CPU Comparison

The Intel Xeon D-1587 and Intel Xeon E-2146G present a fascinating contrast in server processor design. The data reveals a consistent, albeit narrow, victory for the older, higher-core-count D-1587 across all Cinebench tests, while the E-2146G counters with a higher clock speed, a more recent architecture, and integrated graphics. This analysis explores what those benchmark deltas mean in practice, where each processor excels, and which workloads favor one over the other.

Head-to-Head Benchmarks

The head-to-head results are remarkably one-sided in favor of the Intel Xeon D-1587, though the margins are consistently slim. Across all six Cinebench tests, covering both R15, R20, and R23 generations, the D-1587 emerges as the winner with a uniform deltaPct of 3.3% in the multi-core tests and a 3.1% or 3.3% margin in single-core tests. This is a curious outcome given the E-2146G’s substantial clock speed advantage.

In Cinebench R23 multi-core, the D-1587 scores 11585 against the E-2146G’s 11214, a 3.3% lead. The single-core R23 test shows a similar story: 1635 versus 1583, again a 3.3% edge for the D-1587. This pattern repeats in R20 (4865 vs 4709 multi-core; 686 vs 664 single-core) and R15 (1167 vs 1130 multi-core; 164 vs 159 single-core). The consistency of the 3.3% delta across different Cinebench versions suggests a fundamental throughput advantage for the D-1587 that scales with workload intensity.

What makes this interesting is the hardware disparity. The D-1587 runs at a base clock of 1700 MHz with a boost of 2.30 GHz, while the E-2146G operates at 3500 MHz base and boosts to 4.50 GHz. Despite the E-2146G’s 2x+ clock advantage, the D-1587’s 16 cores and 32 threads overwhelm the E-2146G’s 6 cores and 12 threads in these heavily threaded benchmarks. The single-core results are more surprising, the D-1587 still wins, suggesting that its Broadwell architecture with 1.5 MB L3 per core is more efficient per clock than Coffee Lake in this specific test suite.

The avgBenchmarkScore field tells a slightly different story, with the E-2146G at 3410 versus the D-1587’s 3350, a 1.8% advantage for the E-2146G when including Geekbench results. The E-2146G has Geekbench scores (6283 multi-core, 1540 single-core) that the D-1587 lacks in the the benchmark database, which likely lifts its average. This divergence between Cinebench and Geekbench performance hints at workload-specific strengths.

Where Each One Wins

The benchmark data suggests the Intel Xeon D-1587 dominates in sustained, heavily threaded compute workloads. Its 16 cores and 32 threads, combined with 1.5 MB L3 cache per core, make it a clear choice for multi-threaded rendering, scientific simulations, and virtualization environments where parallel throughput is paramount. The 3.3% lead in every Cinebench multi-core test reinforces this, as Cinebench scales well with core count.

The Intel Xeon E-2146G, despite losing every head-to-head benchmark, offers strengths not captured in the Cinebench results. Its 4.50 GHz boost clock suggests superior responsiveness for lightly threaded tasks, though the single-core Cinebench tests contradict this expectation. The E-2146G includes integrated HD Graphics P630, a feature entirely absent from the D-1587, making it viable for workloads requiring basic display output or GPU-accelerated encoding without a discrete card. Its memory bandwidth of 42.7 GB/s is explicitly listed, while the D-1587’s bandwidth is not specified in the the benchmark database.

The E-2146G’s Geekbench scores (6283 multi-core, 1540 single-core) provide additional data points where it demonstrates competitive performance, though no direct comparison is possible without D-1587 Geekbench results. The E-2146G’s 12 MB shared L3 cache versus the D-1587’s per-core 1.5 MB suggests better cache utilization for single-threaded data locality, even if Cinebench doesn't reveal this advantage.

FAQ

Q: Which processor has more cores and threads?

A: The Intel Xeon D-1587 has 16 cores and 32 threads, while the Intel Xeon E-2146G has 6 cores and 12 threads.

Q: What is the clock speed difference between the two?

A: The D-1587 has a base clock of 1700 MHz and boost clock of 2.30 GHz, whereas the E-2146G has a base clock of 3500 MHz and boost clock of 4.50 GHz.

Q: Do both processors support ECC memory?

A: Yes, both the Intel Xeon D-1587 and Intel Xeon E-2146G have ECC memory support enabled.

Q: Which processor wins in Cinebench R23 multi-core?

A: The Intel Xeon D-1587 wins with a score of 11585, compared to the E-2146G’s 11214, a 3.3% difference.

Q: Does the E-2146G have integrated graphics?

A: Yes, the Intel Xeon E-2146G includes HD Graphics P630, while the D-1587 has no integrated graphics listed.

Q: What are the release dates for these processors?

A: The Intel Xeon D-1587 was released on 2016-02-23, and the Intel Xeon E-2146G was released on 2018-07-11.

Specification Differences

The two processors differ in nearly every specification category. The D-1587 uses the Intel BGA 1667 socket, while the E-2146G uses Intel Socket 1151. The D-1587 has 16 cores/32 threads versus the E-2146G’s 6 cores/12 threads. Clock speeds differ dramatically: 1700 MHz base/2.30 GHz boost for the D-1587 against 3500 MHz base/4.50 GHz boost for the E-2146G.

Thermal design power differs as well, with the D-1587 rated at 65 TDP and the E-2146G at 80 TDP. Memory support varies: the D-1587 supports both DDR3 and DDR4, while the E-2146G supports only DDR4. Both use dual-channel memory buses, but the E-2146G has a listed memory bandwidth of 42.7 GB/s, a figure absent for the D-1587. PCIe lane counts differ, with the D-1587 offering Gen 3, 24 Lanes versus the E-2146G’s Gen 3, 16 Lanes.

Cache hierarchies diverge significantly. The D-1587 has 1.5 MB L3 per core, while the E-2146G has 12 MB shared L3. The E-2146G includes integrated HD Graphics P630; the D-1587 has none. Production status differs: the D-1587 is Active, while the E-2146G is End-of-life. The D-1587 has a die size of 246 mm² and 3,200 million transistors; the E-2146G has a 154 mm² die with no transistor count listed. Launch MSRP for the D-1587 is 1549, and for the E-2146G is $317.

Architecture Differences

The architectural divide is generational and structural. The D-1587 is built on Broadwell architecture (codename Broadwell, generation Xeon D / Broadwell-DE), while the E-2146G uses Coffee Lake architecture (codename Coffee Lake-S WS, generation Xeon E / Coffee Lake). Both are fabricated on Intel’s 14 nm process node and use the same foundry.

Core and cache designs differ fundamentally. The D-1587’s 16 cores each have 64 KB L1 and 256 KB L2, with 1.5 MB L3 per core, totaling a massive 24 MB L3 across all cores. The E-2146G’s 6 cores share 12 MB L3, with the same per-core L1 and L2 allocations. This per-core versus shared L3 approach reflects different design philosophies: the D-1587 optimizes for isolated core performance in multi-tenant virtualized environments, while the E-2146G favors shared cache for data reuse across fewer, faster cores.

The D-1587 supports both DDR3 and DDR4 memory, a legacy feature from its 2016 release, while the E-2146G is DDR4-only. The D-1587’s 24 PCIe Gen 3 lanes exceed the E-2146G’s 16 lanes, making it better suited for high-bandwidth I/O configurations. Neither processor has an unlocked multiplier, and both are server/workstation market segments. The D-1587’s higher transistor count (3,200 million) and larger die (246 mm²) reflect its 16-core design, while the E-2146G’s 154 mm² die is more compact.

The Verdict

The data points to distinct use cases. The Intel Xeon D-1587 is the clear choice for workloads that prioritize raw multi-threading and can leverage 16 cores and 32 threads. Its 3.3% lead in every Cinebench multi-core test, despite a significantly lower clock speed, demonstrates superior parallel execution efficiency. The 24 PCIe Gen 3 lanes and dual memory support (DDR3 and DDR4) provide flexibility for diverse server configurations. Its 65 TDP, lower than the E-2146G’s 80 TDP, suggests better power efficiency per core, though no direct power consumption data is available.

The Intel Xeon E-2146G, while losing all head-to-head Cinebench tests, offers a different value proposition. Its 4.50 GHz boost clock and 12 MB shared L3 cache indicate strong performance for lightly threaded or latency-sensitive applications. The integrated HD Graphics P630 eliminates the need for a discrete GPU in display or basic acceleration scenarios. Its newer Coffee Lake architecture, released in 2018 versus 2016, brings architectural refinements that may benefit specific workloads not represented in the Cinebench suite. The E-2146G’s average benchmark score of 3410, which includes Geekbench results, suggests broader workload coverage.

For users choosing between these two, the benchmark results are clear: if Cinebench-style multi-threaded rendering and compute are the primary workloads, the D-1587 wins every test. If single-threaded responsiveness, integrated graphics, or a newer platform are priorities, the E-2146G offers those features, albeit with lower multi-threaded scores. The D-1587’s Active production status versus the E-2146G’s End-of-life status also factors into long-term availability and support considerations. Ultimately, the 3.3% performance delta is consistent but modest, meaning the decision should hinge on the specific feature requirements, cores, graphics, memory support, and PCIe lanes, rather than on benchmark supremacy alone.