In the Lab: 8x AMD Radeon AI PRO R9700, One Fabric
AMD specs the R9700 for 4-way. We ran 8 as one fabric.
Every GPU-to-GPU pair: ~111 GB/s bidirectional, line-rate PCIe 5.0 x16. Flat topology, no cross-domain cliff.
ROCm + PyTorch. No NVLink, no CUDA. ~2× what NVIDIA targets for the RTX PRO 6000.
Full breakdown corespan.ai/resources/blog/8…
The inference era won't be won by the biggest boxes. It'll be won by the fabric that composes GPUs, memory & I/O around each workload, then reclaims it for the next one.
Tokens per dollar, not GPUs per rack.
corespan.ai/resources/blog/s…
Static servers waste compute and power.
AI needs infrastructure that adapts.
Drut’s photonic-native platform makes it possible to be dynamic, efficient, and always optimized for your workload.
🔗 drut.io/why-drut-%3F/f/beyon…
Thank you @OpenComputePrj for hosting the OCP Virtual Workshop. In the session “Break the Rack with Light,” Drut’s Simon McCormack shared how optical switching enables flexible, AI-ready data centers. Watch: youtu.be/PWzpjjfw3yc?feature…
Drut: Co-packaged optics slash AI power use & boost efficiency. Smarter interconnects like CPO are key to scaling inference. More in our blog. drut.io/drut-blog/f/scaling-…
2025 is the turning point for AI infrastructure.
Our latest blog breaks down how purpose-built systems with co‑packaged optics, photonic fabric & DynamicXcelerator are redefining performance.
drut.io/why-drut-%3F/f/2025-…
Drut partners with Ranovus to transform AI/ML & HPC infrastructure!
Powered by Odin® CPO 3.0 (6.4Tbps @ 4pJ/bit) + Drut’s 2500 series, enabling:
Dynamic GPU allocation
>1Tbps per slot
Lower latency & power
Future-ready data centers start here.
🔗 drut.io/drut-collaborates-ra…
As AI, HPC & data-heavy apps evolve, so must the network.
⚡ 15ns port-to-port latency
🔗 0.05–1.9dB insertion loss
📈 384×384 scalability
🌎 Protocol-agnostic, energy-efficient design
Rethink your data center foundation:
drut.io/drut-blog/f/the-powe…
GazettaByte’s latest briefing on how Drut's photonic switching transforms AI and HPC infrastructure.
🔗 Read the full briefing here: lnkd.in/e3PjuDTE








