Intel Core i7-5960X vs Intel Xeon D-1587 Comparison

The Intel Xeon D-1587 and the Intel Core i7-5960X present a fascinating study in contrasting design philosophies. The Xeon D-1587 is a high-density, power-efficient server processor built for scale-out deployments, while the i7-5960X is a desktop enthusiast flagship designed for raw compute and overclocking. Despite their different origins, the benchmark data reveals they land in the same performance tier, making their architectural divergences the primary deciding factor.

Where Each One Wins

The recorded data shows that the Intel Xeon D-1587 wins all six head-to-head benchmark comparisons. However, the margin of victory in every single test is negligible, ranging from 0.0% to 0.2%. In the Cinebench R15 multicore test, the Xeon D-1587 scores 1167 against the i7-5960X's 1165, a 0.2% difference. The Cinebench R20 multicore results are nearly identical, with the Xeon at 4865 and the i7 at 4857, again a 0.2% lead. The Cinebench R23 multicore test shows the Xeon D-1587 at 11585 versus the i7-5960X's 11565, another 0.2% margin.

The single-core tests tell the same story. In Cinebench R15 single-core, both processors score exactly 164, a statistical tie. The Cinebench R20 single-core test shows the Xeon D-1587 at 686 and the i7-5960X at 685, a 0.1% difference. The Cinebench R23 single-core test has the Xeon at 1635 and the i7 at 1632, a 0.2% gap. The data does not indicate a clear use-case split based on benchmark wins, as one processor does not dominate a specific type of workload. Instead, the results suggest that in these CPU-centric rendering tests, the two processors are functionally equivalent, with the Xeon D-1587 holding a razor-thin advantage across the board.

Architecture Differences

The architectures could not be more different in their foundational design. The Intel Xeon D-1587 uses the Broadwell architecture, specifically the Broadwell-DE generation, built on a 14 nm process node. It contains 16 cores and 32 threads. The Intel Core i7-5960X uses the older Haswell architecture, specifically Haswell-E, built on a 22 nm process node. It contains 8 cores and 16 threads. This means the Xeon D-1587 packs twice the core count into a smaller die size of 246 mm², compared to the i7-5960X's 356 mm². The Xeon also has a higher transistor count at 3,200 million versus 2,600 million for the i7.

The cache layouts are fundamentally different. The Xeon D-1587 relies on a per-core cache design, with 64 KB of L1 and 256 KB of L2 per core, plus 1.5 MB of L3 per core, which is not pooled into a single shared block. The i7-5960X uses a shared L3 cache of 20 MB for all cores. The Xeon D-1587 supports both DDR3 and DDR4 memory through a dual-channel memory bus, while the i7-5960X supports only DDR4 but through a quad-channel memory bus, giving it a recorded memory bandwidth of 68.3 GB/s. The Xeon D-1587 supports ECC memory, a critical feature for server reliability, while the i7-5960X does not.

The platform targets also differ sharply. The Xeon D-1587 uses the Intel BGA 1667 socket, is classified as a Server/Workstation part, and remains in Active production status. It has 24 PCIe Gen 3 lanes. The i7-5960X uses the Intel Socket 2011-3, is a Desktop part, and is End-of-life. It offers 40 PCIe Gen 3 lanes. The i7-5960X has an unlocked multiplier, enabling overclocking, while the Xeon D-1587's multiplier is locked. The Xeon D-1587 has a launch MSRP of 1549.

FAQ

Q: Which processor has more cores?

A: The Intel Xeon D-1587 has 16 cores and 32 threads, while the Intel Core i7-5960X has 8 cores and 16 threads.

Q: How do they compare in single-core performance?

A: The recorded single-core benchmarks are virtually identical. The Xeon D-1587 scores 164, 686, and 1635 in Cinebench R15, R20, and R23 respectively, while the i7-5960X scores 164, 685, and 1632 in the same tests.

Q: Does the Core i7-5960X support ECC memory?

A: No, the Intel Core i7-5960X does not support ECC memory. The Intel Xeon D-1587 does support ECC memory.

Q: What is the difference in memory bandwidth?

A: The Intel Core i7-5960X uses a quad-channel memory bus and has a recorded memory bandwidth of 68.3 GB/s. The Intel Xeon D-1587 uses a dual-channel memory bus and has no recorded memory bandwidth figure in the database.

Q: Which processor can be overclocked?

A: The Intel Core i7-5960X has an unlocked multiplier, making it overclockable. The Intel Xeon D-1587 has a locked multiplier.

Q: What is the production status of each processor?

A: The Intel Xeon D-1587 is listed as Active in production, while the Intel Core i7-5960X is listed as End-of-life.

Specification Differences

The two processors diverge on nearly every core specification. The Xeon D-1587 has 16 cores and 32 threads, while the i7-5960X has 8 cores and 16 threads. The base clock of the Xeon D-1587 is 1700.00 MHz, but the i7-5960X starts at a much higher 3.00 GHz. The boost clock follows the same pattern, with the Xeon at 2.30 GHz and the i7 at 3.50 GHz. The Xeon D-1587 has a TDP of 65 watts, while the i7-5960X has a TDP of 140 watts.

The process technology differs, with the Xeon D-1587 on a 14 nm node and the i7-5960X on a 22 nm node. The Xeon D-1587 has a die size of 246 mm² and 3,200 million transistors, while the i7-5960X has a die size of 356 mm² and 2,600 million transistors. The cache hierarchy is also distinct, with the Xeon using per-core L3 of 1.5 MB per core and the i7 using a 20 MB shared L3 cache. Memory support and bus configuration differ, with the Xeon supporting DDR3 and DDR4 on a dual-channel bus and the i7 supporting only DDR4 on a quad-channel bus. The PCIe lane count is 24 for the Xeon and 40 for the i7. The Xeon D-1587 supports ECC memory, the i7-5960X does not. The Xeon has a locked multiplier, the i7 has an unlocked multiplier. The Xeon is an Active server part on the BGA 1667 socket, while the i7 is an End-of-life desktop part on Socket 2011-3.

Head-to-Head Benchmarks

The head-to-head data shows an extraordinary level of parity. In Cinebench R15 multicore, the Xeon D-1587 scores 1167 against the i7-5960X's 1165, a 0.2% lead. This is a surprising result given that the Xeon D-1587 has twice the core count, but the i7-5960X's higher 3.00 GHz base and 3.50 GHz boost clocks clearly compensate for the core deficit. The Cinebench R20 multicore test repeats the pattern: 4865 for the Xeon versus 4857 for the i7, a 0.2% margin. The Cinebench R23 multicore test shows 11585 for the Xeon against 11565 for the i7, again a 0.2% difference.

The single-core results are even tighter. Cinebench R15 single-core is a perfect tie at 164 points each. Cinebench R20 single-core gives the Xeon D-1587 a 0.1% edge at 686 versus 685. Cinebench R23 single-core shows the Xeon at 1635 and the i7 at 1632, a 0.2% difference. These numbers indicate that the per-core performance of the i7-5960X is not superior to the Xeon D-1587 despite the higher clock speeds, which may reflect the architectural efficiency of the newer Broadwell design. The data shows that the Xeon D-1587 wins all six recorded head-to-head tests, but the largest margin of victory is only 0.2%.

The Verdict

The benchmark results indicate that for CPU rendering workloads, the Intel Xeon D-1587 and the Intel Core i7-5960X are effectively interchangeable in performance. The Xeon D-1587 wins all six head-to-head tests, but the margins are so small that they fall within typical run-to-run variance. The choice between these two processors should be based on platform requirements rather than raw benchmark scores. The Xeon D-1587 is the clear pick for server deployments that require ECC memory support, a 65 watt TDP for dense environments, and an active production lifecycle. The i7-5960X is the choice for a desktop build that needs the quad-channel memory bandwidth of 68.3 GB/s, 40 PCIe lanes for expansion, and an unlocked multiplier for overclocking. The Xeon D-1587's higher core count does not translate into a meaningful performance advantage in these tests, but its feature set is oriented toward reliability and efficiency. The i7-5960X offers a more flexible desktop platform with greater memory bandwidth and expansion capability, but it is end-of-life and consumes more power. The data shows a statistical tie in compute performance, so the deciding factors are the non-benchmark specifications.