Manufacturer Part Number: XPRM-X16-SIIB
“Add 4, 8, or 16 full-length, full-height PCI Express slots to your current rackmount workstations or servers - instead of replacing them!"
Xpander Rackmount is a 19" rack-mountable PC appliance which provides multiple servers and workstations with between 4 - 16 full-length PCI Express x16 slots, plus multiple, large capacity power supplies and exhaust fans required for operating high-performance discrete NVIDIA® graphics and GPU cards.
By connecting Cubix’s Xpander Rackmount systems to your current 19" rackmount standalone or blade servers, you instantly transform your legacy PC systems into supercomputers for running GPU-accelerated applications ranging from scientific research and CAE modeling/simulation software, to the latest high-performance digital content creation, linear editing, and digital intermediate solutions - dramatically enhancing performance of these applications by as much as 10x-50x or more beyond the capabilities of today’s multi-core CPU technology.
Xpander Rackmount provides an independent life support system for each server or workstation, and each with its own power and cooling, dedicated to the GPUs (as well as other PCI Express-based I/O) themselves. Because Xpander is a separate add-on module, you don’t tax your existing computer’s cooling capability, power supply(s), or violate its warranty by operating it beyond manufacturer’s specifications. You reduce your capital expense increments since you add two or more PCI Express slots to your existing computers instead of a new rackmount workstation with all of the CPU, memory, storage, and operating system license requirements.
Retrofit Existing PCs
If you are thinking about adding powerful graphics adapters to your existing computer, learn from the mistakes of others. There are a lot of uncertainties. Is there adequate power for one or more additional 200+ Watt GPUs in your current system? How about enough power for two double-wide 200+ Watt adapters? Will adding a high-powered graphics adapter overheat your workstation or server? Do you have enough physical space for one, two or more graphics cards inside your PC? With Cubix GPU-Xpander systems, questions about using an existing PC have answers.
Buying proprietary oversized workstations to run GPU-enabled applications may put you at odds with your Information Technology and Purchasing departments. Why invite conflict? Instead, connect your limited but familiar corporate-approved PC to an Xpander. This combination allows you to garner the best possible return on your PC investments from existing service contracts, familiar sourcing, and corporate standards purchasing perspectives. Contact Cubix for guidance on connecting your current deskside or rackmount computers to Xpander.
Standards-Based Open Architecture
Buying proprietary, oversized PCs to run GPU-enabled applications may put you at odds with your I.T. and Purchasing Departments. Why invite conflict? Instead, use your limited but familiar corporate-approved workstations and servers with Cubix Xpander Rackmount Systems. This combination allows you to garner the best possible return on your PC investments from existing service contracts and corporate standards purchasing perspectives.
Xpander Rackmount models support 2-16 PCI Express-based I/O adapters including, but not limited to:
- NVIDIA GeForce GTX 580/680
- NVIDIA Quadro 4000 / 4000 for Mac / 5000 / 6000
- NVIDIA Tesla C02075, K10
- NVIDIA Quadro DVP components
- Fusion-io® ioFX, and ioDuo SSD cards
- Blackmagic Design® Decklink HD Extreme 3D audio/video capture
- Red Rocket®, and many other 3rd party PCIe-based I/O adapters
Installation is easy. You simply insert a GPU-Xpander PCIe x8 or x16 adapter into your host computer's PCI Express slot, connect the cable to the GPU-Xpander, load the GPU’s software driver on your workstation or server, and you’re ready to harness the power of GPU-computing!
Applications For Xpander Rackmount models include:
- GPU-enabled rendering products from Mental Images®, Chaos Group® V-Ray RT 2.0, BunkSpeed®, Random Control® Arion, and others, along with network queue software from Axceleon™ for management purposes
- GPU-accelerated scientific research applications such as AMBER, MATLAB, and others
- CAE applications such as Autodesk®, Moldflow, ACUSIM AcuSolve 1.8, ANSYS Mechanical, and others
- GPU-accelerated oil and gas applications such as Acceleware® RTM Solver, Headwave, ffA SVI Pro, and others
- Financial applications such as RMS Risk Management Solutions, Numerix Risk, NAG RNG, and others
- GPU-accelerated features in the latest thin-client server software from Microsoft®, VMware®, and Citrix®
- Acceleware ISV applications seismic, electromagnetic, or financial modeling and simulations
- Image Processing applications such as Blackmagic® DaVinci Resolve for Mac, CinnaFilm® Pixel Strings, and others
- Many other GPGPU applications such as Siemens® 4D Ultrasound, Dalsa® Machine Vision, and Manifold® GIS
Xpander Rackmount Series II
Xpander Rackmount Series II is a 2U, external PCI Express (PCIe) expansion chassis that can support 1 server connection and up to 4 single-slot, discrete graphics cards or 2 dual-slot GPU cards per system. This enables massive parallel calculations separate from the server, greatly enhancing system performance and flexibility.
- Enhance existing server I/O infrastructure capabilities without adding servers
- Hardware-acceleration using NVIDIA graphics and GPU cards for faster HD-SDI capture / playback, VDI and thin-client multimedia performance
- Save space, weight and cost while increasing I/O card density
- Install GPUs for blazing fast hardware acceleration purposes, as well as other types of 3rd party I/O controllers.
- Four PCIe Gen 2 x16 expansion slots (supporting two dual-slot, 4 single-slot cards, or a combination of both formats
- Supports one host system in 2U, 19" rack mount system
- A 750W independent power supply, and variable speed. temperature-controlled cooling fans to meet almost any expansion requirement in environments ranging from IT Data centers to the most noise-sensitive digital production sets or post-production workstation offices
- Simple, quick installation with minimal impact per rackmount computer; install a HIC and drivers for devices you install in GPU-Xpander
- Each x8 or x16 cable and connector provides a 250 mating cycle duty cycle
- Add two dual-slot or 4 single-slot GPUs (or combination of each) for supporting a single rackmount computer which has a very limited slot expansion capabilities and limited power supply capacities
- Add 3rd party I/O cards such as RED Rockets, Blackmagic Design Decklink 4k / 3D Extreme, NVIDIA DVP, ATTO
- Add powerful GPUs like NVIDIA Quadro® 4000 / 5000 / 6000, Tesla® C2070 / K10 or GeForce® GTX 580 / 680 graphics cards, while minimizing the ambient noise impact on production sets and broadcast studios
Don't let your rackmount computers' lack of 'watts and slots' hold you back. Get your graphics or HPC processing on with Xpander Rack Mount 16 for up to 4 computers per unit. It will give your studio, engineering firm, or IT Data Center the graphics acceleration and HPC performance it needs at a reasonable cost with little or no interruption to routine or workflow.
GPU-Manager For Microsoft Windows
Cubix GPU-Manager is an application that provides management of graphics processing unit (GPU) clusters in Cubix GPU-Xpander enclosures. The primary aims of GPU-Manager are asset management, health monitoring and technical support data of GPUs. GPU-Manager runs under Microsoft Windows but does not support PCIe devices that are not GPUs. If you run Microsoft Windows on computers connected to GPU-Xpander, you can download GPU-Manager from the Cubix Website here and install it.
- Asset Management - GPU-Manager collects data from distributed GPU clusters into a central management console that lists the following:
- all attached GPU-Xpander enclosures
- the GPUs within each GPU-Xpander enclosure
- GPU capabilities such as number of GPU cores, amount of RAM, etc.
- operating system and software driver version numbers
- Monitoring - GPU-Manager continuously monitors GPU clusters, and provides the following:
- link status, number of lanes, and link speed for each enclosure
- temperature, fan speed and GPU load characteristics for each GPU