Fujitsu ger deep learning-kapacitet för workstations och servrar via stöd för världens mest kraftfulla grafikkort

På tisdagskvällen offentliggjorde NVIDIA sitt nya grafikkort Quadro GV100, som därmed blir världens mest kraftfulla grafikprocessor för professionella användare. Nu meddelar Fujitsu att arbetsstationer i CELSIUS-familjen kommer att ha stöd för kortet, vilket öppnar upp dörren för helt nya möjligheter för professionella användare att använda program som bygger på deep learning.

Det nya grafikkortet Quadro GV100, som bygger på NVIDIAS Volta-arkitektur, blir därmed världens mest kraftfulla grafikprocessor för arbete med databehandling, AI, virtual reality eller avancerad 3D-simulering. Nu annonserar Fujitsu att utvalda workstations i CELSIUS-serien kommer att ha stöd för grafikkortet, vilket öppnar möjligheter för en rad yrkesområden att kunna köra programvara som är beroende av att kunna behandla enorma mängder rådata. Detta möjliggör användning av program som bygger på avancerad AI, maskinlärning samt deep learning, som kan användas i en rad branscher så som hälsa, transport, finans och tillverkning.

- Genom deep learning finns möjligheten att skapa applikationer som på egen hand har förmåga att anpassa sig till nya situationer och miljöer - likt människan genom evolutionen. Detta kan möjliggöra nya lösningar inom allt från att bekämpa sjukdomar till att förutse naturkatastrofer, bygga självkörande bilar eller utveckla smarta städer, säger Dr Joseph Reger, CTO på Fujitsu i Europa.

Sedan december i fjol har också Fujitsus PRIMERGY-servrar stöd för deep learning genom den senaste generationen av NVIDIAS Tesla V100 GPU. Detta innebär att Fujitsu nu kan erbjuda den mest högpresterande grafiken till både serverhallar och workstations.

PRIMERGY x86-servrarna RX2540 M4 and CX2570 M4 samt arbetsstationerna CELSIUS C740, CELSIUS R970 och CELSIUS M770 med stöd för NVIDIA Volta kommer från och med nu finnas tillgängliga på marknaden.

För mer information, läs nedanstående pressrelease

Fujitsu Adds Deep Learning Capabilities to Workstations and Servers

News facts:

• Fujitsu’s high-performance PRIMERGY servers and CELSIUS workstations power Deep Learning applications in offices and data centers

• Latest high-end CELSIUS workstations support the world’s most advanced visual computing platform, NVIDIA® Quadro® GV100

• Extreme memory capacity, scalability and performance will drive Artificial Intelligence and Machine Learning innovation in a wide range of application areas

Munich, March 28, 2018 – Fujitsu accelerates the adoption of Artificial Intelligence (AI) and Deep Learning (DL) with the global availability of ultra-high performance CELSIUS workstations and PRIMERGY server solutions which harness the latest visual computing platforms from NVIDIA. In addition to PRIMERGY servers with Deep Learning support, selected CELSIUS workstations now also support the NVIDIA Quadro GV100 card.

Launched today – the world’s most powerful architecture for high-performance computing, AI, virtual reality, simulations and graphics workloads on professional desktops.

Representing a significant increase in computing performance, Fujitsu server PRIMERGY with the latest NVIDIA Tesla® V100 Graphics processing Unit (GPU) and CELSIUS workstations featuring the new NVIDIA Quadro GV100 card at their core can deliver extraordinary speed and scalability¹. They are the ideal choice for demanding applications that rely on processing huge amounts of raw data, and by including these advanced visual computing capabilities across its workstation and server lines, Fujitsu can provide Deep Learning capabilities for desktops and data centers.

The new x86 PRIMERGY RX2540 M4² and PRIMERGY CX2570 M4³ servers as well as the datacenter-optimized CELSIUS C740 rack workstation, CELSIUS R970 and IF award winner CELSIUS M770 workstations deliver native support for the new generation of Volta-powered NVIDIA GPUs. Enhanced processing capabilities open up further opportunities for advanced AI, Machine Learning and Deep Learning applications, in sectors including healthcare and life sciences, automotive, financial services, science, and manufacturing.

Dr Joseph Reger, Fujitsu Fellow and CTO, EMEIA, Fujitsu says: “Artificial Intelligence, Machine Learning and Deep Learning are key technologies for the digital age and enable previously unknown levels of data insight and solution innovation. Deep learning allows us to build applications that can adapt to evolving situations much like humans do. This inspires exciting innovations – such as preventing disease, predicting hurricanes, building autonomous cars, and enabling smart cities. Our new servers and workstations make flexible and reliable processing power available to our customers – enabling support for new levels of application innovation, high-performance computing, and enterprise intelligence.”

Bob Pette, Vice President of professional visualization, NVIDIA says: “Immense computing power is required to process exponentially growing amounts of data. By partnering with Fujitsu, we are giving customers access to state-of-the-art visual computing platforms that represent an enormous leap forward in pushing processing performance limits.”

Pricing and availability

The PRIMERGY x86 Servers RX2540 M4 and CX2570 M4 as well as the Workstations CELSIUS C740, CELSIUS R970 and CELSIUS M770 with native NVIDIA Volta support are available immediately⁵.Pricing and availability may vary by region.

Notes to editors

¹ Through the 5120 CUDA Cores and additional 640 Tensor Cores for optimizing Machine Learning and Deep Learning computing. The NVIDIA Quadro GV100 card delivers up to 30% faster graphics performance and up to 62% faster render performance.

² The PRIMERGY RX2540 M4 2U dual-socket rack server can be equipped with two NVIDIA Tesla V100 GPUs for PCIe accelerators. Native support for the NVIDIA GPUs is via direct connection to the mainboard to guarantee the highest possible level of system performance, with more than 100 Teraflops per second.

³ The PRIMERGY CX2570 M4 dual-socket scale-out server can be equipped with either two NVIDIA Tesla V100 PCIe based GPUs or four NVIDIA Tesla V100 SXM2 based GPUs including Liquid Cooling. Native support for the NVIDIA GPUs is via direct connection to the mainboard to guarantee the highest possible level of system performance.

⁴ The CELSIUS R970 and CELSIUS M770 workstations each support up to two NVIDIA Quadro GV100 cards with NVLink bridge.

⁵ Restrictions may apply.

Fujitsu PRIMERGY servers and CELSIUS workstations with NVIDIA Volta units support the widely-used Deep Learning frameworks such as Caffe, CNTK, TensorFlow, Theano and Torch.

Machine Learning is the practice of using algorithms to parse data, learn from it, and then make a determination or prediction about something. Deep Learning is a technique for implementing Machine Learning using Artificial Neural Networks, which are inspired by an understanding of the biology of human brains that can learn from training.

Mer information
Läs Fujitsus blogg: http://blog.ts.fujitsu.com
Följ Fujitsu på Twitter: https://twitter.com/Fujitsu_Global
Följ oss på LinkedIn: http://www.linkedin.com/company/fujitsu-sweden
Följ oss på Facebook: http://www.facebook.com/FujitsuSverige
Fujitsus bildbank och mediaresurser: http://mediaportal.ts.fujitsu.com/pages/portal.php
Fujitsus nyhetsrum på MyNewsdesk: http://www.mynewsdesk.com/se/fujitsu

Presskontakt
Bo Lennström,
Produktchef Fujitsu PC, notebook och arbetsstationer
Tel: 08 793 7290
e-post: bo.lennstrom@se.fujitsu.com

Susanne Carlsson,
Senior Marketing Manager, Fujitsu Sverige
Tel: +46 (0)73 978 82 78
e-post: Susanne.carlsson@se.fujitsu.com