Intel Xeon E5-2628L v3 vs Intel Xeon E5-2640 v3 Comparison

# Intel Xeon E5-2628L v3 vs Intel Xeon E5-2640 v3

The data presents a fascinating contradiction: the E5-2628L v3 wins every single benchmark in the head-to-head comparison despite having a lower base clock, lower boost clock, and a higher launch MSRP. The E5-2640 v3, with its 2.60 GHz base and 3.40 GHz boost, looks superior on paper, yet the E5-2628L v3 edges it out in every Cinebench test by margins ranging from 0.7% to 0.9%. This suggests that the 10-core/20-thread configuration of the E5-2628L v3 provides enough parallel throughput advantage to overcome its clock-speed deficit, even in single-core workloads where clock speed traditionally dominates.

The Verdict

The benchmark data is unambiguous: the Intel Xeon E5-2628L v3 wins all six head-to-head comparisons, making it the superior choice for any workload measured by Cinebench. The largest margin is a 0.9% lead in Cinebench R20 single-core (558 vs 553), while the smallest is a 0.7% lead in both R15 and R23 multicore tests. This consistency across all three Cinebench versions—R15, R20, and R23—indicates a genuine, if modest, performance advantage rather than a test-specific anomaly.

The E5-2640 v3 offers nothing in terms of benchmark wins, yet it costs less at launch with an MSRP of $939 compared to $1364 for the E5-2628L v3. The E5-2628L v3 also consumes less power with a 75 W TDP versus 90 W for the E5-2640 v3. For buyers prioritizing raw Cinebench performance and lower power draw, the E5-2628L v3 is the clear winner despite its higher price. However, the E5-2640 v3's higher clock speeds might make it attractive for lightly-threaded workloads not captured in these benchmarks, though the data does not confirm this advantage.

Both processors share the same socket, architecture, process node, and memory support, so platform compatibility is identical. The E5-2628L v3's extra two cores and 5 MB of additional L3 cache likely explain its benchmark dominance, making it the better choice for multi-threaded server workloads. The E5-2640 v3, with its higher clocks, could be considered for scenarios where single-thread latency is critical, but the benchmark data shows the E5-2628L v3 actually wins those tests too.

FAQ

Q: Which processor has more cores?

A: The Intel Xeon E5-2628L v3 has 10 cores and 20 threads, while the Intel Xeon E5-2640 v3 has 8 cores and 16 threads. This gives the E5-2628L v3 a 25% core advantage.

Q: What is the performance difference in Cinebench R23 multicore?

A: The E5-2628L v3 scores 9417 versus 9348 for the E5-2640 v3, a 0.7% difference. This is the smallest multicore margin in the head-to-head results.

Q: Which processor has the higher boost clock?

A: The E5-2640 v3 boosts to 3.40 GHz, while the E5-2628L v3 boosts to only 2.50 GHz. Despite this 0.90 GHz advantage, the E5-2640 v3 loses every benchmark.

Q: Do both processors support ECC memory?

A: Yes, both support ECC memory and share the same DDR4 quad-channel memory bus with 59.7 GB/s bandwidth.

Q: What is the TDP difference between the two?

A: The E5-2628L v3 has a 75 W TDP, while the E5-2640 v3 has a 90 W TDP. The E5-2628L v3 delivers better benchmark performance with 15 W lower power consumption.

Q: Are both processors from the same generation?

A: Yes, both are Haswell-EP generation processors on the 22 nm process with 2,600 million transistors and a 356 mm² die size.

Architecture Differences

Both processors share the Haswell-EP architecture on Intel's 22 nm process, manufactured by Intel with 2,600 million transistors on a 356 mm² die. The fundamental architecture is identical—same socket, same memory controller, same PCIe implementation—but the core configuration and cache hierarchy differ meaningfully.

The E5-2628L v3 packs 10 cores and 20 threads, while the E5-2640 v3 has 8 cores and 16 threads. This 2-core difference translates directly to L3 cache: the E5-2628L v3 has 25 MB shared L3, while the E5-2640 v3 has 20 MB. Both use 64 KB L1 and 256 KB L2 per core, so the per-core cache structure is consistent. The extra L3 cache on the E5-2628L v3 likely helps with data locality in multi-threaded workloads, contributing to its benchmark wins.

The clock strategy differs significantly. The E5-2640 v3 runs at 2.60 GHz base and 3.40 GHz boost, while the E5-2628L v3 runs at 2.00 GHz base and 2.50 GHz boost. The "L" in the E5-2628L v3's name suggests a low-power variant, and the 75 W TDP confirms this—it achieves lower power consumption despite having more cores by running at much lower clocks. This is a classic power-vs-performance tradeoff, but the benchmark data shows the E5-2628L v3 still wins, suggesting the additional cores compensate for the clock deficit in Cinebench workloads.

The memory subsystem is identical: DDR4, quad-channel, 59.7 GB/s bandwidth, with ECC support. PCIe configuration is also the same: Gen 3 with 40 lanes from the CPU. Both are end-of-life products released on the same date, 2014-09-07, and both target the server/workstation market segment. Neither has integrated graphics, and both lack an unlocked multiplier.

Specification Differences

The core count differs: 10 cores/20 threads for the E5-2628L v3 versus 8 cores/16 threads for the E5-2640 v3. This is the most consequential specification difference, as it directly impacts parallel processing capability.

Clock speeds differ substantially. The E5-2628L v3 has a 2.00 GHz base and 2.50 GHz boost, while the E5-2640 v3 has a 2.60 GHz base and 3.40 GHz boost. The E5-2640 v3's boost clock is 0.90 GHz higher, yet this advantage does not translate to benchmark wins.

TDP differs by 15 W: 75 W for the E5-2628L v3 versus 90 W for the E5-2640 v3. This makes the E5-2628L v3 more power-efficient per core and per unit of performance, based on the benchmark results.

L3 cache differs: 25 MB shared for the E5-2628L v3 versus 20 MB shared for the E5-2640 v3. This 5 MB difference likely contributes to the E5-2628L v3's superior multicore performance.

The launch MSRP differs: $1364 for the E5-2628L v3 versus $939 for the E5-2640 v3. The part numbers also differ (SR1XZ for the E5-2628L v3, SR205 for the E5-2640 v3), but all other specifications—socket, architecture, process node, foundry, transistors, die size, L1/L2 cache, memory support, memory bus, memory bandwidth, ECC support, PCIe, market segment, production status, and release date—are identical.

Head-to-Head Benchmarks

The head-to-head results show a clean sweep for the E5-2628L v3 across all six Cinebench tests. In Cinebench R15 multicore, the E5-2628L v3 scores 949 against 942 for the E5-2640 v3, a 0.7% margin. The single-core R15 test shows 133 versus 132, a 0.8% edge. These margins are small but consistent.

Moving to Cinebench R20, the E5-2628L v3 scores 3955 multicore versus 3926, again a 0.7% difference. The single-core R20 test shows 558 versus 553, the largest margin at 0.9%. The E5-2628L v3's biggest win is in R20 single-core, which is surprising given the E5-2640 v3's 0.90 GHz boost clock advantage.

In Cinebench R23, the E5-2628L v3 scores 9417 multicore versus 9348, a 0.7% margin, and 1329 single-core versus 1319, a 0.8% margin. Across all tests, the deltas range from 0.7% to 0.9%, with no test exceeding 1%.

Looking at the average benchmark scores, the E5-2628L v3 has an average score of 2724, while the E5-2640 v3 averages 2703. This 21-point difference (0.8%) aligns with the head-to-head deltas. Both processors sit at the 50th percentile among all CPUs, placing them in the middle of the performance distribution. The nearest rivals for the E5-2628L v3 include the AMD Ryzen 7 4700U (2722, 0.1% behind) and the Intel Core i7-11390H (2715, 0.3% behind), while the E5-2640 v3's nearest rival is the Intel Core i7-1185G7E at 2703 with a 0% delta.

Where Each One Wins

The E5-2628L v3 wins everywhere the data measures. All six head-to-head benchmarks go its way, with wins in both multicore and single-core tests across R15, R20, and R23. The consistency of these wins—never below 0.7% and never above 0.9%—suggests a fundamental architectural advantage rather than a workload-specific quirk. The 10-core/20-thread configuration with 25 MB L3 cache provides more parallelism and cache capacity, which helps even in single-core tests by potentially allowing better scheduling or thermal headroom.

The E5-2640 v3 wins nowhere in the benchmark data. Its higher base and boost clocks (2.60/3.40 GHz versus 2.00/2.50 GHz) do not translate to any measured performance advantage. The only area where it could be considered "better" is its lower launch MSRP at $939 versus $1364, but the data does not quantify whether this price difference is justified by any performance benefit, since the more expensive processor actually performs better.

For use cases, the E5-2628L v3 is preferable for any Cinebench-type workload—rendering, multi-threaded computation, or any application that scales with core count and cache size. Its lower 75 W TDP also makes it attractive for dense server deployments where power efficiency matters. The E5-2640 v3, despite losing all benchmarks, might still serve in scenarios where its higher clock speeds benefit specific legacy applications not represented in the Cinebench suite, but the data cannot confirm this advantage. Given the benchmark results, the E5-2628L v3 is the better processor by every measured metric.