MangoBoost

MangoBoost is a DPU and AI infrastructure provider specialized in developing full-stack, advanced server accelerator solutions.

[MangoBoost @MangoBoost_Inc]

Proud to have teamed up with @AMD on @MLPerf Inference v5.1, with record results on Instinct™ GPUs and the first heterogeneous submission!

👉 Explore AMD’s technical deep dive: https://t.co/epm7HUKfgo

#AI #LLM #Inference #AMD #Instinct #GPU #Server

[MLCommons @MLCommons]

Thank you to our submitters @AMD, Amitash Nanda, @ASUS_OFFICIAL, @Broadcom, @Cisco, @CoreWeave, @DellTech, @gateoverflow, @GigaComputing, @Google, @HPE, @intel , KRAI, @LambdaAPI, @Lenovo, @MangoBoost_Inc, @Azure, @MiTACcomputing,
3/4

[MangoBoost @MangoBoost_Inc]

🚀 Try the AI acceleration solution proven in @MLPerf Inference v5.1, free of charge.

☑️ Run LLM on MI300X & MI325X servers
☑️ Experience LLMBoost™ and GPUBoost™ RDMA
☑️ See how MangoBoost accelerates AI inference

📷 Register: https://t.co/8NxEnHikNd

#AI #LLM #GPU #AMD

[MangoBoost @MangoBoost_Inc]

🚀 MangoBoost, @AMD, @Dell and @Supermicro_SMCI set new @MLPerf records!

✅ 648K tok/s on Llama2-70B (open)
✅ 1st heterogeneous GPU scaling (MI300X + MI325X)
✅ 1st 3rd-party multi-node MI355X showcase

📎 Read more: https://t.co/5P8REcH8Nr

#AI #LLM #AMD #GPU #Dell #SMCI

[MangoBoost @MangoBoost_Inc]

🚀MangoBoost's optimization software LLMBoost and @Supermicro_SMCI deliver 1.96× faster AI inference and 40% faster training with @AMD MI300X and MI325X clusters.

📙Read the whitepaper here: https://t.co/wl25i4y3Vv

#GPU #Server #LLM #AI #GenAI #LLaMa #ML #Software #SW

[MangoBoost @MangoBoost_Inc]

🚀MangoBoost sets a new @MLPerf Storage v2.0 record with DPU-accelerated @NVMexpress NVMe NVMe/TCP!

Up to 7.5x GPU scaling and local SSD performance over 400G Ethernet🔥

📎Read more: https://t.co/QhhhikVlIh

#AI #LLM #Storage #NVMe #TCP #GPU #SSD #Ethernet

[MangoBoost @MangoBoost_Inc]

🚀Mango StorageBoost™ - NTI is now @NVMexpress certified by the UNH Interoperability Labs.

SSD performance and simplicity, powered by NVMe-oF. Enterprise-ready.

#AI #Storage #NVMe #TCP #Initiator #DataCenter #Enterprise #Infra

[MangoBoost @MangoBoost_Inc]

🚀MangoBoost + @AMD Alveo™ U45N = 3× faster NVMe/TCP Initiator

Our StorageBoost™ NTI + U45N offload delivers 5.37M IOPS vs. 1.93M on NVIDIA CX6 (4KB reads) and ~60% better writes, with minimal CPU use.

Learn more: https://t.co/9kzI20A42X

#NVMe #TCP #Storage #AI #DataCenter

[MangoBoost @MangoBoost_Inc]

🔥MangoBoost LLMBoost™ cut Llama2-70B training time by half in @Dell 's test with AMD MI300X.

Multinode AI training, made easy.

🔗 https://t.co/GqW8yCeWZl

#Dell #AMD #GPU #LLM #AI #GenAI #MLPerf #Training

[MLCommons @MLCommons]

2/ Thank you to all 20 submitters:
@AMD , @ASUS, @Cisco , @CoreWeave , @Dell , @GigaComputing , @Google Cloud, @HPE , @IBM , Krai, @LambdaAPI , @Lenovo , @MangoBoost_Inc , @nebiusai , @nvidia , @Oracle , @QuantaQCT , @scitix , @Supermicro_SMCI , @tinycorp .

[MangoBoost @MangoBoost_Inc]

MangoBoost has signed an MoU with AIC to co-explore DPU-powered server & storage innovation.

Together, we aim to deliver scalable, AI-ready infrastructure for the enterprise market.

Efficiency meets performance.

Read more: https://t.co/yR9QjACl60

#DPU #Storage #Server

[MLCommons @MLCommons]

5/
KRAI, @LambdaAPI, @Lenovo, @MangoBoost_Inc, @nvidia, @Oracle, @QuantaQCT, @Supermicro_SMCI, @SMC_FutureAI, & @vmwarevcf
A special congratulations to CoreWeave, FlexAI, GATEOverflow, Lambda, and Mangoboost - who are all first time submitters to MLPerf Inference!

[MangoBoost @MangoBoost_Inc]

🚀 MangoBoost & @AMD set a new @MLCommons MLPerf Inference v5.0 record!

🔥 103,182 TPS (Offline) & 93,039 TPS (Server) on MI300X GPUs – 24% faster than NVIDIA H100

Read more: https://t.co/7BHjJFL5mg
#MLPerf #AI #LLM #Inference #AMD #MI300X #GPU #ROCm

[MangoBoost @MangoBoost_Inc]

🚀 Mango LLMBoost™ is now on Google Cloud Marketplace!

Optimize AI inference with our containerized solution—faster performance, lower costs, and seamless GPU efficiency.

Try it free today! → https://t.co/t8PblJFEos

#Google #GoogleCloud #GCP #Cloud #AI #LLM #Inference #GPU

[MangoBoost @MangoBoost_Inc]

🚀 Mango LLMBoost™ is now on Microsoft Azure!

Optimize GPU efficiency for AI inference with our containerized, deployment-ready solution.

✅ 12.6x AI inference performance
✅ 92% cost savings

Try it FREE today: https://t.co/BskesSwtKd

#Microsoft #Azure #AI #Inference

[MangoBoost @MangoBoost_Inc]

🚀 Chanik Park, ex-Samsung Semiconductor exec, joins Mangoboost as Chief Commercial Officer.

With his global scaling expertise, he's ready to supercharge our growth and redefine data center efficiency.

Welcome aboard, Chanik!

#DPU #DataProcessingUnit #DataCenter

[MangoBoost @MangoBoost_Inc]

🎊Mango StorageBoost™ is now NVM Express™ compliant!

✅Seamless integration
✅Faster performance
✅Future-ready for AI & big data

Unlock your storage potential today: https://t.co/n9436f6Vj5

@NVMexpress

#NVMe #NVMExpress #Storage #DataCenter

[MangoBoost @MangoBoost_Inc]

More information on Mango LLMBoost™: https://t.co/JBgSwQUb1L

[MangoBoost @MangoBoost_Inc]

🚀 Introducing Mango LLMBoost™: 12.6x faster and 92% cheaper AI inference!

- Flexible GPU support (NVIDIA & AMD)
- Multi-model management
- Easy deployment with web APIs
- OpenAI-compatible

Try for FREE on AWS (link in comment)!

#LLM #AI #Inference

[MangoBoost @MangoBoost_Inc]

Try Mango LLMBoost™ on AWS for FREE: https://t.co/upnkdKYGjz