AMD EPYC 9754 vs Intel Xeon 6781P Comparison

AMD
AMD

AMD EPYC 9754

CORE STATE Bergamo
CORE SPECS 128 Cores / 256 Threads
CLOCK SPEED 2.25 Base / 3.1 GHz Turbo
CACHE 256 MB (shared)
MAX TDP 360W
ARCHITECTURE Zen 4
nm
PROCESS 5 nm
LAUNCH DATE 2023
VS
Intel
INTEL

Xeon 6781P

CORE STATE Granite Rapids
CORE SPECS 80 Cores / 160 Threads
CLOCK SPEED 2 Base / 3.8 GHz Turbo
CACHE 336 MB (shared)
MAX TDP 350W
ARCHITECTURE Granite Rapids
nm
PROCESS 5 nm
LAUNCH DATE 2025

PERFORMANCE BENCHMARKS

cinebench_cinebench_r15_multicore
8,460
10,105
cinebench_cinebench_r15_singlecore
1,194
N/A
cinebench_cinebench_r20_multicore
35,254
42,106
cinebench_cinebench_r20_singlecore
4,977
N/A
cinebench_cinebench_r23_multicore
83,939
100,254
cinebench_cinebench_r23_singlecore
11,850
N/A
passmark_data_compression
3,558,043
2,441,690
passmark_data_encryption
231,891
119,623
passmark_extended_instructions
224,322
199,048
passmark_find_prime_numbers
604
1,687
passmark_floating_point_math
588,187
507,406
passmark_integer_math
1,026,896
584,834
passmark_multithread
98,752
117,946
passmark_physics
8,793
17,753
passmark_random_string_sorting
306,481
268,573
passmark_single_thread
2,328
3,152
passmark_singlethread
2,328
3,152

Analysis: AMD EPYC 9754 vs Intel Xeon 6781P

The Verdict

The data splits these two server processors into distinct roles. The AMD EPYC 9754 wins six head-to-head benchmarks, while the Intel Xeon 6781P wins eight. The AMD part dominates in data encryption, integer math, and data compression, with deltas of 93.9%, 75.6%, and 45.7% respectively. The Intel part counters with massive leads in find prime numbers (64.2% ahead), physics (50.5% ahead), and single-thread performance (26.1% ahead). The average benchmark scores tell a similar story: the EPYC 9754 averages 364,371 across all tests, while the Xeon 6781P averages 315,524. The AMD chip sits at the 100th percentile of all CPUs, while the Intel part sits at the 99th. For workloads heavy on encryption, integer processing, or compression, the EPYC 9754 is the clear pick. For tasks dominated by prime-number searches, physics simulation, or lightly threaded code, the Xeon 6781P leads. The EPYC 9754's nearest rival is the Intel Xeon 6960P, which scores 365,194 on average, a mere 0.2% difference. The Xeon 6781P's closest competitor is the AMD Ryzen Threadripper PRO 9985WX at 320,749, which is 1.6% higher.

Architecture Differences

The AMD EPYC 9754 uses the Zen 4c architecture under the Bergamo codename, built on a 5 nm process by TSMC. It packs 128 cores and 256 threads. The Intel Xeon 6781P uses Granite Rapids, also on a 5 nm process but fabricated by Intel, with 80 cores and 160 threads. The die sizes differ dramatically: the EPYC 9754 uses eight chiplets of 73 mm² each, while the Xeon 6781P uses two dies of 598 mm² each. The AMD part has 71,000 million transistors, while the Intel part's transistor count is not listed in the data. Cache layouts diverge significantly. The EPYC 9754 has 64 KB of L1 per core, 1 MB of L2 per core, and 256 MB of shared L3. The Xeon 6781P has 112 KB of L1 per core, 2 MB of L2 per core, and 336 MB of shared L3. The Intel part has more L3 cache and more L2 per core, but the AMD part has more total cores. Memory channels differ: the EPYC 9754 supports twelve-channel DDR5 with 460.8 GB/s bandwidth, while the Xeon 6781P supports eight-channel DDR5 with 409.6 GB/s. Both support ECC memory. PCIe lanes are close: the EPYC 9754 has 128 Gen 5 lanes, while the Xeon 6781P has 136 Gen 5 lanes. The sockets are entirely different — SP5 for AMD, Socket 4710 for Intel. The EPYC 9754 was released on 2023-06-12, while the Xeon 6781P launched later on 2025-02-23. The EPYC 9754 carries a launch MSRP of $11900, and the Xeon 6781P has a launch MSRP of $8960.

Where Each One Wins

The AMD EPYC 9754 wins decisively in data-centric workloads. Its data encryption score of 231,891 is nearly double the Xeon 6781P's 119,623, a 93.9% advantage. Integer math shows a 75.6% lead, with 1,026,896 versus 584,834. Data compression favors AMD by 45.7%, scoring 3,558,043 versus 2,441,690. Floating-point math also leans AMD, with 588,187 versus 507,406, a 15.9% gap. Random string sorting goes to AMD at 306,481 versus 268,573, a 14.1% margin. Extended instructions favor AMD by 12.7%, at 224,322 versus 199,048. These results suggest the EPYC 9754 excels in encryption, compression, and general arithmetic workloads.

The Intel Xeon 6781P wins in multi-core rendering and physics. In Cinebench R15, R20, and R23 multi-core tests, the Xeon 6781P scores 10,105, 42,106, and 100,254 respectively, each 16.3% ahead of the EPYC 9754's 8,460, 35,254, and 83,939. The PassMark multithread score also favors Intel at 117,946 versus 98,752, a 16.3% lead. Find prime numbers is a massive win for Intel: 1,687 versus 604, a 64.2% advantage. Physics shows a 50.5% lead for Intel, with 17,753 versus 8,793. Single-thread performance clearly favors Intel, with 3,152 versus 2,328, a 26.1% gap. These numbers indicate the Xeon 6781P is stronger in rendering, physics simulation, and single-threaded tasks.

FAQ

Q: Which processor has more cores and threads?

A: The AMD EPYC 9754 has 128 cores and 256 threads. The Intel Xeon 6781P has 80 cores and 160 threads.

Q: How do the two compare in single-threaded performance?

A: The Intel Xeon 6781P leads in PassMark single-thread with a score of 3,152 versus the EPYC 9754's 2,328, a 26.1% advantage. The Intel part's boost clock is higher at 3.80 GHz versus 3.10 GHz.

Q: Which processor wins in encryption workloads?

A: The AMD EPYC 9754 wins PassMark data encryption with 231,891 versus 119,623, a 93.9% margin. This is the largest benchmark delta between the two processors.

Q: What are the memory bandwidth differences?

A: The AMD EPYC 9754 has twelve-channel DDR5 memory with 460.8 GB/s bandwidth. The Intel Xeon 6781P has eight-channel DDR5 with 409.6 GB/s bandwidth.

Q: Which processor has more L3 cache?

A: The Intel Xeon 6781P has 336 MB of shared L3 cache. The AMD EPYC 9754 has 256 MB of shared L3 cache.

Q: How do the average benchmark scores compare?

A: The AMD EPYC 9754 averages 364,371 across all benchmarks, while the Intel Xeon 6781P averages 315,524. The AMD part ranks at the 100th percentile of all CPUs, while the Intel part ranks at the 99th.

Head-to-Head Benchmarks

The largest win for the AMD EPYC 9754 comes in PassMark data encryption. The AMD part scores 231,891 against the Intel part's 119,623, a 93.9% advantage. This is followed by integer math, where AMD scores 1,026,896 versus 584,834, a 75.6% lead. Data compression shows AMD at 3,558,043 versus 2,441,690, a 45.7% margin. Floating-point math gives AMD a 15.9% win at 588,187 versus 507,406. Random string sorting goes to AMD at 306,481 versus 268,573, a 14.1% edge. Extended instructions favor AMD by 12.7%, with 224,322 versus 199,048.

The largest win for the Intel Xeon 6781P is in PassMark find prime numbers. Intel scores 1,687 versus AMD's 604, a 64.2% advantage. Physics shows a 50.5% lead for Intel, with 17,753 versus 8,793. Single-thread performance favors Intel by 26.1%, at 3,152 versus 2,328. In Cinebench multi-core tests, Intel wins consistently by 16.3%: R15 scores 10,105 versus 8,460, R20 scores 42,106 versus 35,254, and R23 scores 100,254 versus 83,939. The PassMark multithread score also favors Intel by 16.3%, at 117,946 versus 98,752.

The Cinebench results are particularly telling because they show a consistent 16.3% lead across all three versions of the test. This suggests the Intel architecture has a structural advantage in rendering workloads. The PassMark multithread test confirms the same 16.3% gap. However, the AMD part's wins in encryption and integer math are much larger in percentage terms, which balances the overall benchmark count at 6 wins for AMD and 8 for Intel. The data shows that the EPYC 9754's total average score of 364,371 is higher than the Xeon 6781P's 315,524, despite losing more individual benchmarks. This is because AMD's wins in encryption and integer math are so large that they pull the average up significantly.

Specification Differences

The two processors differ in nearly every core specification. The AMD EPYC 9754 has 128 cores and 256 threads, while the Intel Xeon 6781P has 80 cores and 160 threads. Base clocks are 2.25 GHz for AMD and 2.00 GHz for Intel. Boost clocks reverse the order: 3.10 GHz for AMD and 3.80 GHz for Intel. TDP is close but not identical: 360 W for AMD and 350 W for Intel. The sockets are incompatible: AMD Socket SP5 versus Intel Socket 4710. The process node is the same at 5 nm, but the foundries differ — TSMC for AMD, Intel for Intel. The EPYC 9754 uses eight 73 mm² dies, while the Xeon 6781P uses two 598 mm² dies. The transistor count is 71,000 million for AMD; no transistor count is listed for Intel. Cache per core differs: AMD has 64 KB L1 and 1 MB L2, while Intel has 112 KB L1 and 2 MB L2. Shared L3 is 256 MB for AMD and 336 MB for Intel. Memory channels are twelve for AMD and eight for Intel, with bandwidth at 460.8 GB/s versus 409.6 GB/s. PCIe lanes are 128 for AMD and 136 for Intel, both Gen 5. Both support ECC memory. The release dates differ by over a year: 2023-06-12 for AMD and 2025-02-23 for Intel. The launch MSRP is $11900 for the EPYC 9754 and $8960 for the Xeon 6781P. The Intel part has integrated graphics listed as "N/A," while the AMD part has no integrated graphics listed at all.

DETAILED SPECIFICATIONS

SPECIFICATION
EPYC 9754
6781P
Core Specs
Cores
128
80 -37.5%
Threads
256
160 -37.5%
Base Clock (GHz)
2.25
2 -11.1%
Boost Clock (GHz)
3.1
3.8 +22.6%
Frequency (GHz)
2.25
2 -11.1%
Turbo Clock (GHz)
3.1
3.8 +22.6%
Multiplier
22.5
20 -11.1%
SMP CPUs
2
1 -50.0%
Cache
L1 Cache
64 KB (per core)
112 KB (per core)
L2 Cache
1 MB (per core)
2 MB (per core)
L3 Cache
256 MB (shared)
336 MB (shared)
Power
TDP (W)
360
350 -2.8%
Configurable TDP
320-400 W
—
Architecture
Architecture
Zen 4
Granite Rapids
Codename
Bergamo
Granite Rapids
Generation
EPYC (Zen 4c (Bergamo))
Xeon 6 (Granite Rapids-SP)
Process Size
5 nm
5 nm
Transistors
71,000 million
—
Die Size
8x 73 mm²
2x 598 mm²
Foundry
TSMC
Intel
Memory
Memory Support
DDR5
DDR5
Memory Bus
Twelve-channel
Eight-channel
Memory Bandwidth
460.8 GB/s
409.6 GB/s
ECC Memory
Yes
Yes
Platform
Socket
AMD Socket SP5
Intel Socket 4710
PCIe
Gen 5, 128 Lanes(CPU only)
Gen 5, 136 Lanes(CPU only)
AMD Multi-Die
IO Process Size
6 nm
10 nm
Interconnect
CXL
—
Gen 2.0, 64 Lanes (Shared with PCI-E)
Other
Market
Server/Workstation
Server/Workstation
Production Status
Active
Active
Launch Price
$11900
$8960
Part Number
100-000001234
SRV5J
Package
FC-LGA6096
FC-LGA18N
Tj Max
—
97°C
Bundled Cooler
None
None
View EPYC 9754 Details View Xeon 6781P Details