BOXX Technologies Introduces new APEXX W3 workstation and GX8-P Deep Learning Server
Latest News
April 18, 2018
BOXX Technologies recently announced its new APEXX W3 workstation for high-performance computing applications. Simultaneously, it introduced its GX8-P deep learning server.
Unveiled late last month at NVIDIA’s GPU Technology Conference, the APEXX W3, BOXX says, is ideal for data scientists since it is capable of enabling deep learning development at the user’s deskside. The workstation can support up to four NVIDIA Quadro GV100 graphics cards, which, BOXX explains, can help users iterate and test code rapidly prior to large-scale deep learning deployments while also being an ideal platform for GPU (graphics processing unit) accelerated rendering.
The APEXX W3 is equipped with an Intel Xeon W series processor with up to 18 cores (36 threads). In addition to its support for up to four dual-slot NVIDIA GPUs, the APEXX W3's compact 6.84x15.28x17.8 in. (WxHxD; 17.4x38.8x45.2 cm) chassis can accommodate up to 512GB of system memory. A variety of SSD (solid-state drive) and SATA drive options are offered. It runs under 64-bit Microsoft Windows 10 Professional and comes with a full complement of USB 3.0/3.1 ports as well as networking and other connections.
“Driven by NVIDIA Volta GPU architecture, the Quadro GV100 is the most powerful graphics card ever created for PC workstations,” said Bob Pette, vice president, Professional Visualization, NVIDIA, in a testimonial provided by BOXX. “By integrating it into the new APEXX W3, BOXX has once again demonstrated their commitment to enabling the most advanced, professional GPU capability for complex deep learning and rendering workflows.”
BOXX's new GX8-P deep learning server features a single AMD EPYC processor with up to 32 high-performance cores. This rackmount server is available with up to eight full-size NVIDIA Quadro or Tesla GPUs, or AMD Radeon Pro or Radeon Instinct graphics cards for rapid training of deep neural networks. It can accommodate up to 1TB of memory, and it also provides room for additional InfiniBand or NVMe (non-volatile memory express) storage devices while enabling increased bandwidth and lower latencies between PCIe Gen3 devices than are possible in traditional systems, according to BOXX.
“BOXX is taking the lead with deep learning servers that allow organizations to scale out, as well as solutions like the APEXX W3 which brings deep learning and GPU-accelerated rendering to your deskside,” said Shoaib Mohammad, BOXX VP of Marketing and Business Development in a press statement.
Related to that first point, the BOXX-owned Cirrascale Cloud Services company provides dedicated, multi-GPU cloud-based deep learning solutions. The company offers its cloud-based solutions to large-scale deep learning operators and service providers as well as HPC (high-performance computing) users.
In other news, BOXX also announced that it will soon launch a deskside workstation engineered for deep learning development as well as a variety of professional visualization applications. To be called the APEXX P5, this workstation, BOXX reports, will support up to four dual-slot NVIDIA graphics cards featuring the latest Volta architecture and Tensor Cores, creating an advanced workstation for GPU-centric workflows. The system will come with a single AMD EPYC processor with up to 32 high-performance CPU cores. Further details were unavailable at press time.
Click here for more details on the APEXX W3.
Download the APEXX W3 spec sheet.
Click here for more details on the GX8-P server.
Download the GX8-P Server spec sheet.
Read “NVIDIA Updates Workstation for Deep Learning.”
Read “Graphics Power to Change Engineering Workflows.”
See why DE's Editors selected the APEXX W3 and the GX8-P Server as their Pick of the Week.
Sources: Press materials received from the company and additional information gleaned from the company's website.
Subscribe to our FREE magazine,
FREE email newsletters or both!Latest News
About the Author
Anthony J. LockwoodAnthony J. Lockwood is Digital Engineering’s founding editor. He is now retired. Contact him via [email protected].
Follow DE