Solutions
AI Datacenter Construction
Datacenter construction
Updated
The problem
Datacenter proposals often give conclusions without derivations: how the network is laid out, how much storage is needed and when the investment pays back all rest on trusting a spreadsheet.
Our approach
From a 128-GPU joint test to a thousand-GPU datacenter, we deliver plans you can recompute: network BOM, three-tier storage with KV tiering, power and PUE, and TCO are all modeled in Python for you to rerun with your own parameters.
Key points and sources
- A thousand-GPU plan (128 nodes × 8 MI308X) with a complete TCO model: network, storage, power and depreciation modeled line by line, plus a 10,000-run Monte Carlo sensitivity analysis
- Gated joint testing: G1 delivery acceptance → G2 single-node baseline → G3 main gate (TTFT reduction ≥25%, throughput inside the measured +29–40% bandR2R3) → G4 72-hour stability, with a stop-loss at every gate
- Storage sized on a production basis: one FX100 per 8 nodes, anchored on the measurement that 100 GbE sustains KV read-back for 8 nodesR1R2R3
- Price anchors for networking and key components cite public sources and can be checked item by item
Related test reports
- R1FX100 Comprehensive LLM Inference and Training Benchmark (8 × AMD MI308X)One report, five questions: how much the LMCache parallel-read patch helps, whether KV tiering scales across GPUs, how much faster 8 GPUs load models concurrently, how effective utilization changes under frequent model switching, and how much faster concurrent training checkpoint writes become. Test models are Qwen2.5-32B and 7B.
8 × AMD Instinct MI308X · ROCm 7.2 - R2FX100 KV Cache Benchmark (480B, TP8, Long Context)Qwen3-Coder-480B-FP8 deployed as one 8-GPU instance (TP=8), the standard production topology, under a long-context cold-recovery load. A three-way comparison of FX100, a single local NVMe drive and full recompute without external storage, stepped across concurrency 8 / 16 / 32. Most throughput and TTFT figures quoted on this site come from this report.
8 × AMD Instinct MI308X · ROCm 7.2 - R3FX100 KV Cache Benchmark Summary (480B, TP4×2, All Metrics)The same 8-GPU server redeployed as two 4-GPU instances (TP4×2), with every metric recorded across nine load levels: TTFT p50 / p90 / p99, TPOT, throughput and disk bandwidth. The report also verifies two things: the fs:// shared pool lets both instances hot-share one cache, and two independent runs at the same level deviate by no more than 5%.
8 × AMD Instinct MI308X · ROCm 7.2
Related products
Test jointly first, decide afterwards
Every engagement starts with a gated joint test of about 10 weeks. Four gates are passed in sequence; if one is missed, the project stops there, so you never pay for a benefit that was not demonstrated.
G1Week 2
Delivery acceptance
All materials delivered and inspected; every array passes its power-on self-test with all 24 drives recognized.
G2Week 4
Single-node baseline
The single-node 8-GPU inference baseline is reproduced within ±10% of the R1 figures; a local stress test of one array reaches 83% line efficiency against the measured anchor.
G3Week 7
Main joint-test gate
KV tiering must deliver both a TTFT reduction of ≥25% and a throughput gain inside the measured +29–40% band; with 8 nodes sharing one array, read-back bandwidth approaches the NIC ceiling.
G4Week 10
Stability and sign-off
72 hours of continuous load without interruption; single-drive and single-link fault injection stays invisible to the workload; a signed joint-test report is issued.
Other solutions
- Storage Acceleration: KV Cache Tiering
Storage acceleration
- Enablement for Domestic and Non-NVIDIA Accelerators
Accelerator enablement
- AI Datacenter Efficiency Optimization
Efficiency optimization
- Inference-Stack Software Development
Software development
Test jointly first, decide afterwards
Test jointly first, decide afterwards
Every engagement starts with a gated joint test of about 10 weeks. Four gates are passed in sequence; if one is missed, the project stops there, so you never pay for a benefit that was not demonstrated.
- Contact
- Wang Limin
- Phone
- +86 131 6181 8898
- mingxin@agentmail.to
The inbox is monitored and triaged by AI around the clock