High-throughput compute for local AI models, 3D rendering, and heavy compilation
Heavy production environments demand sustained parallel execution without hitting hardware bottlenecks. The Ryzen Threadripper 9980X provides 128 execution threads, allowing software developers, 3D technical artists, and data engineers to run complex offline simulations, compile massive codebases, and train local machine learning models simultaneously. Because the CPU features an unlocked multiplier, system builders can fine-tune clock frequencies and power limits to match specialised workstation demands on supported TRX50 motherboards.
64 Zen 5 cores with 256 MB L3 cache for multi-threaded efficiency
Built on AMD's Zen 5 architecture using a TSMC 4nm FinFET manufacturing process across 9 package dies, the Threadripper 9980X pairs single-thread responsiveness with high multi-core throughput. A 3.2 GHz base clock provides solid baseline performance across all 64 cores, while individual cores can dynamically scale up to a 5.4 GHz boost clock when thread-bound processing requires maximum speed. To keep data moving efficiently to the execution pipelines, the processor incorporates a multi-tier cache structure:
- L1 Cache: 5,120 KB total for immediate instruction delivery
- L2 Cache: 64 MB dedicated cache across the core dies
- L3 Cache: 256 MB of shared memory to prevent memory stalls in large datasets
This cache arrangement directly supports advanced instruction sets including AVX-512, AVX2, AES, FMA3, and SHA extensions for acceleration in scientific computation and cryptographic processing.
88 usable PCIe 5.0 lanes for multi-GPU and high-speed storage arrays
Modern workstations rely heavily on direct PCI Express bandwidth for add-in compute accelerators, high-speed networking, and storage controllers. The sTR5 platform provides 92 total PCIe 5.0 lanes, with 88 usable lanes directly exposed for system expansion. This bandwidth capacity allows users to run multiple full-speed PCIe 5.0 graphics cards alongside dedicated RAID storage arrays without dynamic lane sharing or bandwidth throttling.
The companion I/O die, fabricated on a TSMC 6nm process, manages direct interface connectivity across the system platform:
| Usable PCIe 5.0 Lanes | 88 Lanes |
| NVMe RAID Support | Boot, RAID 0, RAID 1, RAID 5, RAID 10 |
| Native USB Connectivity | 4 x USB 3.2 Gen 2 ports |
| Native SATA Ports | 2 x SATA ports |
Quad-channel DDR5 ECC RDIMM support up to 6400 MT/s
Data integrity is critical in enterprise workstation tasks where an uncorrected memory error can halt hours of processing. The Threadripper 9980X features a quad-channel memory controller supporting ECC (Error-Correcting Code) DDR5 RDIMMs at transfer rates up to 6400 MT/s. The four memory channels double available bandwidth compared to standard dual-channel platforms, feeding heavy memory-bound applications without latency spikes.
Hardware-level security and management features are integrated directly into the processor silicon. AMD Memory Guard provides real-time system memory encryption to protect sensitive data against physical access, supported by an embedded AMD Secure Processor and DASH management protocols for remote administration.
350W TDP on the sTR5 socket requiring dedicated cooling and discrete graphics
Operating with a Thermal Design Power (TDP) of 350W and a maximum operating temperature rating of 95°C, the Ryzen Threadripper 9980X requires a dedicated thermal cooling solution built for the sTR5 socket. The processor contains no integrated graphics, requiring a discrete graphics card for video output. It installs directly into sTR5 socket motherboards paired with the TRX50 chipset, providing a scalable foundation for professional workstation builds.