's two extra cores should give it a substantial multicore advantage, but the E3-1545M v5's higher clocks (2.90 GHz base, 3.80 GHz boost versus 2.40 GHz and 3.20 GHz) and newer Skylake architecture largely offset that core deficit. The E5-2620 v3 wins all six recorded benchmarks, yet the E3-1545M v5 never falls more than 1.6% behind. This suggests that for single-threaded workloads, the E3-1545M v5 nearly matches the E5-2620 v3, and for multi-threaded workloads, the extra cores only provide a modest edge.

Specification Differences

The two processors differ on nearly every measurable specification. The E5-2620 v3 has 6 cores and 12 threads; the E3-1545M v5 has 4 cores and 8 threads. Base clocks run 2.40 GHz on the E5-2620 v3 versus 2.90 GHz on the E3-1545M v5. Boost clocks run 3.20 GHz versus 3.80 GHz, favoring the E3-1545M v5 by 0.60 GHz in both cases. TDP differs dramatically: 85 watts for the E5-2620 v3 versus 45 watts for the E3-1545M v5, making the E3-1545M v5 far more power-efficient.

Socket types are incompatible: Intel Socket 2011-3 for the E5-2620 v3, Intel BGA 1440 for the E3-1545M v5. Architecture and process node differ as well: Haswell-EP on 22 nm versus Skylake-H on 14 nm. The E5-2620 v3's die measures 356 mm² with 2,600 million transistors; the E3-1545M v5's die measures 171 mm² with 2,300 million transistors. L3 cache totals 15 MB on the E5-2620 v3 versus 8 MB on the E3-1545M v5.

Memory support separates the two: the E5-2620 v3 supports DDR4 only, while the E3-1545M v5 supports both DDR3 and DDR4. Memory bus width is quad-channel on the E5-2620 v3 versus dual-channel on the E3-1545M v5, producing bandwidth figures of 59.7 GB/s and 34.1 GB/s respectively. PCIe lane counts are 40 versus 16, both Gen 3. Integrated graphics exist only on the E3-1545M v5 as Iris Pro Graphics P580; the E5-2620 v3 has none. Release dates differ as well: the E5-2620 v3 launched in September 2014, while the E3-1545M v5 launched in January 2016. The launch MSRP for the E5-2620 v3 is $417, and for the E3-1545M v5 it is $679.

FAQ

Q: Which processor is faster in multi-core Cinebench tests?

A: The Intel Xeon E5-2620 v3 wins all three multi-core tests. Its Cinebench R15 multicore score is 664 versus 654, R20 multicore is 2769 versus 2728, and R23 multicore is 6595 versus 6497. The margin is consistently 1.5% in favor of the E5-2620 v3.

Q: Does the E3-1545M v5's higher clock speed make it faster in single-core tests?

A: No. Despite base and boost clocks of 2.90 GHz and 3.80 GHz, compared to 2.40 GHz and 3.20 GHz on the E5-2620 v3, the E5-2620 v3 still wins every single-core test. The margins are 1.1% in R15 (93 versus 92), 1.6% in R20 (391 versus 385), and 1.5% in R23 (931 versus 917).

Q: How do their average benchmark scores compare?

A: The E5-2620 v3 has an average benchmark score of 1907, while the E3-1545M v5 scores 1879. This places the E5-2620 v3 approximately 1.5% higher on average. Both processors sit at the 43rd percentile among all CPUs in the database.

Q: What memory bandwidth does each processor support?

A: The E5-2620 v3 supports quad-channel DDR4 memory with 59.7 GB/s of bandwidth. The E3-1545M v5 supports dual-channel DDR3 or DDR4 memory with 34.1 GB/s of bandwidth. The E5-2620 v3 offers roughly 75% more theoretical memory bandwidth.

Q: Which processor includes integrated graphics?

A: Only the Intel Xeon E3-1545M v5 includes integrated graphics, specifically Iris Pro Graphics P580. The Intel Xeon E5-2620 v3 has no integrated graphics, requiring a discrete GPU for display output.

Q: What are the TDP differences between the two?

A: The E5-2620 v3 has a TDP of 85 watts, while the E3-1545M v5 has a TDP of 45 watts. The E3-1545M v5 consumes nearly half the power, making it better suited for compact or thermally constrained systems.

Where Each One Wins

The E5-2620 v3 wins every recorded benchmark, but the context matters. Its victories in Cinebench R15, R20, and R23, both single-core and multi-core, make it the default choice for pure compute performance. The 6-core, 12-thread configuration, combined with 15 MB of L3 cache and quad-channel memory, suggests strength in multi-threaded server workloads such as virtualization, database processing, and content rendering. The 40 PCIe Gen 3 lanes and 59.7 GB/s memory bandwidth support heavy I/O configurations, making it suitable for dual-socket servers or workstations with multiple GPUs and storage arrays. Its launch MSRP of $417 also positions it as a lower-cost entry into the Xeon E5 platform.

The E3-1545M v5 wins on efficiency and integration. Its 45 watt TDP, less than half the E5-2620 v3's 85 watts, makes it attractive for small-form-factor workstations, embedded systems, or mobile workstations where heat and power budgets are tight. The integrated Iris Pro Graphics P580 eliminates the need for a discrete GPU in basic display or light graphics workloads. Its support for both DDR3 and DDR4 memory offers flexibility in system design, though the dual-channel bus and 34.1 GB/s bandwidth limit memory-intensive tasks. The higher base and boost clocks (2.90 GHz and 3.80 GHz) provide a clock-for-clock advantage, though the recorded benchmarks show this does not translate into a performance win. The E3-1545M v5's BGA 1440 socket means it is soldered to the motherboard, which suits OEM designs but prevents CPU upgrades.

For single-threaded responsiveness, the two chips are virtually indistinguishable, with the E5-2620 v3 leading by roughly 1.5% in every test. For multi-threaded workloads, the E5-2620 v3's extra cores provide a consistent, though modest, advantage. The database's percentile ranking places both at 43, indicating neither excels beyond the middle of the CPU distribution. The E5-2620 v3 is the pick for anyone maximizing compute throughput on a socketed server platform. The E3-1545M v5 is the pick for anyone needing a low-power, integrated-GPU solution in a compact form factor, accepting a small performance deficit in exchange for substantial power savings and platform simplicity.