AMD Ryzen 9 PRO 9955

The AMD Ryzen 9 PRO 9955 is a 12-core, 24-thread Granite Ridge processor built on TSMC’s 4 nm process, with a base clock of 3.40 GHz and a boost clock of 5.40 GHz. It targets the server/workstation segment, yet fits the mainstream AMD Socket AM5 platform, bridging enterprise-class compute demands with consumer motherboard compatibility. The chip carries a 120 W TDP, supports DDR5 memory in a dual-channel configuration, and includes Radeon Graphics as integrated silicon, making it a flexible option for both professional workstations and high-end desktop builds.

Platform and Compatibility

The Ryzen 9 PRO 9955 uses AMD Socket AM5, which means it drops into the same motherboards designed for AMD’s 9000 series desktop processors. This socket provides a clear upgrade path from earlier Ryzen generations, though the exact chipset feature set is not specified in the data. The processor supports DDR5 memory only, with a dual-channel memory bus delivering a peak bandwidth of 89.6 GB/s. For memory-intensive workloads such as large database operations or scientific simulations, that bandwidth is a critical enabler, and the chip’s support for ECC memory adds a reliability layer typically reserved for server platforms.

PCIe connectivity is handled through Gen 5 with 24 lanes available from the CPU alone. This is a high-bandwidth interface for modern GPUs, NVMe storage, and accelerators, though the data does not specify how those lanes are partitioned. The integrated Radeon Graphics means the CPU can output video without a discrete GPU, which is useful for headless servers or troubleshooting workstations. The processor is not multiplier-unlocked, indicating that overclocking is not a primary intended use case; instead, stability and sustained performance are the focus.

The market segment is listed as Server/Workstation, yet the AM5 socket typically serves consumer and prosumer boards. This dual nature suggests the chip is positioned for compact workstations or single-socket servers that leverage consumer platform costs while retaining PRO-grade features like ECC support and long-term availability. The production status is Active, and the release date is June 29, 2026, though no launch MSRP is provided in the data. The part number is 100-000001971, and the transistor count is 16,630 million spread across a die size of 2x 70.6 mm².

Power and Thermals

The Ryzen 9 PRO 9955 carries a TDP of 120 W. This is a moderately high power envelope, but it is not extreme for a 12-core processor. For cooling, this TDP class implies the need for a capable air cooler or a modest liquid cooler, but not a massive radiator setup. The data does not specify a separate maximum boost power or thermal throttling behavior, so the analysis must rely on the 120 W figure alone. In a workstation chassis with good airflow, a high-end tower cooler should suffice; in a compact server case, a low-profile cooler with sufficient static pressure would be required.

The 4 nm process from TSMC contributes to efficiency, as smaller nodes typically reduce power draw for a given performance level. With a boost clock of 5.40 GHz, the chip can reach high frequencies under light loads, but sustained all-core workloads will likely push the TDP limit. The lack of a multiplier unlock means users cannot manually adjust voltage and frequency, so thermal management falls to the motherboard’s default power delivery settings. For long-running compute tasks, the 120 W envelope suggests consistent performance without the need for exotic cooling, but users should ensure their case has adequate exhaust to handle the heat output.

Single-Thread vs Multi-Thread Behavior

The benchmark data reveals a stark split between single-thread and multi-thread performance. The PassMark single-thread score is 4597, which is a strong result, indicating excellent responsiveness in lightly threaded tasks. In contrast, the multi-thread score is 54866, which is roughly 11.9 times higher than the single-thread score. This ratio is lower than what a 12-core/24-thread chip might theoretically achieve if scaling were perfect, but it reflects real-world inefficiencies in thread scheduling and memory contention.

For workloads that rely on single-thread performance, such as legacy applications, spreadsheet macros, or certain simulation tools that are not parallelized, the 5.40 GHz boost clock and the efficient Zen 5 architecture will deliver snappy performance. The single-thread score places the chip in the top tier of processors, as evidenced by its 97th percentile ranking across all CPUs. On the multi-thread side, the 54866 score shows substantial parallel capability, suitable for video rendering, compiler jobs, or financial modeling. However, the data also includes other PassMark sub-tests that reveal specific strengths: integer math scores 182312, floating-point math scores 121509, and extended instructions score 54903. These are all compute-heavy metrics that benefit from the 12 full cores and 24 threads.

Notably, the find prime numbers score is only 461, which is a workload that often stresses memory latency and integer division; this low score relative to other metrics suggests that the chip is not optimized for that specific algorithm, but it does not detract from overall capability. The data compression score of 684470 is exceptionally high, indicating strong performance in archiving or data throughput tasks. Random string sorting scores 71928, which reflects memory and cache efficiency. The physics score is 3332, which is modest but acceptable for a workstation chip. The encryption score of 33754 shows adequate cryptographic throughput, likely aided by AES-NI instructions.

How It Compares

The nearest rival is the Intel Xeon w7-2595X, which has an average benchmark score of 108663. The Ryzen 9 PRO 9955 scores 110612, a delta of 1.8% higher. This is a narrow margin, meaning the two processors are effectively interchangeable in average performance. The AMD chip offers a lower TDP (120 W vs the Xeon’s unspecified but likely higher), and it uses a mainstream socket, which could reduce platform costs. However, the Xeon may offer more PCIe lanes or memory channels, but those details are not in the data.

The AMD Ryzen 9 9850HX is another rival, with an average score of 106413. The PRO 9955 leads by 3.9%. The 9850HX is a mobile processor, so it has lower power limits and cooling constraints. The PRO 9955’s advantage in average score reflects its desktop-class power envelope and higher boost clock. For users comparing a laptop workstation versus a desktop, the PRO 9955 will sustain higher performance under load, but the 9850HX offers portability.

The Intel Xeon w7-3555 scores 106192, placing it 4.2% behind the Ryzen 9 PRO 9955. This is a more pronounced gap, suggesting the AMD chip has a clear edge in average throughput. The Xeon w7-3555 is a server-class part, so it may have better memory bandwidth or reliability features, but the benchmark average favors the AMD part.

The Intel Xeon 6527P is the only rival that beats the Ryzen 9 PRO 9955, with an average score of 115190. The delta is -4%, meaning the Xeon 6527P is 4% faster. This is a significant lead, likely due to the Xeon’s higher core count or memory bandwidth, but those specs are not in the data. For workloads that scale heavily with cores, the Xeon 6527P would be the better choice, but it likely comes with a higher power draw and platform cost.

Who Should Consider It

The Ryzen 9 PRO 9955 is suited for professionals who run a mix of single-threaded and multi-threaded applications. The single-thread score of 4597 ensures that day-to-day interactions, such as compiling code, running database queries, or using CAD tools with single-threaded solvers, remain responsive. The multi-thread score of 54866, combined with the high data compression score of 684470, makes it a strong candidate for data engineering tasks, file archiving, and video encoding.

For gamers, the chip is overkill in core count, but the high single-thread score and boost clock of 5.40 GHz will drive high frame rates in most titles. The integrated Radeon Graphics means a discrete GPU is not required for basic display output, but for serious gaming, a dedicated GPU is mandatory. The 97th percentile ranking across all CPUs indicates that this processor outperforms the vast majority of installed hardware, so it will not bottleneck modern GPUs.

Content creators who use software like Blender, Premiere Pro, or After Effects will benefit from the multi-threaded performance, though the 4% gap to the Xeon 6527P suggests that for heavy 3D rendering, a larger core-count Xeon might be faster. However, the PRO 9955’s AM5 platform offers lower platform costs and better single-thread performance, which is advantageous for interactive editing. Office workers and general productivity users will find the chip overpowered, but its ECC memory support and long-term stability make it a safe choice for financial analysis or data-heavy spreadsheets.

The PRO branding implies enterprise features like manageability and security, though specifics are not in the data. For users who need a single-socket workstation with ECC memory and a mainstream upgrade path, this chip is a balanced option. It is not for overclockers, as the multiplier is locked, but it is for those who value reliability and sustained performance.

FAQ

Q: What socket does the AMD Ryzen 9 PRO 9955 use?

A: It uses AMD Socket AM5, which is compatible with motherboards designed for the 9000 series processors.

Q: Does the Ryzen 9 PRO 9955 support ECC memory?

A: Yes, ECC memory is supported, which is beneficial for server and workstation workloads where data integrity is critical.

Q: What is the TDP of this processor?

A: The TDP is 120 W, which implies a capable air cooler or a modest liquid cooler is sufficient.

Q: How does it perform in single-threaded tasks compared to multi-threaded tasks?

A: The single-thread score is 4597, while the multi-thread score is 54866. This indicates strong single-thread performance for responsive tasks and robust multi-thread performance for parallel workloads.

Q: What is the memory bandwidth of the Ryzen 9 PRO 9955?

A: The memory bandwidth is 89.6 GB/s, achieved through dual-channel DDR5 memory support.

Q: Does the chip have integrated graphics?

A: Yes, it includes Radeon Graphics, allowing basic display output without a discrete GPU.

Q: Is the processor multiplier unlocked?

A: No, the multiplier is locked, so overclocking is not supported.

Q: What is the process node and foundry?

A: It is built on TSMC’s 4 nm process, with a transistor count of 16,630 million.

Benchmark Performance

The average benchmark score for the Ryzen 9 PRO 9955 is 110612, which places it in the 97th percentile of all CPUs. This is a high ranking, indicating that the chip outperforms nearly all processors in the database. The score is derived from a suite of PassMark tests that measure various aspects of performance, from integer math to data compression.

Against its nearest rivals, the performance picture is nuanced. The Intel Xeon w7-2595X trails by just 1.8%, with an average score of 108663. This is a negligible difference in real-world terms; most applications will see no meaningful distinction. The AMD Ryzen 9 9850HX is 3.9% slower, scoring 106413, which is a more noticeable gap, especially for sustained workloads where the desktop PRO 9955 can maintain higher clocks. The Intel Xeon w7-3555 is 4.2% behind with a score of 106192, reinforcing the PRO 9955’s position as a solid mid-range workstation performer. Conversely, the Intel Xeon 6527P leads by 4%, scoring 115190. This is the only rival that outperforms the PRO 9955 by a margin that matters for heavy compute tasks.

Breaking down the sub-scores, the data compression test shows a score of 684470, which is exceptionally high and suggests that the chip excels at tasks involving large datasets and archiving. The integer math score of 182312 and floating-point math score of 121509 are both strong, indicating good arithmetic throughput for scientific and engineering applications. The extended instructions score of 54903 reflects the efficiency of AVX-512 or similar instruction sets, which is useful for modern workloads like AI inference. The encryption score of 33754 is moderate, but sufficient for secure communication protocols.

The single-thread score of 4597 is a standout, as it ranks the chip among the best for single-core performance. This is critical for applications that are not parallelized, such as older games or legacy business software. The multi-thread score of 54866, while lower than some higher-core-count rivals, is still robust for a 12-core part. The physics score of 3332 is lower than expected, which might indicate that this workload is not well-optimized for the chip’s architecture, but it does not impact overall usability.

The find prime numbers score of 461 is the weakest sub-score, but that test is known to be sensitive to memory latency and branch prediction, and the low value does not correlate with poor real-world performance. The random string sorting score of 71928 shows good memory and cache behavior, which is essential for database operations. Overall, the benchmark data paints a picture of a balanced processor that leads its immediate rivals in average score, except for the Xeon 6527P, and offers top-tier single-thread performance with respectable multi-thread capability. For users prioritizing memory bandwidth and data compression, this chip is particularly strong. The 4% deficit to the Xeon 6527P is the only notable weakness, and it affects only the most parallel workloads.