Intel Xeon D-1528 vs Intel Xeon E5-2623 v3 Comparison

FAQ

Q: Which CPU has more cores and threads?

A: The Intel Xeon D-1528 has 6 cores and 12 threads, while the Intel Xeon E5-2623 v3 has 4 cores and 8 threads. The D-1528 offers a 50% advantage in core count and thread count.

Q: How do the two chips compare in raw multicore performance?

A: Across all three Cinebench versions, the Xeon D-1528 leads. In R15 multicore, it scores 584 versus 568 (a 2.7% lead). In R20, it scores 2435 versus 2367 (2.8% ahead). In R23, it scores 5799 versus 5637 (2.8% ahead). The D-1528 wins every multicore test, but the margins are modest.

Q: Which CPU has the higher clock speed?

A: The Xeon E5-2623 v3 has a base clock of 3.00 GHz and a boost clock of 3.50 GHz. The Xeon D-1528 has a base clock of 1900.00 MHz (1.90 GHz) and a boost of 2.50 GHz. The E5-2623 v3 runs at significantly higher frequencies.

Q: How does power consumption differ between these two?

A: The Xeon D-1528 has a TDP of 35 watts, while the Xeon E5-2623 v3 has a TDP of 105 watts. The D-1528 uses dramatically less power, making it far better suited for dense or low-power deployments.

Q: Do both support ECC memory?

A: Yes, both the Xeon E5-2623 v3 and the Xeon D-1528 support ECC memory. However, the E5-2623 v3 uses quad-channel DDR4 with a memory bandwidth of 59.7 GB/s, while the D-1528 supports dual-channel DDR3 and DDR4 and has no listed bandwidth figure.

Q: Which processor has the higher overall benchmark percentile?

A: The Xeon D-1528 sits at the 40th percentile among all CPUs, while the Xeon E5-2623 v3 sits at the 39th percentile. The D-1528 also has a higher average benchmark score: 1677 versus 1630.

The Verdict

The data points to the Xeon D-1528 as the better performer in every measured benchmark. It wins all six Cinebench tests, both single-core and multi-core, with margins ranging from 2.7% to 3.7%. Its average benchmark score of 1677 is 47 points higher than the E5-2623 v3's 1630, and it lands at the 40th percentile versus the 39th percentile for the older chip.

The D-1528 also brings a substantial efficiency advantage. Its 35W TDP is one-third that of the E5-2623 v3's 105W TDP. For builders prioritizing low power draw, passive cooling, or compact systems, this is a decisive factor. The D-1528 is an active product, while the E5-2623 v3 is end-of-life.

However, the E5-2623 v3 is not without a case. It runs at a much higher clock speed (3.00 GHz base, 3.50 GHz boost versus 1.90 GHz base, 2.50 GHz boost), which can matter for lightly threaded workloads that respond to raw frequency. It also supports quad-channel memory with a rated bandwidth of 59.7 GB/s, and it uses the Socket 2011-3 platform, which offers 40 PCIe lanes versus 24 on the D-1528.

The choice depends on the platform priorities. If the goal is maximum performance per watt in a small form factor, the D-1528 is the clear winner. If the system requires the expandability of a large socket with more PCIe lanes and quad-channel memory, the E5-2623 v3 is the appropriate pick. For most general server and workstation loads based purely on the recorded benchmarks, the D-1528 is the stronger processor.

Head-to-Head Benchmarks

The head-to-head results are consistent: the Xeon D-1528 wins every single test. In Cinebench R15 multicore, the D-1528 scores 584 against 568 for the E5-2623 v3, a 2.7% advantage. In R15 single-core, it scores 82 versus 79, a 3.7% lead, which is the largest margin in the entire comparison.

Moving to Cinebench R20, the D-1528 again leads in multicore with 2435 versus 2367, a 2.8% difference. In single-core, it scores 343 versus 333, a 2.9% edge. The pattern holds in R23, where the D-1528 posts 5799 multicore against 5637, a 2.8% lead, and 818 single-core against 795, a 2.8% lead.

Notably, the E5-2623 v3 does not win a single benchmark. Despite its higher boost clock of 3.50 GHz, it trails in every single-core test. This suggests that the D-1528's newer 14 nm Broadwell architecture and additional cores more than compensate for its lower clock speed. Even in single-threaded workloads, where clock speed usually dominates, the D-1528's advantage holds.

The largest single-core delta is 3.7% in R15, while the multicore deltas hover around 2.7% to 2.8%. These are not large margins, but they are consistent. In practical terms, a user upgrading from the E5-2623 v3 to the D-1528 would see a small but measurable improvement across all Cinebench workloads, with the added benefit of far lower power consumption.

Specification Differences

The two processors differ in nearly every major specification category. The E5-2623 v3 uses 4 cores and 8 threads, while the D-1528 uses 6 cores and 12 threads. Clock speeds favor the E5-2623 v3: 3.00 GHz base and 3.50 GHz boost versus 1.90 GHz base and 2.50 GHz boost.

Power consumption is a major separator. The E5-2623 v3 has a 105W TDP, while the D-1528 has a 35W TDP. The socket types are incompatible: the E5-2623 v3 uses Intel Socket 2011-3, a large LGA platform, while the D-1528 uses Intel BGA 1667, a soldered embedded package.

Memory support differs as well. The E5-2623 v3 supports DDR4 only, with a quad-channel bus and a rated bandwidth of 59.7 GB/s. The D-1528 supports both DDR3 and DDR4, but only in dual-channel configuration, with no bandwidth figure listed. Both support ECC memory.

PCIe lane counts favor the E5-2623 v3, which offers 40 CPU-attached Gen 3 lanes. The D-1528 offers 24 Gen 3 lanes. Neither chip has integrated graphics.

The die and process details also differ. The E5-2623 v3 is built on Intel's 22 nm process with 2,600 million transistors on a 356 mm² die, while the D-1528 is built on 14 nm with 3,200 million transistors on a 246 mm² die. The D-1528 packs more transistors into a smaller area, reflecting the newer process node.

Cache layouts are different as well. Both have 64 KB L1 and 256 KB L2 per core, but the L3 differs: the E5-2623 v3 has 10 MB shared, while the D-1528 has 1.5 MB per core. With 6 cores, the D-1528's total L3 comes to 9 MB, slightly less than the E5-2623 v3's 10 MB.

The release timeline also separates them. The E5-2623 v3 launched in September 2014 with a launch MSRP of $444. The D-1528 launched in October 2015 with a launch MSRP of $320. The E5-2623 v3 is end-of-life, while the D-1528 remains active.

Architecture Differences

The E5-2623 v3 is built on the Haswell architecture, specifically Haswell-EP, and uses a 22 nm process node. The D-1528 is built on Broadwell, specifically Broadwell-DE, and uses a 14 nm process node. This process shrink allows the D-1528 to pack 3,200 million transistors onto a 246 mm² die, compared to 2,600 million on a 356 mm² die for the E5-2623 v3.

The D-1528's newer architecture and smaller process node explain its efficiency advantage. It delivers higher benchmark scores while consuming only 35W, a 70W reduction from the E5-2623 v3's 105W TDP. The Broadwell generation also brings architectural refinements that improve instructions per clock, which helps the D-1528 stay competitive in single-threaded tests despite its much lower clock speed.

Cache architecture differs in an important way. The E5-2623 v3 uses a shared 10 MB L3 cache, while the D-1528 uses 1.5 MB per core. This per-core allocation can be beneficial for workloads that need dedicated cache slices, while the shared design on the E5-2623 v3 may be better for data shared across cores. In practice, the benchmark results show the D-1528's approach works well enough to edge ahead.

The memory controller also reflects the platform difference. The E5-2623 v3's quad-channel DDR4 controller provides 59.7 GB/s of bandwidth, which is a significant advantage for memory-intensive server workloads. The D-1528's dual-channel controller, supporting both DDR3 and DDR4, offers less raw bandwidth but more flexibility in memory type selection.

The socket and platform architecture are fundamentally different. The E5-2623 v3 is a socketed processor for the 2011-3 platform, which supports 40 PCIe Gen 3 lanes and is designed for large servers and workstations with extensive expansion needs. The D-1528 is a soldered BGA part intended for embedded and compact systems, with 24 PCIe Gen 3 lanes. This makes the D-1528 a system-on-chip style solution, while the E5-2623 v3 is a traditional server CPU.

Where Each One Wins

The Xeon D-1528 wins in every measured benchmark, so the performance case is clear. It is the better choice for compute workloads that can use six cores and twelve threads. Its 35W TDP makes it ideal for low-power servers, network appliances, storage systems, and other compact deployments where heat dissipation and power budgets are tight. The fact that it is still an active product means it remains available for new designs.

The D-1528 also has the advantage in single-threaded performance, despite its lower clock speed. Its 3.7% lead in R15 single-core and 2.8% lead in R23 single-core show that the Broadwell architecture's efficiency offsets the clock deficit. For workloads that are only lightly threaded, the D-1528 still comes out ahead.

The Xeon E5-2623 v3 wins on platform features rather than raw benchmark scores. Its quad-channel memory bus with 59.7 GB/s of bandwidth is a genuine advantage for memory-bound server applications such as databases, virtualization hosts, or large in-memory workloads. The 40 PCIe Gen 3 lanes also allow for more expansion cards, NVMe drives, or GPU accelerators than the D-1528's 24 lanes.

The E5-2623 v3's higher clock speeds, while not enough to win benchmarks, can still be relevant in niche scenarios where frequency scaling behaves differently than Cinebench suggests. Some legacy software may respond well to the 3.50 GHz boost clock, and the socketed 2011-3 platform offers upgrade paths and repairability that a soldered BGA chip cannot match.

For a new server build, the D-1528 is the default recommendation. For an existing 2011-3 platform upgrade, the E5-2623 v3 remains a capable option, especially if the system already has a motherboard with many PCIe devices installed. Builders prioritizing density and efficiency should pick the D-1528. Builders needing maximum memory bandwidth and expansion should pick the E5-2623 v3.