Intel Core i7-11390H vs Intel Xeon E5-2628L v3 Comparison

This comparison pits a 10-core, 20-thread Haswell-EP server processor against a 4-core, 8-thread Tiger Lake-H mobile chip. The data shows a clear sweep for the older Xeon in every benchmark listed, but the margin and the context matter more than the raw win count.

Head-to-Head Benchmarks

The Xeon E5-2628L v3 wins all six Cinebench comparisons, and it does so with remarkable consistency. In Cinebench R15 multi-core, the Xeon scores 949 against the Core i7-11390H’s 828, a 14.6% advantage. The single-core R15 result is nearly identical in percentage terms: 133 versus 116, also a 14.7% lead for the Xeon. This is the first notable surprise — a server part from 2014 beating a 2021 mobile chip in single-threaded performance, which is usually where newer architectures dominate.

Moving to Cinebench R20, the pattern holds. The Xeon posts 3955 multi-core versus 3451 for the i7, another 14.6% gap. In R20 single-core, the Xeon’s 558 beats the i7’s 487 by 14.6%. The consistency across R15 and R20 suggests this is not a workload-specific quirk but a fundamental performance differential in Cinebench’s rendering engine.

Cinebench R23 confirms the trend with the same 14.6% delta in both multi-core (9417 versus 8218) and single-core (1329 versus 1160). The near-identical percentage deltas across all six tests — 14.6% or 14.7% — indicate that whatever advantage the Xeon holds, it applies uniformly to both lightly threaded and heavily threaded tasks in this benchmark suite.

The average benchmark score data reinforces this. The Xeon averages 2724 across all its benchmarks, while the i7-11390H averages 2715. That is a 0.3% difference in the Xeon’s favor, which is nearly negligible. Both processors sit at the 50th percentile of all CPUs in the database. The Xeon’s nearest rival is the AMD Ryzen 7 4700U at 2722 (0.1% ahead), while the i7-11390H’s nearest rival is the Intel Xeon D-1577 at 2715 (0% delta). In other words, these two chips are statistically tied in overall average performance, despite the Xeon winning every head-to-head Cinebench test by a solid margin.

Where Each One Wins

Given the head-to-head sweep, the Xeon E5-2628L v3 is the clear pick for Cinebench-style rendering workloads. Whether you are looking at R15, R20, or R23, and whether you are using one core or all of them, the Xeon comes out ahead by roughly 14.6%. This makes it the better choice for anyone whose primary application is CPU-based rendering, 3D modeling, or video encoding that scales with Cinebench-like multi-threaded performance.

The Core i7-11390H, despite losing every listed benchmark, still has a role. It offers integrated Iris Xe graphics with 96 execution units, which the Xeon lacks entirely. For a system that needs display output without a discrete GPU, or for light GPU-accelerated tasks, the i7 is the only functional option. The Xeon has no integrated graphics, so it requires a separate graphics card just to post a display.

The i7 also has a much higher boost clock — 5.00 GHz versus 2.50 GHz on the Xeon — and a smaller 28W TDP versus 75W. While the benchmark data does not show this translating into a win in any Cinebench test, the lower power draw makes the i7 suitable for compact, battery-powered laptops. The Xeon’s 75W TDP and server socket are not realistic for portable use. The i7 is also listed as Active production status, while the Xeon is End-of-life, meaning the i7 remains a viable new purchase for mobile builds.

Architecture Differences

The two processors come from different eras and design philosophies. The Xeon E5-2628L v3 is built on Intel’s 22nm Haswell-EP process, with a die size of 356 mm² and 2,600 million transistors. It has 10 cores and 20 threads, with 64 KB of L1 cache per core, 256 KB of L2 per core, and 25 MB of shared L3 cache. It supports DDR4 memory across a quad-channel bus, yielding 59.7 GB/s of memory bandwidth, and includes ECC memory support. The Xeon uses Intel Socket 2011-3 and provides 40 PCIe Gen 3 lanes from the CPU.

The Core i7-11390H is a Tiger Lake-H part on Intel’s 10nm process, with a much smaller die at 146.1 mm². It has 4 cores and 8 threads, with a larger 80 KB L1 per core and 1.25 MB L2 per core, but only 12 MB of shared L3. It also supports DDR4, but across a dual-channel bus, delivering 51.2 GB/s of memory bandwidth — about 14% less than the Xeon. ECC memory is not supported. It uses Intel BGA 1449 and provides 20 PCIe Gen 4 lanes from the CPU, which is newer but half the lane count.

The cache hierarchy is telling. The Xeon’s 25 MB L3 is more than double the i7’s 12 MB, which helps with larger working sets. The i7’s per-core L2 is much larger at 1.25 MB versus 256 KB, but with only four cores, the total cache footprint is smaller. The Xeon’s memory bandwidth advantage (59.7 GB/s versus 51.2 GB/s) is modest but consistent with its quad-channel design. The i7 compensates with PCIe Gen 4 support, but the Xeon offers double the lanes at Gen 3.

The Xeon is a server/workstation part, released in September 2014, with a launch MSRP of $1364. The i7 is a mobile part, released in June 2021, with a launch MSRP of $426. The Xeon’s part number is SR1XZ; the i7’s is SRKSL. Neither has an unlocked multiplier.

The Verdict

From the data alone, the Xeon E5-2628L v3 is the faster processor in every benchmark category measured. If your workload is Cinebench — or anything that behaves like it — the Xeon delivers a 14.6% performance advantage across the board. This is a meaningful edge for rendering tasks, and it comes with the benefits of ECC memory support and quad-channel bandwidth.

The Core i7-11390H, however, is not a loser in all contexts. It is the only one of the two that can run without a discrete GPU, thanks to its Iris Xe integrated graphics. It draws significantly less power at 28W versus 75W, making it the only viable choice for laptop or small-form-factor builds. It also has a much higher boost clock at 5.00 GHz, which may matter in short bursts of single-threaded work, even though the Cinebench single-core results do not reflect that advantage.

Choose the Xeon if you are building a workstation or server where rendering performance, memory bandwidth, and ECC support are priorities, and where a separate GPU is already part of the plan. The Xeon’s End-of-life status means it is a used-market part, but the benchmark data shows it still competes at the 50th percentile of all CPUs. Choose the i7 if you need a mobile processor with integrated graphics, low power consumption, and active production status. The i7’s performance is within 0.3% of the Xeon’s average score, and its 28W TDP makes it appropriate for thin-and-light laptops where the Xeon physically cannot fit.

The verdict is straightforward: for raw compute in a stationary system, the Xeon wins. For portable or graphics-inclusive systems, the i7 is the only option that works.

FAQ

Q: Which processor is faster in multi-core Cinebench R23?

A: The Intel Xeon E5-2628L v3 scores 9417, which is 14.6% higher than the Core i7-11390H’s 8218.

Q: Does the Core i7-11390H win any of the listed head-to-head benchmarks?

A: No. The Xeon E5-2628L v3 wins all six Cinebench tests, with deltas ranging from 14.6% to 14.7%.

Q: Which CPU supports ECC memory?

A: Only the Intel Xeon E5-2628L v3 supports ECC memory. The Core i7-11390H does not.

Q: What is the difference in memory bus width?

A: The Xeon uses a quad-channel DDR4 bus with 59.7 GB/s bandwidth, while the i7 uses a dual-channel bus with 51.2 GB/s bandwidth.

Q: Can either CPU run without a discrete graphics card?

A: The Core i7-11390H has integrated Iris Xe graphics with 96 execution units. The Xeon E5-2628L v3 has no integrated graphics, so it requires a separate GPU for display output.

Q: How do their average benchmark scores compare?

A: The Xeon averages 2724, while the i7 averages 2715. That is a 0.3% difference in favor of the Xeon, and both sit at the 50th percentile of all CPUs.

Specification Differences

| Specification | Intel Xeon E5-2628L v3 | Intel Core i7-11390H |

|---|---|---|

| Cores | 10 | 4 |

| Threads | 20 | 8 |

| Base Clock | 2000.00 MHz | 2.90 GHz |

| Boost Clock | 2.50 GHz | 5.00 GHz |

| TDP | 75W | 28W |

| Socket | Intel Socket 2011-3 | Intel BGA 1449 |

| Architecture | Haswell | Tiger Lake |

| Codename | Haswell-EP | Tiger Lake-H |

| Process Node | 22 nm | 10 nm |

| Die Size | 356 mm² | 146.1 mm² |

| Transistors | 2,600 million | Not specified |

| L1 Cache | 64 KB (per core) | 80 KB (per core) |

| L2 Cache | 256 KB (per core) | 1.25 MB (per core) |

| L3 Cache | 25 MB (shared) | 12 MB (shared) |

| Memory Bus | Quad-channel | Dual-channel |

| Memory Bandwidth | 59.7 GB/s | 51.2 GB/s |

| ECC Memory | Yes | No |

| PCIe | Gen 3, 40 Lanes (CPU only) | Gen 4, 20 Lanes (CPU only) |

| Integrated Graphics | None | Iris XE 96EU |

| Market Segment | Server/Workstation | Mobile |

| Production Status | End-of-life | Active |

| Release Date | 2014-09-07 | 2021-06-20 |

| Launch MSRP | $1364 | $426 |

| Part Number | SR1XZ | SRKSL |