Intel Xeon E5-1660 v3 vs Intel Xeon E5-2618L v3 Comparison

Both processors are 8-core, 16-thread Haswell-EP parts on the same 22 nm process, sharing the same 2,600 million transistor count and 356 mm² die size. The data shows a remarkably close contest, with the Intel Xeon E5-1660 v3 winning all six head-to-head Cinebench tests, yet the margins are so thin that the average benchmark scores differ by only 16 points. The E5-1660 v3 posts an average score of 3082 against the E5-2618L v3's 3066, a gap of just over 0.5%. This raises an immediate question: with such similar performance, what separates these two Haswell-EP siblings, and why would a buyer choose one over the other?

FAQ

Q: How much faster is the Intel Xeon E5-1660 v3 in multi-core workloads?

A: The E5-1660 v3 leads in all three multi-core Cinebench tests, but by tiny margins. It scores 1074 versus 1068 in R15 (0.6% ahead), 4475 versus 4452 in R20 (0.5% ahead), and 10656 versus 10601 in R23 (0.5% ahead). These deltas are within run-to-run variance, making the multi-core performance effectively identical.

Q: What about single-core performance—does the higher boost clock help?

A: The E5-1660 v3 has a 3.50 GHz boost clock versus 3.40 GHz on the E5-2618L v3, and it wins every single-core test. However, the advantage is again marginal: 151 versus 150 in R15 single-core (0.7%), 631 versus 628 in R20 (0.5%), and 1504 versus 1496 in R23 (0.5%). The data suggests the extra 0.10 GHz boost yields only a fractional real-world gain.

Q: Do the two CPUs have the same cache configuration?

A: Yes. Both feature 64 KB of L1 cache per core, 256 KB of L2 cache per core, and a shared 20 MB L3 cache. The identical cache hierarchy, combined with the same core count, explains why their benchmark scores are so close.

Q: What is the most significant difference between these two processors?

A: Thermal design power (TDP) is the largest gap. The E5-1660 v3 has a 140 W TDP, while the E5-2618L v3 is rated at just 75 W. This 65 W difference represents a substantial reduction in power draw for the latter, despite near-identical performance. The E5-2618L v3's lower TDP likely stems from its lower base clock of 2.30 GHz versus 3.00 GHz on the E5-1660 v3.

Q: Are both processors still in production?

A: No. The E5-1660 v3 is marked as end-of-life, while the E5-2618L v3 is listed as active. Both were released on the same date in September 2014, but they now occupy different positions in Intel's lifecycle. The E5-1660 v3 also has an unlocked multiplier, whereas the E5-2618L v3 is locked.

Q: How do these CPUs compare to the AMD Ryzen 7 1800X?

A: The E5-1660 v3 is 0.4% ahead of the Ryzen 7 1800X in average benchmark score (3082 versus 3068), while the E5-2618L v3 is 0.1% behind the same rival (3066 versus 3068). Both Xeons effectively trade blows with the Ryzen 7 1800X, according to the nearestRivals data.

Architecture Differences

Both processors share the same fundamental Haswell-EP design: 22 nm process, 8 cores, 16 threads, and identical cache sizes (64 KB L1 per core, 256 KB L2 per core, 20 MB shared L3). The transistor count (2,600 million) and die size (356 mm²) are also identical, indicating they are likely the same physical die binned for different power and clock targets. The architecture itself is unchanged between the two—the differences lie in how Intel configured the silicon.

The most consequential architectural distinction is the memory bandwidth. The E5-1660 v3 is rated at 68.3 GB/s, while the E5-2618L v3 drops to 59.7 GB/s, an 8.6 GB/s deficit. Both use quad-channel DDR4, so this difference likely reflects the lower base and boost clocks of the E5-2618L v3, which limits memory controller throughput. This could matter in memory-bound workloads, but the Cinebench results show it does not affect CPU rendering performance.

Clock speeds further differentiate the pair. The E5-1660 v3 runs at a 3.00 GHz base and 3.50 GHz boost, while the E5-2618L v3 operates at 2.30 GHz base and 3.40 GHz boost. The 0.70 GHz base clock gap is substantial and explains the TDP difference—the E5-2618L v3's 75 W rating versus 140 W for the E5-1660 v3. The unlocked multiplier on the E5-1660 v3 also allows overclocking, a feature the E5-2618L v3 lacks entirely.

Both chips share the Intel Socket 2011-3 interface, support DDR4 memory, enable ECC, and provide PCIe Gen 3 with 40 lanes (CPU only). Neither has integrated graphics. The production status differs, with the E5-1660 v3 at end-of-life and the E5-2618L v3 still active, which suggests Intel sees ongoing demand for the lower-power part in server deployments.

The Verdict

The benchmark data paints a clear picture: these are near-identical performers with a decisive power consumption split. The E5-1660 v3 wins all six Cinebench tests, but by margins of 0.5% to 0.7%—fractions that will not be perceptible in real workloads. Its average benchmark score of 3082 is only 0.5% higher than the E5-2618L v3's 3066. Anyone choosing between these two purely on performance would be splitting hairs.

The verdict hinges on the TDP and production status. The E5-2618L v3's 75 W TDP is nearly half the E5-1660 v3's 140 W, making it the obvious choice for dense server environments, power-constrained racks, or systems where cooling is limited. Its active production status also means it remains available for new builds, whereas the E5-1660 v3 is end-of-life. The E5-1660 v3 counterargument is the unlocked multiplier—it offers overclocking headroom that the E5-2618L v3 cannot match, plus a marginally higher boost clock. The data suggests the E5-1660 v3 is for enthusiasts or workstation users who want the last drop of performance and can tolerate 140 W, while the E5-2618L v3 is for datacenter operators who prioritize efficiency and availability.

Specification Differences

| Specification | Intel Xeon E5-1660 v3 | Intel Xeon E5-2618L v3 |

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

| Base Clock | 3.00 GHz | 2.30 GHz |

| Boost Clock | 3.50 GHz | 3.40 GHz |

| TDP | 140 W | 75 W |

| Memory Bandwidth | 68.3 GB/s | 59.7 GB/s |

| Multiplier Unlocked | Yes | No |

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

| Launch MSRP | N/A | $779 |

| Part Number | SR20N | SR200 |

Both CPUs share identical core counts (8), threads (16), cache configuration (64 KB L1, 256 KB L2, 20 MB L3), socket (Intel Socket 2011-3), architecture (Haswell), process node (22 nm), foundry (Intel), memory support (DDR4 quad-channel), ECC support, PCIe configuration (Gen 3, 40 lanes), and release date (September 2014). The differences listed above are the only specification-level gaps between them.

Head-to-Head Benchmarks

The E5-1660 v3 sweeps all six head-to-head Cinebench tests, but the margin of victory never exceeds 0.7%. In Cinebench R15 multi-core, it scores 1074 against 1068, a 0.6% lead. The R15 single-core test shows 151 versus 150, a 0.7% gap that represents the largest delta in the entire comparison. Moving to R20, the multi-core score is 4475 versus 4452 (0.5%), and single-core is 631 versus 628 (0.5%). The R23 results mirror this pattern: 10656 versus 10601 in multi-core (0.5%) and 1504 versus 1496 in single-core (0.5%).

The consistency of these deltas—all hovering around 0.5%—suggests the boost clock advantage of the E5-1660 v3 is the sole differentiator. The 0.10 GHz boost advantage translates to roughly half a percent in Cinebench, which matches the observed scores. The base clock difference (0.70 GHz) does not appear to affect these tests because Cinebench is bursty and often runs at boost frequencies. The memory bandwidth gap (68.3 GB/s versus 59.7 GB/s) also fails to manifest in Cinebench, indicating these workloads are not memory-bandwidth-limited.

Interestingly, the E5-1660 v3's nearestRivals include the Intel Core i7-6850K (deltaPct 0.1%), Intel Core i7-8700B (deltaPct -0.1%), and Intel Xeon W-2133 (deltaPct -0.2%). The E5-2618L v3's rivals include the AMD Ryzen 7 4980U (deltaPct 0%) and AMD Ryzen 3 PRO 7330U (deltaPct 0.1%). Both Xeons occupy the same performance tier as these CPUs, but the E5-1660 v3 edges out the Ryzen 7 1800X by 0.4%, while the E5-2618L v3 trails it by 0.1%.

Where Each One Wins

The E5-1660 v3 wins in every single benchmark category: multi-core and single-core across Cinebench R15, R20, and R23. Its 3.50 GHz boost clock gives it the edge in lightly threaded tasks, as evidenced by the single-core wins. The unlocked multiplier is a speculative win—it allows overclocking, which could widen the performance gap beyond the stock 0.5% to 0.7% margins, though the data does not quantify this potential. For workloads that demand the absolute maximum Cinebench score, the E5-1660 v3 is the pick, albeit by a razor-thin margin.

The E5-2618L v3 wins where performance is not the metric—power efficiency and availability. Its 75 W TDP is 46.4% lower than the 140 W rating of the E5-1660 v3, a massive reduction that makes it the superior choice for multi-socket servers, blade chassis, or any environment where heat dissipation and power budgets are critical. The active production status ensures it can be sourced for new systems, while the E5-1660 v3 is end-of-life. The lower base clock of 2.30 GHz also suggests it may sustain boost clocks longer under sustained loads, given the reduced thermal headroom required, though the Cinebench data does not directly confirm this. For datacenter operators who prioritize density and reliability over a 0.5% performance bump, the E5-2618L v3 is the clear winner.