AMD EPYC 8534P vs Intel Xeon 6980P Comparison

The Verdict

The Intel Xeon 6980P and AMD EPYC 8534P occupy very different positions in the server CPU landscape, and the benchmark data reflects that. The Intel Xeon 6980P wins 14 of the 17 head-to-head benchmark comparisons, while the AMD EPYC 8534P takes 3. The Intel part is the clear overall performance leader, with its average benchmark score of 251,516 placing it in the 99th percentile of all CPUs. The AMD EPYC 8534P, with an average score of 185,092, sits in the 98th percentile, a respectable but clearly lower tier.

The verdict depends entirely on workload priorities. For compute-heavy, multi-threaded applications, data compression, encryption, and floating-point work, the Intel Xeon 6980P is the stronger choice, often by substantial margins. The data shows the Xeon is 32% ahead in data compression, 73.3% ahead in floating-point math, and 90.3% ahead in extended instructions. However, the AMD EPYC 8534P wins decisively in single-threaded performance, with a 31.1% lead in PassMark single-thread scoring, and also edges out Intel in the physics benchmark by 8.6%. Organizations prioritizing single-thread responsiveness or physics simulation workloads should examine the AMD part closely, despite its lower core count.

The Intel Xeon 6980P's closest rivals in the database include the AMD EPYC 9684X, which outperforms it by 5.8%, and the AMD Ryzen Threadripper 9970X, which leads by 10.1%. Meanwhile, the AMD EPYC 8534P's nearest rival is the AMD Ryzen Threadripper PRO 9975WX, which scores only 1.3% higher, and the Intel Xeon 6740E, which trails by 1.4%. These comparisons suggest the Xeon 6980P faces stiffer competition at the top end, while the EPYC 8534P sits in a more contested mid-range tier.

Architecture Differences

The architectural gap between these two processors is substantial. The Intel Xeon 6980P uses the Granite Rapids architecture, specifically the Xeon 6 generation built on a 5 nm process at Intel's own foundry. It packs 128 cores and 256 threads, with a die composed of three 598 mm² slices. Its cache hierarchy includes 112 KB of L1 per core, 2 MB of L2 per core, and a massive 504 MB of shared L3 cache.

The AMD EPYC 8534P belongs to the EPYC 8004 series, built on the Zen 4c architecture with the codename Siena. It uses a 5 nm process from TSMC and contains 64 cores and 128 threads, exactly half the Intel part's core count. The AMD chip is built from four 73 mm² dies, with a total of 35,500 million transistors. Its cache configuration is more modest: 64 KB of L1 per core, 1 MB of L2 per core, and 128 MB of shared L3.

The process node is the same on paper, 5 nm, but the foundry differs: Intel fabricates its own chip, while AMD relies on TSMC. The Intel die is far larger in aggregate, at roughly 1,794 mm² across three slices, compared to AMD's four smaller dies totaling 292 mm². The transistor count for the Intel part is not recorded in the database, but the physical size difference suggests a fundamentally different design approach. The Xeon 6980P also carries a 500 W TDP, while the EPYC 8534P is rated at 200 W, a gap that reflects the Intel part's much larger core count and cache pool.

Memory architecture diverges sharply as well. The Intel Xeon 6980P supports twelve-channel DDR5 with 614.4 GB/s of memory bandwidth, while the AMD EPYC 8534P uses six-channel DDR5 with 230.4 GB/s. Both support ECC memory and provide Gen 5 PCIe with 96 lanes from the CPU. Neither has integrated graphics. The Intel part launched on September 23, 2024, roughly a year after the AMD EPYC 8534P, which arrived on September 17, 2023.

FAQ

Q: Which processor has more cores and threads?

A: The Intel Xeon 6980P has 128 cores and 256 threads. The AMD EPYC 8534P has 64 cores and 128 threads, exactly half of Intel's counts.

Q: How do their memory bandwidth figures compare?

A: The Intel Xeon 6980P supports twelve-channel DDR5 memory with 614.4 GB/s of bandwidth. The AMD EPYC 8534P uses six-channel DDR5 with 230.4 GB/s, meaning Intel provides roughly 2.7 times the memory bandwidth in the recorded specifications.

Q: Which processor wins in single-threaded performance?

A: The AMD EPYC 8534P wins decisively in single-threaded tests. In PassMark single-thread and singlethread benchmarks, it scores 2441 compared to Intel's 1681, a 31.1% advantage.

Q: What is the L3 cache difference between the two?

A: The Intel Xeon 6980P has 504 MB of shared L3 cache, while the AMD EPYC 8534P has 128 MB of shared L3 cache. Intel's L3 is nearly four times larger.

Q: How does the Intel Xeon 6980P perform in data compression?

A: The Intel Xeon 6980P scores 2,364,519 in PassMark data compression, which is 32% higher than the AMD EPYC 8534P's score of 1,791,742.

Q: Which processor has the higher boost clock?

A: The Intel Xeon 6980P has a boost clock of 3.90 GHz, while the AMD EPYC 8534P boosts to 3.10 GHz. Intel's base clock is 2.00 GHz compared to AMD's 2.30 GHz, so AMD has the higher base clock but Intel reaches a higher maximum.

Specification Differences

The recorded specifications show distinct trade-offs between the two processors. The Intel Xeon 6980P offers 128 cores versus the AMD EPYC 8534P's 64 cores, and 256 threads versus 128 threads. Clock speeds differ in both directions: Intel has a 2.00 GHz base clock and 3.90 GHz boost, while AMD starts at 2.30 GHz base and reaches only 3.10 GHz boost. The TDP gap is significant, with Intel rated at 500 W and AMD at 200 W.

Sockets are incompatible: Intel uses Socket 7529, while AMD uses Socket SP6. Cache sizes diverge across all levels. Intel provides 112 KB L1 per core, 2 MB L2 per core, and 504 MB shared L3. AMD provides 64 KB L1 per core, 1 MB L2 per core, and 128 MB shared L3. The memory bus width differs at twelve channels versus six, producing the bandwidth gap noted earlier. Both support DDR5, ECC memory, and Gen 5 PCIe with 96 lanes.

The physical packaging also differs. Intel uses three dies of 598 mm² each, while AMD uses four dies of 73 mm² each. The AMD part lists 35,500 million transistors; no transistor count is recorded for Intel. Manufacturing responsibilities differ as well, with Intel's foundry producing the Xeon and TSMC producing the EPYC. The launch MSRP for the Intel Xeon 6980P is $12460, and for the AMD EPYC 8534P it is $4950. Neither processor has an unlocked multiplier.

Head-to-Head Benchmarks

The head-to-head data reveals a pattern of Intel dominance in most workloads, with AMD winning in specific niches. Across all Cinebench tests, the Intel Xeon 6980P wins every round with a consistent 3.4% margin. In Cinebench R15 multicore, Intel scores 6367 against AMD's 6160. In R15 singlecore, Intel leads 898 to 869, a 3.3% edge. The R20 multicore test shows Intel at 26533 versus AMD's 25668, and R20 singlecore has Intel at 3745 versus 3623. Cinebench R23 multicore gives Intel 63175 against AMD's 61115, while R23 singlecore shows Intel at 8918 versus 8628. The consistency of the 3.4% delta in multicore tests across all Cinebench versions suggests a stable architectural advantage rather than workload-specific variation.

The PassMark suite tells a more dramatic story. In data compression, Intel scores 2,364,519 against AMD's 1,791,742, a 32% lead. Data encryption is closer, with Intel at 125,246 versus AMD's 121,728, a 2.9% margin. Extended instructions show the largest gap: Intel scores 214,794 versus AMD's 112,860, a 90.3% advantage. Find prime numbers gives Intel 555 versus AMD's 278, a 99.6% lead, meaning Intel completes the task in roughly half the time. Floating-point math favors Intel heavily, 501,720 versus 289,443, a 73.3% gap. Integer math gives Intel 637,476 versus AMD's 514,526, a 23.9% lead. Multithread performance is close, with Intel at 74,324 versus AMD's 71,900, a 3.4% margin. Random string sorting favors Intel at 240,792 versus 129,479, an 86% lead.

The AMD EPYC 8534P secures its wins in physics and single-threaded tests. In PassMark physics, AMD scores 3,667 against Intel's 3,350, an 8.6% advantage. In PassMark single-thread and singlethread, AMD scores 2,441 against Intel's 1,681, a 31.1% lead. These are the only three benchmark wins for AMD out of the 17 head-to-head comparisons.

Where Each One Wins

The Intel Xeon 6980P is the choice for workloads that scale with core count and cache capacity. Its 128 cores, 256 threads, and 504 MB of L3 cache drive massive advantages in data compression, where it leads by 32%, and in floating-point math, where the 73.3% gap indicates strong vector processing capability. Extended instructions show a 90.3% lead, making the Intel part well-suited for cryptographic or SIMD-heavy workloads. The find prime numbers test, with a 99.6% advantage, points to raw integer throughput that dwarfs the AMD part. Random string sorting, which typically stresses memory access patterns and cache behavior, shows an 86% lead, likely benefiting from the twelve-channel memory bus and 614.4 GB/s bandwidth. The Xeon 6980P also holds a narrow but consistent edge in all Cinebench multicore tests at 3.4%, and in encryption at 2.9%.

The AMD EPYC 8534P wins where single-thread performance and specific physics workloads matter. Its 31.1% lead in PassMark single-thread scoring suggests higher per-core efficiency, despite lower boost clocks on paper. The physics benchmark, where AMD leads by 8.6%, may reflect a particular instruction mix or memory access pattern that favors the Zen 4c architecture. The EPYC 8534P also offers a much lower TDP at 200 W versus 500 W, which the data records but does not directly benchmark. For deployments where per-core licensing, lower power envelopes, or single-thread latency are the priority, the AMD part has clear appeal.

The broader competitive context reinforces this split. The Intel Xeon 6980P's nearest rivals include the AMD EPYC 9684X, which beats it by 5.8%, and the AMD Ryzen Threadripper 9970X, which leads by 10.1%. These are higher-performing alternatives, suggesting the Xeon 6980P is positioned just below the absolute top tier. The AMD EPYC 8534P, meanwhile, trades blows with the AMD Ryzen Threadripper PRO 9975WX, which is only 1.3% ahead, and the Intel Xeon 6740E, which trails by 1.4%. The AMD part sits in a tightly contested segment where small margins separate competitors, making its single-thread advantage a differentiator. The data does not support a universal recommendation; it supports a workload-specific one.