Intel Xeon D-2733NT vs Intel Xeon E5-2698B v3 Comparison

The Intel Xeon E5-2698B v3 and Intel Xeon D-2733NT occupy the same performance tier in the database, but they represent radically different generations and design philosophies. The E5-2698B v3 is a 16-core Haswell-EP part from the end-of-life era of Socket 2011-3, while the D-2733NT is an active 8-core Ice Lake-D system-on-chip soldered to a BGA 2579 package. Their average benchmark scores are nearly identical—4004 for the E5-2698B v3 versus 3984 for the D-2733NT—yet the data reveals that the older chip wins every single head-to-head Cinebench test, albeit by a razor-thin 0.5% margin. This parity is remarkable given the E5-2698B v3’s double core count and its 135 W TDP, compared to the D-2733NT’s 80 W envelope. The question is not which is faster—the E5-2698B v3 is consistently ahead—but whether that marginal lead justifies the older platform’s trade-offs in power, process node, and platform longevity.

FAQ

Q: Which processor has more cores and threads?

A: The Intel Xeon E5-2698B v3 has 16 cores and 32 threads, while the Intel Xeon D-2733NT has 8 cores and 16 threads. The E5-2698B v3 offers double the parallel execution resources on paper.

Q: What is the performance difference in multi-core workloads?

A: The E5-2698B v3 leads by 0.5% in all three multi-core Cinebench tests: 1395 vs 1388 in R15, 5814 vs 5785 in R20, and 13843 vs 13775 in R23. The margin is consistent but negligible in real-world terms.

Q: How do their single-core scores compare?

A: The E5-2698B v3 also wins single-core tests by 0.5%: 196 vs 195 in R15, 820 vs 816 in R20, and 1954 vs 1944 in R23. Despite the D-2733NT’s newer Ice Lake architecture, it cannot overcome the older chip’s higher boost clock of 3.40 GHz versus 3.20 GHz.

Q: What are their respective average benchmark scores and percentiles?

A: The E5-2698B v3 has an average benchmark score of 4004 and sits at the 57th percentile of all CPUs. The D-2733NT averages 3984, also at the 57th percentile. Their nearest rivals are nearly identical in score, with the E5-2698B v3 trailing the AMD Ryzen 5 PRO 4655G by 0.2% and the D-2733NT matching the Intel Xeon E-2336 at 0% delta.

Q: How do their memory subsystems differ?

A: Both support DDR4 with quad-channel memory buses, but the D-2733NT has a higher theoretical memory bandwidth of 85.3 GB/s versus 68.3 GB/s for the E5-2698B v3. Both support ECC memory.

Q: Which processor is more power-efficient?

A: The D-2733NT has a TDP of 80 W, significantly lower than the E5-2698B v3’s 135 W. The D-2733NT also uses a 10 nm process node compared to the E5-2698B v3’s 22 nm node, which contributes to its lower power draw.

The Verdict

The data points to a clear but counterintuitive conclusion: the older, larger, and more power-hungry Intel Xeon E5-2698B v3 is the better performer in every benchmark recorded, yet the margin is so thin that it barely matters. With six wins out of six head-to-head tests, the E5-2698B v3 is technically the superior CPU for raw Cinebench throughput. However, the D-2733NT’s 0.5% deficit in every test means that in practice, users would be hard-pressed to notice a difference. The real decision hinges on platform characteristics. The E5-2698B v3 offers 16 cores and 32 threads for heavily parallel server workloads, plus 40 MB of shared L3 cache and 40 PCIe Gen 3 lanes. The D-2733NT counters with 8 cores, 15 MB of L3 cache, 32 PCIe Gen 4 lanes, and a much lower 80 W TDP. For a system builder prioritizing absolute performance and core count, the E5-2698B v3 is the pick. For those prioritizing power efficiency, newer architecture, and an active production status, the D-2733NT is the rational choice, especially given its 85.3 GB/s memory bandwidth advantage.

Head-to-Head Benchmarks

The head-to-head results are monotonous in their consistency: the Intel Xeon E5-2698B v3 wins every single test by exactly 0.5%. In Cinebench R15 multi-core, the E5-2698B v3 scores 1395 against the D-2733NT’s 1388—a 7-point gap. Single-core R15 sees 196 versus 195, a 1-point difference. Moving to R20, the multi-core scores are 5814 and 5785, while single-core is 820 versus 816. The R23 tests follow the same pattern: the E5-2698B v3 posts 13843 multi-core and 1954 single-core, while the D-2733NT manages 13775 and 1944, respectively.

The most striking aspect of these numbers is not the winner but the near-total parity. The E5-2698B v3 has 16 cores and a 3.40 GHz boost clock; the D-2733NT has 8 cores and a 3.20 GHz boost clock. Yet the older chip only leads by 0.5% across the board. This suggests that the D-2733NT’s Ice Lake architecture—built on a 10 nm process versus 22 nm—delivers vastly superior instructions-per-clock efficiency. The D-2733NT’s per-core L2 cache is 1.25 MB compared to 256 KB for the E5-2698B v3, and its L1 cache is 80 KB per core versus 64 KB. These architectural advantages allow the 8-core D-2733NT to nearly match the 16-core E5-2698B v3 in both single-core and multi-core workloads.

The average benchmark scores reinforce this picture. The E5-2698B v3 averages 4004, and its nearest rival is the AMD Ryzen 5 PRO 4655G at 4010 (a -0.2% delta). The D-2733NT averages 3984, with its closest rival being the Intel Xeon E-2336 at 3985 (a 0% delta). Both chips sit in the same performance neighborhood, and their head-to-head delta of 0.5% is smaller than the variance between some of their nearest rivals.

Specification Differences

The two processors diverge sharply on core configuration. The E5-2698B v3 offers 16 cores and 32 threads, while the D-2733NT halves that to 8 cores and 16 threads. Base clocks are close—2.00 GHz for the E5-2698B v3 versus 2.10 GHz for the D-2733NT—and boost clocks are 3.40 GHz versus 3.20 GHz. The TDP gap is significant: 135 W for the E5-2698B v3 versus 80 W for the D-2733NT. Socket compatibility is entirely different, with the E5-2698B v3 using Intel Socket 2011-3 and the D-2733NT using Intel BGA 2579 (a soldered design). The E5-2698B v3’s part number is SR21T, while the D-2733NT is SRLCJ. Production status also differs: the E5-2698B v3 is end-of-life, whereas the D-2733NT is active, with a release date of 2022-02-23. Neither chip has a launch MSRP listed in the data.

Architecture Differences

The architectural gap is generational. The E5-2698B v3 is built on the Haswell-EP architecture using a 22 nm process node from Intel’s foundry, with a die size of 356 mm² and 2,600 million transistors. The D-2733NT uses the Ice Lake-D architecture on a 10 nm process node, also from Intel’s foundry, though its die size and transistor count are not listed. Cache hierarchies differ substantially: the E5-2698B v3 has 64 KB of L1 and 256 KB of L2 per core, plus 40 MB of shared L3; the D-2733NT has 80 KB of L1 and 1.25 MB of L2 per core, but only 15 MB of shared L3. The D-2733NT’s larger per-core L2 cache is a hallmark of its newer design, allowing it to compensate for its lower core count. Both support DDR4 memory and ECC, but the D-2733NT’s memory bandwidth is 85.3 GB/s versus 68.3 GB/s for the E5-2698B v3. PCIe capabilities also differ: the E5-2698B v3 provides 40 lanes of Gen 3, while the D-2733NT provides 32 lanes of Gen 4. Neither chip has integrated graphics, and neither is multiplier-unlocked.

Where Each One Wins

The Intel Xeon E5-2698B v3 wins in raw multi-core throughput across every Cinebench version, from R15’s 1395 to R23’s 13843. Its 16 cores and 32 threads make it the better choice for heavily threaded server and workstation workloads that can saturate all available cores. The 40 MB of shared L3 cache also provides a larger pool of fast memory for data-intensive applications. Additionally, the E5-2698B v3 wins every single-core test—196, 820, and 1954 in R15, R20, and R23 respectively—meaning it holds a small but universal performance edge in the benchmarks recorded. For users who need the absolute highest Cinebench scores, the E5-2698B v3 is the winner.

The Intel Xeon D-2733NT wins on efficiency and platform modernity. Its 80 W TDP is 55 W lower than the E5-2698B v3’s 135 W, making it suitable for dense, power-constrained deployments. The 10 nm process node and Ice Lake architecture deliver near-identical performance with half the cores, underscoring a massive architectural advantage. The D-2733NT also offers higher memory bandwidth at 85.3 GB/s and PCIe Gen 4 support, which can benefit storage and network I/O in edge or network-attached environments. Its active production status means it remains a current product, whereas the E5-2698B v3 is end-of-life. The D-2733NT’s 15 MB of L3 is smaller, but its per-core L2 cache of 1.25 MB is five times larger than the E5-2698B v3’s, which may help in latency-sensitive workloads. Ultimately, the D-2733NT is the pick for new designs where power and platform longevity matter more than a 0.5% performance lead.