Intel Core i7-990X vs Intel Xeon E5-1620 v3 Comparison

The Intel Core i7-990X and Intel Xeon E5-1620 v3 are two end-of-life processors from different eras, yet their benchmark scores place them in a near-perfect statistical tie. The data shows the i7-990X, a 2011 Gulftown desktop part, edges out the 2014 Haswell-EP workstation Xeon in every single head-to-head test, but by margins so slim—all under 3%—that the two effectively trade blows in overall average score. The Xeon E5-1620 v3 counters with a newer architecture, DDR4 memory support, and ECC capabilities, making this a fascinating clash between core count and platform modernity.

FAQ

Q: Which processor has more cores and threads?

A: The Intel Core i7-990X has 6 cores and 12 threads, while the Intel Xeon E5-1620 v3 has 4 cores and 8 threads. This gives the i7-990X a 50% advantage in core count and a 50% advantage in thread count.

Q: How close are their average benchmark scores?

A: The i7-990X has an average benchmark score of 1713, while the Xeon E5-1620 v3 scores 1709. That is a difference of just 0.2%, placing them virtually at parity in the database.

Q: What is the biggest single benchmark margin between them?

A: The largest delta in any head-to-head test is 2.8%, which occurs in Cinebench R23 single-core, where the i7-990X scores 857 versus the Xeon's 834. The other five tests show deltas between 2.4% and 2.7%.

Q: Do both processors support ECC memory?

A: No. The Xeon E5-1620 v3 supports ECC memory, while the Core i7-990X does not. This is a critical distinction for workstation reliability use cases.

Q: What generation of PCIe does each support?

A: The Core i7-990X supports PCIe Gen 2, while the Xeon E5-1620 v3 supports PCIe Gen 3 with 40 lanes (CPU only). The newer standard offers higher bandwidth for expansion cards.

Q: Are both processors in the same performance percentile?

A: Yes, both the i7-990X and the Xeon E5-1620 v3 sit at the 40th percentile against all CPUs in the database, confirming their similar overall standing.

Architecture Differences

The architectural gap between these two parts is substantial, spanning three years and a full process node shrink. The Core i7-990X is built on Intel's 32nm Westmere architecture (codename Gulftown), while the Xeon E5-1620 v3 uses the 22nm Haswell architecture (codename Haswell-EP). This node difference is reflected in transistor counts: the Xeon packs 2,600 million transistors on a 356 mm² die, versus 1,170 million transistors on a 239 mm² die for the i7-990X. Despite the Xeon's larger die and higher transistor density, it uses fewer cores (4 versus 6), suggesting a more complex per-core design.

The cache layouts also differ. Both share 64 KB of L1 and 256 KB of L2 per core, but the i7-990X has 12 MB of shared L3 cache, while the Xeon E5-1620 v3 has 10 MB of shared L3. Interestingly, the Xeon has less L3 despite its larger die, a trade-off likely tied to its different memory architecture. The Xeon supports quad-channel DDR4 memory with a rated bandwidth of 68.3 GB/s, whereas the i7-990X uses triple-channel DDR3 with no official bandwidth figure in the data. This memory subsystem overhaul is a major generational advantage for the Xeon.

The sockets are incompatible: the i7-990X uses Intel Socket 1366, while the Xeon E5-1620 v3 uses Intel Socket 2011-3. The market segments differ too—the i7-990X is a desktop part with an unlocked multiplier, while the Xeon is a server/workstation part with a locked multiplier. The Xeon also supports ECC memory and offers PCIe Gen 3 with 40 lanes, versus Gen 2 for the i7-990X. Power envelopes are close but not identical: the i7-990X has a 130W TDP, while the Xeon E5-1620 v3 has a 140W TDP. The i7-990X launched in February 2011, and the Xeon followed in September 2014.

Head-to-Head Benchmarks

The benchmark results tell a consistent story of narrow victory for the older i7-990X. Across all six Cinebench tests, the i7-990X wins, but the margins are remarkably tight. In Cinebench R15 multi-core, the i7-990X scores 611 against the Xeon's 595, a 2.7% lead. The single-core R15 test shows 86 versus 84, a 2.4% edge. Moving to R20, the multi-core result is 2549 versus 2482 (2.7%), and single-core is 359 versus 350 (2.6%). In R23, the multi-core gap is 6071 versus 5910 (2.7%), and the single-core margin expands slightly to 2.8% with scores of 857 versus 834.

The pattern is clear: the i7-990X's two extra cores and four extra threads provide a consistent, if small, advantage in multi-threaded workloads. However, the fact that it also wins single-core tests—where core count is irrelevant—is more surprising. This suggests the i7-990X's higher boost clock of 3.73 GHz (versus 3.60 GHz on the Xeon) may be doing the heavy lifting in single-threaded scenarios. The base clocks are nearly identical (3.47 GHz versus 3.50 GHz), so the boost differential of 0.13 GHz appears to be enough to tip the scales.

Geekbench results are only available for the i7-990X, showing a multi-core score of 2612 and single-core of 556. Without a corresponding Xeon score, these cannot be compared directly, but they contribute to the i7-990X's average benchmark score of 1713. The Xeon's average of 1709 comes solely from its Cinebench results. In the nearest rivals list, the i7-990X is 0.2% ahead of the Xeon, while the Xeon is 0.2% behind the i7-990X—a mirror image that reinforces their near-identical standing. Both CPUs share a 40th percentile ranking against all CPUs.

The Verdict

The data points to a clear, if counterintuitive, conclusion: the older Core i7-990X outperforms the newer Xeon E5-1620 v3 in every measured benchmark, but the differences are so marginal that they are unlikely to be perceptible in real-world use. The i7-990X's wins range from 2.4% to 2.8%, which is within the range of run-to-run variance for many workloads. The Xeon E5-1620 v3 does not win a single head-to-head test, yet its platform features—DDR4, quad-channel memory, ECC support, PCIe Gen 3—offer capabilities the i7-990X simply does not have.

For a pure performance-per-core and multi-threaded score perspective, the i7-990X is the better chip. Its 6 cores and 12 threads deliver a consistent edge in Cinebench, and its unlocked multiplier allows for overclocking, which the locked Xeon cannot offer. However, the Xeon's 68.3 GB/s memory bandwidth, ECC support, and 40 PCIe Gen 3 lanes make it the more robust platform for stability-critical or I/O-heavy workstation tasks. The choice hinges on whether raw compute scores or platform features matter more. The i7-990X wins the benchmark war; the Xeon wins the platform war.

Specification Differences

The two processors differ in nearly every major specification category. The i7-990X has 6 cores and 12 threads, while the Xeon E5-1620 v3 has 4 cores and 8 threads. Base clocks are effectively equal (3.47 GHz versus 3.50 GHz), but boost clocks differ (3.73 GHz versus 3.60 GHz). TDP is 130W for the i7-990X and 140W for the Xeon. The socket is different (1366 versus 2011-3), as is the architecture (Westmere versus Haswell) and process node (32nm versus 22nm). Transistor count is 1,170 million versus 2,600 million, and die size is 239 mm² versus 356 mm². L3 cache is 12 MB versus 10 MB. Memory support is DDR3 triple-channel versus DDR4 quad-channel, with the Xeon offering 68.3 GB/s bandwidth and ECC support. PCIe is Gen 2 versus Gen 3 with 40 lanes. The market segment is Desktop versus Server/Workstation, and the multiplier is unlocked versus locked. Release dates are February 2011 versus September 2014.

Where Each One Wins

The Intel Core i7-990X wins in all six Cinebench benchmarks, making it the clear choice for any workload that relies on raw multi-threaded or single-threaded compute as measured by these tests. Its 6-core/12-thread configuration provides a structural advantage in heavily parallel tasks, and its higher boost clock helps in lightly threaded scenarios. The unlocked multiplier is another win for enthusiasts who want to push performance beyond stock settings. If the goal is maximizing benchmark scores on a legacy platform, the i7-990X is the winner.

The Intel Xeon E5-1620 v3 wins in platform features rather than benchmarks. It supports ECC memory, which is essential for error-sensitive server or workstation environments where data integrity is paramount. Its quad-channel DDR4 memory with 68.3 GB/s bandwidth offers a significant upgrade in memory throughput compared to the i7-990X's triple-channel DDR3. The 40 PCIe Gen 3 lanes provide more bandwidth and connectivity for modern expansion cards, GPUs, and NVMe storage. The Xeon's newer 22nm architecture and higher transistor count also suggest better power efficiency per transistor, though its TDP is 10W higher. For users building a workstation that prioritizes memory capacity, ECC reliability, or PCIe expansion, the Xeon E5-1620 v3 is the more appropriate selection despite losing every compute benchmark.