AMD EPYC 4564P vs Intel Xeon 6736P Comparison

Head-to-Head Benchmarks

The benchmark data presents a striking contrast between these two server processors. Across the Cinebench suite, the AMD EPYC 4564P consistently delivers a 30.6% advantage over the Intel Xeon 6736P in every single test. This includes both single-core and multi-core workloads, with the AMD part scoring 5603 versus 4290 in Cinebench R15 multi-core, 23348 versus 17875 in R20 multi-core, and 55592 versus 42561 in R23 multi-core. The single-core results follow the same pattern: 791 versus 605 in R15, 3296 versus 2523 in R20, and 7848 versus 6008 in R23.

The most dramatic gap appears in PassMark's single-thread test, where the EPYC 4564P scores 4292 compared to the Xeon's 2024, a 112.1% difference. This is more than double the performance, which suggests fundamental differences in how each processor handles single-threaded execution. The AMD chip also wins the PassMark multi-thread test with 64357 against 50072, a 28.5% margin.

The Intel Xeon 6736P does claim victories in four benchmarks. The most significant is PassMark physics, where it scores 6531 against the AMD part's 3392, a 48.1% advantage. The Xeon also wins floating point math with 149770 versus 141017 (5.8% ahead), find prime numbers with 392 versus 365 (6.9% ahead), and random string sorting with 103723 versus 103202, a narrow 0.5% margin.

In the remaining workloads, the AMD EPYC 4564P shows consistent superiority. Data compression favors AMD at 858105 versus 796658 (7.7% ahead), data encryption goes to AMD at 50708 versus 46236 (9.7% ahead), extended instructions show AMD winning 63136 versus 55563 (13.6% ahead), and integer math sees AMD ahead at 228761 versus 206833 (10.6% ahead). Overall, the AMD part wins 13 benchmarks while the Intel part wins 4.

Where Each One Wins

The data reveals a clear split in workload characteristics. The AMD EPYC 4564P dominates in scenarios that reward high clock speeds and efficient instruction execution. Its Cinebench sweep, particularly the single-core results, indicates strong performance in lightly threaded applications, database queries, and latency-sensitive tasks. The 112.1% lead in PassMark single-thread performance suggests that software relying on single-threaded responsiveness will see dramatic improvements with the AMD part.

The AMD processor also handles memory-intensive and cryptographic workloads well. Its wins in data compression, encryption, and extended instructions point to strong integer and data-processing capabilities. The multithreaded benchmark victory at 28.5% further reinforces that the AMD chip scales effectively across its 16 cores and 32 threads.

The Intel Xeon 6736P carves out its niche in physics simulation and floating-point mathematics. The 48.1% advantage in PassMark physics indicates that scientific computing, simulation workloads, and physics engines will favor the Intel part. The floating point math win, while smaller at 5.8%, reinforces this pattern. The narrow win in random string sorting and the find prime numbers result suggest that certain algorithmic patterns, possibly those benefiting from the Xeon's larger core count of 36 cores and 72 threads, can overcome the clock speed disadvantage.

For workloads that are purely core-count driven and can fully utilize 36 cores, the Intel part shows it can compete in specific mathematical domains. However, the AMD processor's wins in general multithreaded performance, despite having fewer cores, indicate that its higher clock speeds and architectural efficiency compensate significantly.

FAQ

Q: Which processor has the higher boost clock speed?

A: The AMD EPYC 4564P has a boost clock of 5.70 GHz, while the Intel Xeon 6736P reaches 4.10 GHz. This 1.6 GHz difference helps explain the AMD part's 30.6% lead across all Cinebench tests.

Q: How do the core counts compare between these two chips?

A: The Intel Xeon 6736P has 36 cores and 72 threads, more than double the AMD EPYC 4564P's 16 cores and 32 threads. Despite this, the AMD processor wins the PassMark multi-thread test with 64357 versus 50072.

Q: Which processor supports more PCIe lanes?

A: The Intel Xeon 6736P provides Gen 5 with 88 lanes, while the AMD EPYC 4564P offers Gen 5 with 28 lanes. This represents a substantial expansion capability difference for peripheral connectivity.

Q: What is the memory bandwidth difference?

A: The Intel Xeon 6736P has an eight-channel memory bus delivering 409.6 GB/s, compared to the AMD EPYC 4564P's dual-channel configuration at 83.2 GB/s. The Intel part offers nearly five times the memory bandwidth.

Q: How do the average benchmark scores compare?

A: The AMD EPYC 4564P has an average benchmark score of 95183, while the Intel Xeon 6736P averages 87864. Both processors sit at the 96th percentile among all CPUs, but the AMD part holds the higher average score.

Q: Does either processor include integrated graphics?

A: The AMD EPYC 4564P includes Radeon Graphics, while the Intel Xeon 6736P has no integrated graphics (listed as N/A). This could matter for systems requiring basic display output without a discrete GPU.

Specification Differences

The two processors diverge significantly in their base specifications. The AMD EPYC 4564P operates at a 4.50 GHz base clock with a 5.70 GHz boost, while the Intel Xeon 6736P runs at 2.00 GHz base and 4.10 GHz boost. Power consumption differs as well, with the AMD part rated at 170W TDP versus the Intel part's 205W TDP.

Socket compatibility is entirely different: the AMD chip uses AMD Socket AM5, while the Intel chip requires Intel Socket 4710. The memory configuration shows the Intel part with an eight-channel bus and 409.6 GB/s bandwidth, compared to the AMD part's dual-channel setup and 83.2 GB/s. PCIe lane counts also differ substantially, with Intel offering 88 Gen 5 lanes versus AMD's 28 Gen 5 lanes.

The release dates place the AMD processor first at 2024-05-20, with Intel following on 2025-02-23. The launch MSRP for the AMD EPYC 4564P is $699, while the Intel Xeon 6736P has a launch MSRP of $3351. Die size varies considerably: the AMD chip uses 2x 71 mm² dies, while the Intel chip uses a single 598 mm² die. The AMD processor's part number is 100-000001476, and the Intel part's is SRVNW.

Architecture Differences

The architectural foundations of these processors reflect different design philosophies. The AMD EPYC 4564P uses Zen 4 architecture under the Raphael codename, part of the EPYC 4004 series. It is built on a 5 nm process at TSMC, with 13,140 million transistors distributed across two 71 mm² dies. The Intel Xeon 6736P uses Granite Rapids architecture, part of the Xeon 6 generation, fabricated on Intel's 5 nm process with a single 598 mm² die.

Cache hierarchies differ notably. The AMD part provides 64 KB of L1 cache per core, 1 MB of L2 cache per core, and 64 MB of shared L3 cache. The Intel part offers 112 KB of L1 per core, 2 MB of L2 per core, and 144 MB of shared L3 cache. This gives the Intel chip more than double the L3 cache, which may contribute to its physics and floating-point advantages.

Both processors support DDR5 memory and ECC memory. The AMD chip integrates Radeon Graphics, while the Intel chip has no integrated graphics. Neither processor has an unlocked multiplier. The AMD EPYC 4564P belongs to the EPYC 4004 series with Raphael codename, while the Intel Xeon 6736P belongs to the Xeon 6 family with Granite Rapids-SP codename.

The Verdict

The benchmark data paints a clear picture for most workloads. The AMD EPYC 4564P is the stronger performer across the majority of tests, winning 13 of 17 head-to-head comparisons. Its 30.6% advantage across every Cinebench test, combined with a 112.1% lead in single-thread performance, makes it the obvious choice for applications sensitive to clock speed and per-core efficiency. The AMD part also delivers higher average benchmark scores (95183 versus 87864) while consuming less power (170W versus 205W) and having a lower launch MSRP.

The Intel Xeon 6736P, despite having more than double the cores and threads, only wins in specific domains: physics simulation, floating-point math, prime number finding, and random string sorting. Its 48.1% physics advantage and 5.8% floating-point lead suggest that computational science workloads could benefit from the Intel architecture. The Intel part also offers substantially more memory bandwidth (409.6 GB/s versus 83.2 GB/s) and more PCIe lanes (88 versus 28), which matters for memory-bound or I/O-heavy server configurations.

The data suggests that buyers prioritizing general server performance, single-threaded responsiveness, and overall benchmark leadership should choose the AMD EPYC 4564P. Those running specialized physics or floating-point workloads, or requiring extreme memory bandwidth and PCIe expansion, would find the Intel Xeon 6736P's specific strengths compelling. Both processors sit at the 96th percentile among all CPUs, indicating high overall capability, but the AMD part's benchmark victories and lower launch MSRP give it a broader appeal for most server deployments.