Take a look at report highlights outcomes from 4-, 6-, and 8-GPU configurations on real-world agentic AI workloads in manufacturing and monetary companies.
Principled Applied sciences (PT) has launched a brand new report evaluating the agentic AI efficiency of the Dell PowerEdge XE7740, a purpose-built enterprise AI server. PT testing measured how three GPU configurations of the server, powered by Intel Xeon 6747P processors and NVIDIA RTX PRO 6000 Blackwell Server Version GPUs, dealt with real-world agentic AI workloads within the manufacturing and monetary companies industries.
PT examined three configurations of the Dell PowerEdge XE7740: one with 4 GPUs, one with 6 GPUs, and one with 8 GPUs. On a monetary companies workload with a 30-second latency SLA, the 8-GPU configuration of the PowerEdge XE7740 supported as much as 74 concurrent AI brokers and sustained 1,179 tokens per second. On a producing workload with a 90-second SLA, the 8-GPU configuration of the PowerEdge XE7740 supported 15 concurrent brokers and delivered greater than double the throughput of the 4-GPU configuration.
Agentic AI represents a big change in enterprise AI infrastructure demand. The place a standard chatbot question would possibly contain a single inference name, a single agent activity can spawn tons of. That workload enhance falls on each the GPU and the CPU, making server platform selection essential for organizations planning manufacturing deployments of agentic AI.
Additionally Learn: AiThority Interview with Matej Bukovinski, Chief Know-how Officer at Nutrient
Key takeaways from the report included:
• Extra GPUs imply extra brokers—growing GPU rely additionally elevated the variety of brokers the server may assist
• Agentic AI calls for CPU in addition to GPU energy—PT noticed a variety of CPU utilization, peaking at 99% throughout activity task and agent orchestration
• On-premises possession has the potential to supply a value benefit—in comparison with Amazon Bedrock, the PowerEdge XE7740 supplied a decrease price per million tokens ($/Mtok)
To check the server, PT used “a customized agentic benchmark that runs real looking AI-agent workflows end-to-end on a computing resolution.” As PT notes within the report, “As a result of agentic AI brokers can do a variety of labor, a lot of it industry- and company-specific, we constructed the benchmark to simulate a variety of real looking industry-specific workflows. For this research, we used the monetary companies and manufacturing workflows, every of which includes eight workflow situations.”
Additionally Learn: AI techniques – Interoperable AI techniques: Connecting fashions throughout platforms
[To share your insights with us, please write to psen@itechseries.com ]
