B300 and MI355X Both Sustain 52 Agents, but the Curves Differ Wildly
ryanshrout · x · 2026-09-19
Signal65's PINNACLE testing (cited by Ryan Shrout) shows that on a fully loaded 8-GPU node running MiniMax-M3 on agentic workloads, NVIDIA B300 and AMD MI355X both sustain 52 concurrent agents at the service level (every agent ≥10 tokens/s).
But the underlying curves differ sharply: B300 is 3.7x faster per agent at 8 agents, 2.4x at 16, and still 1.4x at the ceiling — then falls off a cliff one step later. MI355X starts lower, declines gently, and keeps first token flat at 0.2s.
Takeaway: headcount says how many agents a node holds; the curve says what those agents feel like and where the node stops being productive. Both belong in buying decisions, along with cost per correct task.
Related event: Signal65 Benchmarks Show AMD MI355X Doubles MI300X in Concurrent Agents(3 posts)→
More from Infra
- Fire at Taiwan Panasonic PCB materials plant may tighten CCL supply further — zephyr_z9 · 2026-11-04
- AMD CEO Lisa Su says revenue to grow over 80% in 2027, data center to top 80% of sales — AccBalanced · 2026-11-04
- Self-improving software is inevitable: closing the agent-observability loop — bibryam · 2026-11-04
- Gigabyte's AI TOP motherboard packs 2TB RAM and 9 GPUs for home AI labs — vincent_koc · 2026-11-04
- Building a 7-GPU 132GB VRAM local inference rig: consolidate on EPYC or split across two systems? — infectiousstupidity · 2026-11-04
- imec neural probe integrates photonics directly into 130nm CMOS backend — jwt0625 · 2026-11-04