AMD EPYC 9275F vs Intel Xeon 6745P Comparison

Head-to-Head Benchmarks

The head-to-head data reveals a clear split: the AMD EPYC 9275F wins 11 of the 16 recorded comparisons, while the Intel Xeon 6745P takes 5. But the margins tell a more interesting story than the win count. The AMD part's victories are often narrow, while the Intel part's wins are frequently decisive.

In the Cinebench suite, the EPYC 9275F edges out the Xeon across all five tests, but by the slimmest of margins. Each of the three multicore tests (R15, R20, R23) and both single-core tests shows a delta of only -0.5% for the Intel part. The scores are nearly identical: 7250 versus 7214 in R15 multicore, 30209 versus 30062 in R20 multicore, and 71927 versus 71578 in R23 multicore. Single-core R15 lands at 1023 versus 1018, and R20 single-core at 4264 versus 4244. These are statistical ties in practical terms, yet the AMD chip claims the win in every one.

PassMark's multithread test follows the same pattern: the EPYC 9275F scores 84620 against the Xeon's 84210, another -0.5% delta. The picture shifts dramatically when looking at specialized workloads. The Xeon 6745P dominates floating point math with 267438 against 201888, a 32.5% advantage. It also leads extended instructions by 14.2% (108326 versus 94889), data compression by 11.6% (1352801 versus 1212560), encryption by 6.4% (66665 versus 62664), and integer math by 6% (336926 versus 317777).

The AMD EPYC 9275F counters with its own decisive wins. The physics test shows a 49.2% gap in its favor, scoring 12089 against the Intel part's 6144. Prime number finding goes to AMD by 31.3% (991 versus 681). Single-thread PassMark scores favor the EPYC by 9.4% (3810 versus 3450), and random string sorting goes to AMD by 7.3% (144037 versus 133528).

The data suggests two very different performance profiles. The Intel Xeon 6745P is built for throughput in math-heavy, compression-heavy, and encryption-heavy tasks. The AMD EPYC 9275F excels in latency-sensitive, single-thread-bound, and physics-style workloads, while also delivering marginally better all-core rendering scores.

Architecture Differences

The two processors come from fundamentally different design philosophies. The Intel Xeon 6745P uses Granite Rapids architecture, built on Intel's 5 nm process, while the AMD EPYC 9275F uses Zen 5 architecture, produced by TSMC on a 4 nm node. The Intel die is substantial, listed as 2x 598 mm², whereas the AMD chip uses 8x 70.6 mm² chiplets, with the combined transistor count for the AMD part listed at 66,520 million.

Core counts diverge significantly. The Xeon packs 32 cores and 64 threads, while the EPYC offers 24 cores and 48 threads. That is 8 fewer cores and 16 fewer threads for the AMD part, yet it still manages to win most benchmarks. Cache configurations differ as well. The Intel part carries 112 KB of L1 per core, 2 MB of L2 per core, and a massive 336 MB of shared L3. The AMD part has 80 KB of L1 per core, 1 MB of L2 per core, and 256 MB of shared L3.

Clock speeds favor AMD. The EPYC 9275F runs at a 4.10 GHz base and 4.80 GHz boost, compared to the Xeon's 3.10 GHz base and 4.30 GHz boost. That 1.0 GHz base clock gap and 0.5 GHz boost gap help explain the AMD chip's single-thread advantage, despite the Intel part's larger core count.

Memory subsystems also differ. Both support DDR5 and ECC, but the EPYC uses a twelve-channel memory bus with 576.0 GB/s bandwidth, while the Xeon uses an eight-channel bus with 409.6 GB/s bandwidth. PCIe lanes favor AMD as well: 128 Gen 5 lanes versus 88 Gen 5 lanes for Intel. The sockets are incompatible, with Intel using Socket 4710 and AMD using Socket SP5.

The release dates are close: the EPYC 9275F launched on 2024-10-09, and the Xeon 6745P followed on 2025-02-23. Both are active production parts, both lack integrated graphics, both are locked, and both target the server and workstation segment.

Where Each One Wins

The benchmark data points to distinct use cases. The Intel Xeon 6745P is the clear choice for floating point math, where its 32.5% lead over the EPYC 9275F is the largest margin in the entire comparison. It also wins data compression by 11.6%, extended instructions by 14.2%, encryption by 6.4%, and integer math by 6%. Workloads that hammer the FPU, compress large datasets, or perform heavy encryption will favor the Intel part.

The AMD EPYC 9275F claims the physics benchmark by a massive 49.2%, suggesting a strong advantage in simulation and physics-based workloads. Prime number finding favors AMD by 31.3%, which points to integer-heavy algorithmic work with low thread counts. The 9.4% single-thread PassMark advantage means the EPYC is better suited to latency-sensitive applications, database queries, and lightly threaded tasks.

Random string sorting goes to AMD by 7.3%, and the Cinebench suite, though narrow, consistently favors the EPYC across both single and multicore tests. The multithread PassMark score also edges toward AMD by 0.5%. For rendering workloads, the AMD chip is slightly ahead despite having fewer cores, which suggests better per-core efficiency.

The average benchmark scores in the database tell a similar story. The Xeon 6745P has an average benchmark score of 154858, placing it in the 98th percentile of all CPUs. The EPYC 9275F averages 133174, putting it in the 97th percentile. The Intel part's nearest rivals include the AMD EPYC 9375F at -4.7%, the AMD EPYC 7663 at -4.4%, the AMD EPYC 9355P at -3.4%, and the Intel Xeon 676X at -2.3%. The AMD part's nearest rivals are the Intel Xeon 6737P at -5.3%, the Intel Xeon 6730P at 6.7%, the AMD EPYC 9354 at 5%, and the Intel Xeon 6710E at 2.5%.

Specification Differences

The two processors differ across nearly every major specification. The Xeon 6745P has 32 cores and 64 threads, while the EPYC 9275F has 24 cores and 48 threads. Base clocks are 3.10 GHz versus 4.10 GHz, and boost clocks are 4.30 GHz versus 4.80 GHz, both favoring AMD. TDP ratings are 300 W for Intel and 320 W for AMD.

The process node differs: 5 nm for Intel, 4 nm for AMD. The foundries are Intel and TSMC respectively. Die size is 2x 598 mm² for Intel and 8x 70.6 mm² for AMD. The AMD part reports 66,520 million transistors, while the Intel part has no transistor count listed.

Cache hierarchies diverge: L1 is 112 KB per core for Intel versus 80 KB per core for AMD, L2 is 2 MB per core versus 1 MB per core, and L3 is 336 MB shared versus 256 MB shared. Memory bus width differs at eight channels for Intel and twelve channels for AMD. Memory bandwidth is 409.6 GB/s versus 576.0 GB/s. PCIe lanes are 88 for Intel and 128 for AMD, both Gen 5.

Socket types are Intel Socket 4710 and AMD Socket SP5. The architectures are Granite Rapids and Zen 5. The codenames are Granite Rapids and Turin. The generations are Xeon 6 (Granite Rapids-SP) and EPYC (Zen 5 (Turin)). The AMD part belongs to the EPYC 9005 series. Part numbers are SRWPAQ7L9 for Intel and 100-000001144 for AMD. The launch MSRP is $5250 for the Xeon and $3439 for the EPYC. Both have no integrated graphics, both are locked, and both are active production parts.

FAQ

Q: Which processor has more cores and threads?

A: The Intel Xeon 6745P has 32 cores and 64 threads, while the AMD EPYC 9275F has 24 cores and 48 threads.

Q: Which processor is faster in single-threaded workloads?

A: The AMD EPYC 9275F wins the PassMark single-thread test with 3810 against the Intel Xeon 6745P's 3450, a 9.4% advantage. It also wins both Cinebench R15 and R20 single-core tests by 0.5% margins.

Q: How do the two compare in floating point math performance?

A: The Intel Xeon 6745P leads by 32.5% in PassMark floating point math, scoring 267438 versus the AMD EPYC 9275F's 201888.

Q: What is the memory bandwidth difference?

A: The AMD EPYC 9275F provides 576.0 GB/s over a twelve-channel DDR5 memory bus, while the Intel Xeon 6745P provides 409.6 GB/s over an eight-channel DDR5 bus.

Q: Which processor has a higher boost clock?

A: The AMD EPYC 9275F boosts to 4.80 GHz, compared to the Intel Xeon 6745P's 4.30 GHz. The AMD part also has a higher base clock at 4.10 GHz versus 3.10 GHz.

Q: How do the average benchmark scores compare?

A: The Intel Xeon 6745P has an average benchmark score of 154858 in the 98th percentile, while the AMD EPYC 9275F averages 133174 in the 97th percentile.

The Verdict

The data supports a clear split decision. The AMD EPYC 9275F is the better all-rounder for general server workloads, winning 11 of 16 head-to-head benchmarks. Its physics score leads by 49.2%, its prime number finding by 31.3%, and its single-thread PassMark by 9.4%. It also delivers superior memory bandwidth at 576.0 GB/s and 128 PCIe lanes, both of which matter for I/O-heavy deployments. The Cinebench results, while narrow, favor AMD in every rendering test, making it the safer choice for mixed workloads that include both rendering and general compute.

The Intel Xeon 6745P is the specialist. Its 32.5% lead in floating point math, 14.2% lead in extended instructions, and 11.6% lead in data compression make it the pick for compute-heavy scientific workloads, compression pipelines, and encryption tasks. The larger L3 cache of 336 MB and the 32-core configuration provide a throughput advantage that shows up in the average benchmark score of 154858, which sits above the EPYC's 133174.

The choice depends on workload profile. If the work is dominated by FPU math, compression, or encryption, the Intel Xeon 6745P is the data-backed winner. If the work involves physics simulation, single-threaded performance, or memory-bandwidth-hungry applications, the AMD EPYC 9275F is the stronger option. The EPYC's lower launch MSRP of $3439, compared to $5250 for the Xeon, is a secondary consideration that the performance metrics alone do not fully justify in either direction, but the benchmark wins speak for themselves.