Intel Xeon D-1587

Intel Xeon D-1587 is a 16-core, 32-thread Broadwell-DE processor aimed squarely at the server and workstation segment, with a 65W TDP and a launch MSRP of 1549. Its benchmark profile places it near the middle of the pack for all CPUs, at the 56th percentile, with an average benchmark score of 3282. The data shows a processor that trades raw single-thread speed for a high core count, making its performance context heavily dependent on workload scaling.

Benchmark Performance

The Cinebench results for the Intel Xeon D-1587 reveal a clear performance profile. In Cinebench R23, the processor scores 11349 points in multi-core and 1602 points in single-core. This represents a multi-core to single-core ratio of roughly 7.1:1, indicating that the chip is designed to excel in heavily threaded environments while accepting lower performance in lightly threaded tasks. The R20 scores follow the same pattern: 4766 multi-core and 672 single-core.

Relative to its nearest rivals, the Xeon D-1587's average benchmark score of 3282 is nearly identical to the AMD Ryzen Embedded V2516 (3277), the Intel Xeon E5-2650 v4 (3276), the Intel Xeon E-2136 (3274), and the AMD Ryzen 3 3300X (3272). The deltaPct values for all these rivals are extremely small — 0.2% or 0.3% — which means the D-1587 effectively sits in a statistical tie with each of them on average. This is a striking result, as the D-1587 achieves this parity through a completely different architecture: 16 slower Broadwell cores versus the newer, lower-core-count designs of its rivals.

The Cinebench R15 scores reinforce the multi-core focus. The D-1587 scores 1143 multi-core and 161 single-core, again showing a significant gap between the two metrics. The data suggests that the D-1587's average score is buoyed entirely by its multi-core throughput; its single-core scores are not competitive with the more modern cores in the rival list, but the sheer core count compensates in multi-threaded benchmarks.

How It Compares

AMD Ryzen Embedded V2516: The D-1587 edges out the V2516 by a 0.2% margin in average score, which is negligible in real-world terms. This is a notable comparison because the V2516 is a much newer, more power-efficient embedded part with far fewer cores. The D-1587's advantage comes from its 16 cores versus the V2516's 6 cores, but the V2516's newer architecture likely provides superior per-core efficiency, making the overall performance parity a testament to the D-1587's brute-force core count rather than architectural efficiency.

Intel Xeon E5-2650 v4: The D-1587 also leads the E5-2650 v4 by 0.2% in average score. Both are Broadwell-generation parts, but the E5-2650 v4 is a larger, higher-end server chip with 12 cores. The D-1587 manages to match it with more cores at a lower TDP class, which is a useful data point for dense server deployments. The performance is effectively identical, so the choice between them would come down to platform features rather than raw compute.

Intel Xeon E-2136: The 0.2% lead over the E-2136 is again within the margin of error. The E-2136 is a 6-core Coffee Lake part with a much higher clock speed. The D-1587's 16 cores allow it to keep pace in multi-threaded workloads, but the E-2136 will win decisively in single-threaded tasks. The average score masks this divergence; the D-1587 is the better choice for parallel throughput, while the E-2136 is better for latency-sensitive, lightly threaded applications.

AMD Ryzen 3 3300X: The D-1587 is 0.3% ahead of the Ryzen 3 3300X on average. The 3300X is a 4-core, 8-thread desktop processor with a high boost clock. This comparison highlights the D-1587's unique position: it competes with a budget desktop chip on average score, but achieves this via 4x the cores and 4x the threads. For workloads that scale perfectly, the D-1587 will pull ahead; for single-threaded tasks, the 3300X will dominate. The average score hides these divergent behaviors.

Power and Thermals

The Intel Xeon D-1587 carries a 65W TDP, which is notably low for a 16-core server processor. This is a critical data point for system design, as it implies the chip can be cooled by a capable air cooler or a low-profile server heatsink, without requiring exotic liquid cooling or high-static-pressure fans. The 65W figure places it in the same thermal class as many mainstream desktop CPUs, despite having double the core count of typical consumer parts.

The low TDP suggests that sustained all-core workloads will be limited by power delivery and thermal headroom rather than raw electrical draw. Benchmark results indicate that the chip is able to maintain its 2.30 GHz boost clock under load, but the data does not specify sustained clock behavior. For a server chassis with constrained airflow, the 65W TDP is a significant advantage, allowing for high-density deployments in 1U or small-form-factor enclosures. The 14nm process node and 3,200 million transistors on a 246 mm² die contribute to this efficiency.

Who Should Consider It

The Xeon D-1587 is best suited for workloads that leverage its 16 cores and 32 threads. In Cinebench R23 multi-core, the score of 11349 indicates strong performance for video rendering, 3D modeling, and batch processing tasks that can utilize all available threads. For content creation, the chip will handle multi-threaded encodes and compiles with ease, though single-threaded tasks like legacy plugin processing or light scripting will feel slower than a modern high-clock part.

For gaming, the data is less favorable. The R23 single-core score of 1602 is low by modern standards, and most games rely heavily on single-thread performance. The D-1587 would bottleneck in gaming scenarios, particularly at high frame rates, making it a poor choice for a gaming rig. Its market segment is explicitly Server/Workstation, and the benchmark profile aligns with that: it is a compute-density part, not a latency-optimized one.

Office and general productivity workloads that use a single thread for most interactions will not see a benefit from the 16 cores. The chip's strength lies in parallel execution — virtualization hosts, database servers, and scientific computing. For those use cases, the 65W TDP and high core count make it an attractive option for always-on systems where power consumption and heat are concerns.

Single-Thread vs Multi-Thread Behavior

The split between the D-1587's multi-core and single-core scores is the defining characteristic of this processor. In Cinebench R23, the multi-core score of 11349 is roughly 7.1 times the single-core score of 1602. This is a much higher ratio than typical desktop processors, which often sit between 4:1 and 5:1. The data shows that the D-1587 prioritizes throughput over responsiveness.

This behavior translates directly to real workloads. Applications that are parallelized — such as video encoders, 3D renderers, and scientific simulators — will see near-linear scaling up to 32 threads, making the D-1587 a strong performer. Conversely, applications that are latency-bound or single-threaded, such as web servers with low concurrency or interactive design tools, will leave most of the chip's capabilities idle. The 2.30 GHz boost clock is modest, and the base clock of 1700.00 MHz is low, so even lightly threaded tasks will run at speeds far below what a modern desktop chip offers.

The practical advice is to match workloads to the chip's profile. If a task can use all 32 threads, the D-1587 delivers performance comparable to much newer CPUs. If a task cannot, the chip will feel sluggish. This is a binary processor: excellent at parallel work, poor at serial work.

Platform and Compatibility

The Xeon D-1587 uses the Intel BGA 1667 socket, which means it is soldered to the motherboard and not upgradeable as a standalone processor. This is a significant platform consideration, as the entire system must be replaced to upgrade the CPU. The chip is based on the Broadwell architecture on a 14nm process, with a die size of 246 mm² and 3,200 million transistors.

Memory support includes both DDR3 and DDR4, with a dual-channel memory bus and ECC memory support. This dual-memory compatibility is unusual and provides flexibility for system builders using older or newer memory types. The platform includes PCIe Gen 3 with 24 lanes from the CPU, which is adequate for a server workload involving multiple NVMe drives or network cards, though it is not a high-lane-count platform for heavy GPU or accelerator expansion.

The processor has no integrated graphics, so a discrete GPU is required for any display output. The production status is Active, and the release date is February 2016, meaning this is an older but still supported platform. The lack of an unlocked multiplier means there is no overclocking headroom; the chip runs at its specified base and boost clocks. For upgrade path, the BGA 1667 socket limits options to the Xeon D series, so buyers should consider their long-term compute needs before committing to this platform, as the motherboard and CPU are a single unit.