Principled Tech

Facts matter. Partner with PT for all your marketing, learning, and testing needs.

[Principled Tech @PrincipledTech]

Half the RAM, better performance: In our tests, an @HP #EliteBook 8 G2i with an @intel Core Ultra 7 355 processor and 16 GB of RAM delivered up to 59.1% better productivity scores than prior-gen 32GB EliteBooks. Get the details: https://t.co/IJXEMseKhR
#PCs

[Principled Tech @PrincipledTech]

Game on: In PT testing, @surface Laptop Ultra with NVIDIA #RTXSpark delivered up to 44% more frames per second in Cyberpunk 2077 than an Apple MacBook Pro with M5 Pro. See the results, commissioned by @Microsoft. https://t.co/rJdSDVbvd1
#PCGaming

[Principled Tech @PrincipledTech]

We found an @AMD #RyzenAI Max+ powered mobile workstation can pay for itself in under 16 months running AI coding agents locally instead of in the cloud, saving $6,573 in cloud spend over 3 years. https://t.co/YjVf2MwJja
#OnDeviceAI

[Principled Tech @PrincipledTech]

Worried about #agenticAI costs per token? In our tests, a @Dell #PowerEdge XE9780LAP with @intel Xeon 6 CPUs and 8 @nvidia B300 GPUs spent as little as $0.25/Mtok on coding and $0.43/Mtok on manufacturing workloads. https://t.co/4f7ouf6mbq
#Tokenomics

[Principled Tech @PrincipledTech]

Could a single-socket server refresh help a growing business get more from SQL Server Standard Edition? We compared two @Dell #PowerEdge R4715 configurations with a legacy HPE server: https://t.co/eOUVtYfyjk

[Principled Tech @PrincipledTech]

Do you need GPUs to run agentic AI? On a financial-services workload, a @Dell PowerEdge R770 with #Intel Xeon 6 processors completed 110 agent tasks per minute within a 30-second service target: https://t.co/96B7ckcFtD
#AgenticAI

[Principled Tech @PrincipledTech]

Boost student multitasking: A @MediaTek Kompanio 540 processor-based #Chromebook offered up to 64% faster app responsiveness while running background apps on battery vs. Intel processor-based Chromebooks in our testing: https://t.co/q74ulTRbwi #EdTech

[Principled Tech @PrincipledTech]

Ship code faster with AI-driven fixes: up to 14 coding agents for as little as $0.25 per million tokens on the @Dell PowerEdge XE9780LAP with @intel Xeon 6979P CPUs and NVIDIA B300 GPUs. https://t.co/0gTJT4gL4n
#AgenticAI

[Principled Tech @PrincipledTech]

Refreshing your @Dell PC fleet in waves? PT hands-on testing showed IT can deploy updates, apps, and patches to @AMD or Intel processor-powered devices from the same portal, with no extra time, effort, or logins. https://t.co/WCG4QKFv6m
#RyzenPRO

[Principled Tech @PrincipledTech]

See how @AzureSQL Database Hyperscale delivered up to 52% higher performance and 51% better performance per dollar than Amazon Aurora PostgreSQL in PT's hands-on database tests: https://t.co/6GhTN2rjsd
#managedDBservice #SQL #cloud

[Principled Tech @PrincipledTech]

Agentic coding for $0.25 per million tokens: PT found a @Dell PowerEdge XE9780LAP with @intel Xeon 6 CPUs and eight NVIDIA B300 GPUs supported 14 AI agents on a complex code generation workload. https://t.co/AwzSomNAYO
#AgenticAI

[Principled Tech @PrincipledTech]

Get up to 2.4x the SQL Server NOPM on a @DellServers #PowerEdge R4715 w/ 32-core @AMD EPYC CPUs vs. an older HPE server. Plus, see up to 87% more performance per watt & 67% more NOPM per dollar: https://t.co/6YFaZJPTzp #SQL

[Principled Tech @PrincipledTech]

Same 8 NVIDIA RTX PRO 6000 GPUs, very different results. In our tests, a @googlecloud G4 VM delivered up to 17.6% more output token throughput and 26.7% faster TTFT than AWS EC2 G7e. The key? All-to-all P2P GPU fabric. https://t.co/yUNhKOE1O4
#CloudPerformance #AIInference

[Principled Tech @PrincipledTech]

New report: Put agentic AI to work on familiar, general purpose servers: A @Dell #PowerEdge R770 with #Intel #Xeon 6 processors completed 110 agentic AI tasks per minute through OpenVINO, no GPU required: https://t.co/96B7ckcFtD

[Principled Tech @PrincipledTech]

Running LLMs locally? In llama.cpp testing on the Qwen3.5 27B model, @surface Laptop Ultra with NVIDIA #RTXSpark reached the first token up to 2.1x faster than an Apple MacBook Pro with M5 Pro. https://t.co/rJdSDVbvd1
#AI

[Principled Tech @PrincipledTech]

We found that a new @HP EliteBook 8 G2i with an @intel Core Ultra 7 355 processor and 16 GB of RAM outperformed two prior-gen 32GB EliteBooks, including up to 111.6% better video editing performance. Read more: https://t.co/IJXEMseKhR
#laptops

[Principled Tech @PrincipledTech]

Stretch your IT budget: A @Dell #PowerEdge R4715 with a 32-core @AMD #EPYC processor delivered up to 67% more SQL Server performance per dollar than a legacy HPE ProLiant DL320 Gen11 in our tests: https://t.co/eOUVtYfyjk

[Principled Tech @PrincipledTech]

New report: Benchmark testing: Current-generation @Windows 11 PCs vs. five-year-old Windows PCs across five device collections https://t.co/YFbOBnHP5K

[Principled Tech @PrincipledTech]

Stay on the call when Wi-Fi isn't available. In PT's cellular-only Teams testing, the @HP #EliteBook X G2a held up to 21.3 dBm more signal in the weakest spots and cut audio round-trip time by up to 41% on a commute: https://t.co/yThsUXMpeH

[Principled Tech @PrincipledTech]

See how @surface Laptop Ultra and other Windows laptops with NVIDIA #RTXSpark compared to an Apple MacBook Pro with M5 Pro in AI, gaming, and 3D tests commissioned by @Microsoft https://t.co/rJdSDVbvd1