Supercharge your next cluster with Tesla Kepler K10, K20 or K40 GPUs
Microway NVIDIA® Tesla™ GPU Powered High Density Clusters
Microway’s preconfigured Tesla GPU clusters deliver supercomputing performance at a lower power, lower cost, and using many fewer systems than standard CPU-only clusters. These clusters are powered by NVIDIA Tesla Kepler K10, K20/K20X and K40 GPUs. With up to 12 GB of memory, 2880 CUDA cores per GPU, support for IEEE 754 single & double floating point precision, and a fast GDDR5 interface to memory, Tesla GPUs speed the transition to energy-efficient parallel computing and scale to solve the world’s most important computing challenges more quickly and accurately. Successfully deployed in demanding applications at research institutes, universities, and enterprises, Tesla GPUs power the most powerful supercomputers worldwide.
Installed Software
A Microway Tesla preconfigured cluster comes preinstalled with:
- Linux distribution of your choice, including Red Hat, SuSE, Debian, Fedora and Gentoo
- NVIDIA CUDA Toolkit, Libraries and SDK
- Bright Cluster Manager or Microway Cluster Management Software (MCMS) integrated with optional MPI Link-Checker™ Fabric Validation Suite
- User-level applications and libraries optional
GPU Speedups
Applications running CUDA-based code on Tesla GPUs see speedups of up to 250x in domains ranging from MATLAB to computational fluid dynamics, molecular dynamics, quantum chemistry, imaging, signal processing and bio-informatics. See NVIDIA’s full list of GPU-accelerated applications.
GPU Software Development Tools
It is easier than ever to harness GPUs in your applications. Many libraries include pre-optimized GPU support; language bindings are available for Python, Java, C#, C, C++, Fortran and more. Visit NVIDIA’s GPU Software Development Tools site.
Maximum GPU Memory and Compute Performance: Tesla K40 GPU Accelerator
Integrated in Microway NumberSmasher GPU Servers and GPU Clusters
Specifications
- 2880 CUDA GPU cores (GK110b)
- 4.2 TFLOPS single; 1.4 TFLOPS double-precision
- 12GB GDDR5 memory
- Memory bandwidth up to 288 GB/s
- PCI-E x16 Gen3 interface to system
- GPU Boost increased clock speeds
- Supports Dynamic Parallelism and HyperQ features
- Passive heatsink relies on chassis cooling of specially-designed GPU servers
Tesla K20 and K20X GPU Accelerators
Integrated in Microway NumberSmasher and Navion GPU Servers and GPU Clusters
Specifications
- 2496 CUDA cores in Tesla K20 (GK110)
2688 CUDA cores in Tesla K20X (GK110) - K20: 3.52 TFLOPS single, 1.17 TFLOPS double
K20X: 3.95 TFLOPS single, 1.32 TFLOPS double - 5GB GDDR5 memory per Tesla K20
6GB GDDR5 memory per Tesla K20X - Memory bandwidth up to:
208GB/sec in Tesla K20
250GB/sec in Tesla K20X - PCI-E x16 Gen2 interface to system
- Supports Dynamic Parallelism and HyperQ features
- Active and Passive heatsinks available for installation in workstations and specially-designed GPU servers
Single-Precision Optimized Tesla K10 GPU Accelerator
Integrated in Microway NumberSmasher and Navion GPU Servers and GPU Clusters
Specifications
- 2 x GK104 GPU chips on a single board (PCB)
- Total of 3072 CUDA cores and 8GB memory
(each GPU provides 1536 CUDA cores and 4GB GDDR5 memory) - 4.58 TFLOPS single-precision
0.19 TFLOPS double-precision
(each GPU provides 2.29 TFLOPS single-precision; 95 GFLOPS double-precision) - Memory bandwidth of up to 160 GB/s per GPU chip (320 GB/s in aggregate)
- PCI-E x16 Gen3 interface to system
- Passive heatsink relies on chassis cooling of specially-designed GPU servers
Sample Microway Tesla GPU Cluster Specifications
One CPU and Two GPUs with 1U NumberSmasher compute nodes
| GPUs per Node | (2) Tesla K10, K20/K20X or K40 |
|---|---|
| Sample Cluster Size | 16 Nodes (32 GPUs) in a 24U rackmount cabinet |
| Base Platform | 1U Rackmount Server (1) Intel Xeon E5-1600v2 or E5-2600v2 series “Ivy Bridge” CPU |
| System Memory per Node | 256 GB DDR3 1866MHz / 1600MHz |
| Total GPU Memory per Node |
16GB (Tesla K10) 10GB (Tesla K20) 12GB (Tesla K20X) 24GB (Tesla K40) |
| Head Node | Dual Xeon Server (1U – 4U) with up to 768 GB memory Optional NVIDIA Quadro Graphics |
| Storage | Head Node: up to Compute Nodes: up to Optional Storage Servers or Parallel HPC Storage System |
| Network | Dual Gigabit Ethernet built-in Optional 10Gb or 40Gb Ethernet |
| Interconnect (optional) | ConnectX-3 56Gb FDR or 40Gb QDR InfiniBand Fabric |
| Cabinet | 24U APC NetShelter Cabinet or scale up with 42U APC NetShelter Cabinet |
| Green HPC Features | High-efficiency (80PLUS Platinum-Level) power supplies Optional low-power Intel CPUs Software/firmware to reduce power consumption on idle cores Cooling paradigms designed to reduce power consumption |
Two CPUs and Three GPUs with 1U NumberSmasher compute nodes
| GPUs per Node | (3) Tesla K10, K20/K20X or K40 |
|---|---|
| Sample Cluster Size | 16 Nodes (48 GPUs) in a 24U rackmount cabinet |
| Base Platform | 1U Rackmount Server with optional redundant power supplies (2) Intel Xeon E5-2600v2 series “Ivy Bridge” CPUs |
| System Memory per Node | 256 GB DDR3 1866MHz / 1600MHz |
| Total GPU Memory per Node |
24GB (Tesla K10) 15GB (Tesla K20) 18GB (Tesla K20X) 36GB (Tesla K40) |
| Head Node | Dual Xeon Server (1U – 4U) with up to 768 GB memory Optional NVIDIA Quadro Graphics |
| Storage | Head Node: up to Compute Nodes: up to Optional Storage Servers or Parallel HPC Storage System |
| Network | Dual Gigabit Ethernet built-in Optional 10Gb or 40Gb Ethernet |
| Interconnect (optional) | ConnectX-3 56Gb FDR or 40Gb QDR InfiniBand Fabric |
| Cabinet | 24U APC NetShelter Cabinet or scale up with 42U APC NetShelter Cabinet |
| Green HPC Features | High-efficiency (80PLUS Platinum-Level) power supplies Optional low-power Intel CPUs Software/firmware to reduce power consumption on idle cores Cooling paradigms designed to reduce power consumption |
Two CPUs and Four GPUs with 2U NumberSmasher compute nodes
| GPUs per Node | (4) Tesla K10, K20/K20X or K40 |
|---|---|
| Sample Cluster Size | 8 Nodes (32 GPUs) in a 24U rackmount cabinet |
| Base Platform | 2U Rackmount Server with redundant power supplies (2) Intel Xeon E5-2600v2 series “Ivy Bridge” CPUs |
| System Memory per Node | 512 GB DDR3 1866MHz / 1600MHz |
| Total GPU Memory per Node |
32GB (Tesla K10) 20GB (Tesla K20) 24GB (Tesla K20X) 48GB (Tesla K40) |
| Head Node | Dual Xeon Server (1U – 4U) with up to 768 GB memory Optional NVIDIA Quadro Graphics |
| Storage | Head Node: up to Compute Nodes: up to Optional Storage Servers or Parallel HPC Storage System |
| Network | Dual Gigabit Ethernet built-in Optional 10Gb or 40Gb Ethernet |
| Interconnect (optional) | ConnectX-3 56Gb FDR or 40Gb QDR InfiniBand Fabric |
| Cabinet | 24U APC NetShelter Cabinet or scale up with 42U APC NetShelter Cabinet |
| Green HPC Features | High-efficiency (80PLUS Platinum-Level) power supplies Optional low-power Intel CPUs Software/firmware to reduce power consumption on idle cores Cooling paradigms designed to reduce power consumption |
Two CPUs and Four GPUs (plus a 5th slot) with 4U NumberSmasher compute nodes
| GPUs per Node | (4) Tesla K10, K20/K20X or K40 Optional NVIDIA Quadro Graphics in 5th slot |
|---|---|
| Sample Cluster Size | 4 Nodes (16 GPUs) in a 24U rackmount cabinet |
| Base Platform | 4U Rackmount Server with redundant power supplies (2) Intel Xeon E5-2600v2 series “Ivy Bridge” CPUs |
| System Memory per Node | 512 GB DDR3 1866MHz / 1600MHz |
| Total GPU Memory per Node |
32GB (Tesla K10) 20GB (Tesla K20) 24GB (Tesla K20X) 48GB (Tesla K40) |
| Head Node | Dual Xeon Server (1U – 4U) with up to 768 GB memory Optional NVIDIA Quadro Graphics |
| Storage | Head Node: up to Compute Nodes: up to Optional Storage Servers or Parallel HPC Storage System |
| Network | Dual Gigabit Ethernet built-in Optional 10Gb or 40Gb Ethernet |
| Interconnect (optional) | ConnectX-3 56Gb FDR or 40Gb QDR InfiniBand Fabric |
| Cabinet | 24U APC NetShelter Cabinet or scale up with 42U APC NetShelter Cabinet |
| Green HPC Features | High-efficiency (80PLUS Platinum-Level) power supplies Optional low-power Intel CPUs Software/firmware to reduce power consumption on idle cores Cooling paradigms designed to reduce power consumption |
Additional GPU Server Options
Microway offers a wide variety of servers optimized for GPUs. Any of these may be custom-integrated and combined into a unique configuration to fit your requirements. View a full list of Microway’s GPU-Accelerated servers.
External GPU Acceleration Systems
An external GPU chassis allows addition of GPU compute capability to your existing HPC infrastructure. We work with a number of vendors and can supply a GPU system compatible with almost any of your servers or workstations.
These external systems support both PCI-Express generation 2.0 and generation 3.0 GPU devices. Up to 8 GPUs may be connected to a single server via 1-meter or 2-meter cables.
Contact us to learn more about compatibility and pricing.
Supported for Life
Our technicians and sales staff consistently ensure that your entire experience with Microway is handled promptly, creatively, and professionally.
Telephone support is available for the lifetime of your cluster by Microway’s experienced technicians. After the initial warranty period, hardware warranties are offered on an annual basis. Out-of-warranty repairs are available on a time & materials basis.
