CPU Comparison

The AMD A10-7890K and Intel Xeon E5620 are both end-of-life processors from different eras, yet benchmark data shows them landing in nearly the same performance tier. The A10-7890K wins every single head-to-head Cinebench test, but the margins are razor-thin, ranging from 0.5% to 0.7%. Both CPUs sit at the 28th percentile of all CPUs, and their average benchmark scores are 1035 for the AMD part versus 1029 for the Intel part. The verdict from the data is clear: if you need a socketed desktop chip with integrated graphics and an unlocked multiplier, the A10-7890K is the choice; if you require ECC memory support and a triple-channel memory bus for a server or workstation, the Xeon E5620 is the only option that fits.

The Verdict

The data shows a statistical dead heat in raw compute performance. The AMD A10-7890K edges out the Intel Xeon E5620 in all five recorded Cinebench runs, but the largest margin is just 0.7% in Cinebench R15 multicore (303 vs 301) and Cinebench R23 multicore (3009 vs 2989). The single-core results are equally tight, with the A10 leading 178 vs 177 in R20 single-core and 424 vs 422 in R23 single-core. These deltas are within run-to-run variance, so the benchmarks effectively declare a tie in processing capability.

The deciding factors are not raw performance but platform features. The AMD A10-7890K is built for desktop use with a Radeon R7 integrated GPU, a 4.00 GHz base clock boosting to 4.30 GHz, and an unlocked multiplier for overclocking. The Intel Xeon E5620 is a server/workstation part with a lower 2.40 GHz base clock (2.67 GHz boost), but it compensates with 8 threads from 4 cores, ECC memory support, and a triple-channel DDR3 memory bus. The A10 has a 95W TDP while the Xeon draws 80W, making the Intel part more power-efficient on paper.

Pick the A10-7890K if you are building a desktop system where integrated graphics and overclocking matter more than memory reliability features. Pick the Xeon E5620 if your workload demands ECC memory validation and you can leverage the triple-channel memory bandwidth, even though that bandwidth figure is not provided in the data. The benchmark scores will not decide this matchup; platform compatibility and feature requirements will.

FAQ

Q: Which processor has a higher average benchmark score?

A: The AMD A10-7890K scores 1035 on average, which is 6 points higher than the Intel Xeon E5620’s 1029. That is a 0.6% difference, placing both at the 28th percentile of all CPUs.

Q: Does the Intel Xeon E5620 support ECC memory?

A: Yes. The Xeon E5620 has ECC memory support enabled, while the AMD A10-7890K does not support ECC memory. This is a critical differentiator for server or workstation reliability requirements.

Q: What are the memory channel differences between the two?

A: The AMD A10-7890K uses a dual-channel DDR3 memory bus with a theoretical bandwidth of 34.1 GB/s. The Intel Xeon E5620 uses a triple-channel DDR3 memory bus, though no bandwidth figure is listed for it in the data.

Q: Which processor has integrated graphics?

A: Only the AMD A10-7890K has integrated graphics, specifically a Radeon R7 GPU. The Intel Xeon E5620 has no integrated graphics, requiring a discrete GPU for display output.

Q: Are both CPUs unlocked for overclocking?

A: No. The AMD A10-7890K has an unlocked multiplier, allowing user overclocking. The Intel Xeon E5620 has a locked multiplier, preventing multiplier-based overclocking.

Q: How do their thread counts differ?

A: The AMD A10-7890K has 4 cores and 4 threads, while the Intel Xeon E5620 has 4 cores and 8 threads thanks to Hyper-Threading. Despite the Xeon’s thread advantage, the A10 wins every multicore benchmark in the head-to-head data.

Architecture Differences

The architectural gap between these two parts is substantial, reflecting their different release eras and design goals. The AMD A10-7890K uses the Steamroller architecture on a 28 nm process node from GlobalFoundries, packing 2,411 million transistors into a 245 mm² die. It is codenamed Godaveri and belongs to the A10 generation. In contrast, the Intel Xeon E5620 uses the Westmere architecture on Intel’s 32 nm process, with 1,170 million transistors on a 239 mm² die. The Xeon is part of the Westmere-EP generation.

Cache hierarchies differ significantly. The A10-7890K has 256 KB of L1 cache and 4 MB of L2 cache, with no L3 cache at all. The Xeon E5620 has a per-core L1 allocation of 64 KB per core, 256 KB of L2 per core, and a shared 12 MB L3 cache. The Xeon’s 12 MB L3 gives it a substantial cache advantage, yet this does not translate into benchmark wins in the recorded tests.

The memory architecture also diverges. The A10 supports dual-channel DDR3 with a listed bandwidth of 34.1 GB/s, while the Xeon uses triple-channel DDR3 with no bandwidth figure provided. The Xeon supports ECC memory; the A10 does not. PCIe support also differs, with the A10 offering Gen 3 with 16 lanes from the CPU only, while the Xeon provides Gen 2 PCIe. These architectural choices reflect the A10’s desktop gaming focus versus the Xeon’s server reliability positioning.

Specification Differences

The core specifications show a clear split between a high-clock desktop part and a lower-clock workstation part. The AMD A10-7890K runs at a base clock of 4.00 GHz and boosts to 4.30 GHz, while the Intel Xeon E5620 operates at 2.40 GHz base and 2.67 GHz boost. The A10’s clocks are 67% higher at base and 61% higher at boost, which explains how it keeps pace with the Xeon’s thread advantage.

Thread counts differ: the A10 has 4 threads total, while the Xeon has 8 threads from 4 cores. TDP ratings also differ, with the A10 at 95W and the Xeon at 80W. Sockets are incompatible: the A10 uses AMD Socket FM2+ while the Xeon uses Intel Socket 1366. Process nodes differ by 4 nm, with the A10 on 28 nm and the Xeon on 32 nm. The A10 has an unlocked multiplier, the Xeon does not. The A10 includes integrated Radeon R7 graphics; the Xeon has none. Memory channel support differs (dual vs triple channel), and ECC support is present only on the Xeon. PCIe generations differ (Gen 3 vs Gen 2). The A10 is marked for Desktop market segment, while the Xeon is marked for Server/Workstation.

Head-to-Head Benchmarks

The head-to-head data shows the AMD A10-7890K winning all five recorded Cinebench tests, but by margins so narrow they are nearly negligible. In Cinebench R15 multicore, the A10 scores 303 against the Xeon’s 301, a 0.7% delta. Cinebench R20 multicore shows 1263 vs 1255, a 0.6% advantage for AMD. The single-core results are even closer: R20 single-core is 178 vs 177 (0.6% delta), and R23 single-core is 424 vs 422 (0.5% delta). The largest gap appears in Cinebench R23 multicore, where 3009 vs 2989 is a 0.7% difference.

These results are remarkable given the architectural disparity. The Xeon has twice the thread count (8 vs 4) and a 12 MB L3 cache, yet it still loses multicore tests. The A10 compensates with its much higher clocks, 4.00 GHz base versus 2.40 GHz, and its 4 MB L2 cache. The Xeon’s thread advantage does not overcome the clock deficit in these Cinebench workloads, which appear to favor higher per-core frequency over parallel throughput at this scale.

The A10 also wins single-core tests, which is expected given its 4.30 GHz boost versus the Xeon’s 2.67 GHz boost. The 0.5% to 0.6% deltas in single-core are smaller than the multicore deltas, suggesting that the Xeon’s architecture is relatively competitive per-clock in single-threaded work, but the A10’s clock advantage still carries the day. Overall, the AMD part wins 5 benchmarks to 0, but the data supports a conclusion of performance parity rather than dominance.

Where Each One Wins

The AMD A10-7890K wins in every benchmark category recorded, so its advantage is universal across Cinebench R15, R20, and R23 in both single-core and multicore tests. This makes it the better choice for any application that relies on raw CPU throughput as measured by Cinebench, which includes 3D rendering and similar multi-threaded workloads. The A10 also wins on platform features for desktop users: it has integrated Radeon R7 graphics, an unlocked multiplier for overclocking, and a 34.1 GB/s memory bandwidth figure that is explicitly listed. For users who want a single socket that handles both CPU and basic graphics tasks, the A10 is the clear winner.

The Intel Xeon E5620 does not win a single benchmark in the head-to-head data, but it wins on platform capabilities that the benchmarks do not measure. It supports ECC memory, which the A10 does not, making it the only choice for error-correcting memory in reliability-critical systems. Its triple-channel memory bus provides a memory architecture advantage over the A10’s dual-channel setup, even though no bandwidth number is provided for the Xeon. The Xeon also has a lower TDP at 80W versus 95W, making it more power-efficient. With 8 threads, the Xeon may handle heavily threaded server workloads better than the A10’s 4 threads, although the Cinebench data does not show this advantage materializing. For server or workstation builds requiring ECC, triple-channel memory, or Socket 1366 compatibility, the Xeon is the only viable option despite losing every benchmark test.