HYPERPC G9 PRO HYDRO Maintenance: 96 Cores and 6 GPUs
Client's Objective
This is an unusual case: the client did not come to us for a new build, but to save an existing project. The client owns an uncompromising workstation built for the most complex mathematical calculations, machine learning, and rendering.
- Scope of activity: high-load computing, neural network workloads, and Big Data.
- Key issue: the workstation had operated around the clock at full load for more than a year without scheduled maintenance.
- Risk: temperatures of key components were rising rapidly, threatening thermal throttling and interruptions to critical workloads.
Configuration
- Processor:AMD Ryzen Threadripper PRO 7995WX [up to 5.1GHz, 96 cores]
- Graphics cards:6 × NVIDIA RTX 6000 Ada Generation [48GB, 18176 CUDA]
- Motherboard:ASUS Pro WS WRX90E-SAGE SE [DDR5]
- RAM:1024GB RDIMM ECC [DDR5, 4800MHz, 4 × 256GB]
- Storage:4 × 4TB Samsung 990 PRO and 6 × 18TB Seagate Exos X managed by Broadcom MegaRAID
- Power supplies:2 × 2050W SilverStone HELA 2050R [80+ Platinum]
- Case and cooling:Lian Li V3000 PLUS Black, dual custom liquid cooling loops with two 480mm radiators
Cause of Overheating
The HYPERPC G9 PRO HYDRO removes heat from ultra-powerful components through two independent liquid cooling loops with separate pumps and reservoirs.
After a year of continuous extreme loads, the EK-CryoFuel Clear coolant had naturally degraded. The resulting sediment blocked the microchannels in the processor water block, restricting coolant flow and causing temperatures to rise sharply.
Completed work
HYPERPC service engineers carried out a complete set of restoration procedures:
- completely disassembled both liquid cooling loops;
- flushed the radiators, hoses, and distribution plate;
- disassembled and cleaned seven water blocks — six for the graphics cards and one for the processor;
- refilled the system with fresh coolant and restored normal circulation.
Coolant degradation is a natural process. Custom liquid cooling systems that operate around the clock under heavy load should be serviced every 10–12 months.
Result
The workstation returned to its original temperature levels. The customer received a fully serviced, stable, and quiet system ready for more months of continuous computing.
This project demonstrated in practice that regular servicing is not merely a recommendation, but an essential condition for reliable operation of premium equipment under continuous extreme load.
Does Your Workstation Need Maintenance?
We will diagnose and carry out deep technical maintenance on your high-performance system, restore its cooling, and prepare the equipment for reliable operation under load.