Persönliche Beratung

+49 421 40 966 0 oder per Mail

Produkt Filter Produkt Filter

Deep Learning / Artificial Intelligence

Deep Learning ist eines der am schnellsten wachsenden Gebiete der Computer-Branche und wohl das schnellste im Bereich Machine Learning.


Bei Deep Learning werden viel-schichtige neuronale Netzwerke (Deep Neural Networks - DNNs) genutzt, um ein hohes Maß an Abstraktion und Darstellung auf Basis von Daten wie z.B. Bilder, Tönen oder Text zu erlernen. Im Gegensatz zum klassischen Machine Learning, wo jede Merkmals- und Mustererkennung noch per Hand geschrieben werden mussten, nutzen bei Deep Learning neuronale Netze Algorithmen und große Datenmengen(Big Data) um sich diese Erkennung quasi selbst beizubringen. Die Einsatzgebiete Reichen von der Wissenschaft wie Biologie oder Physik, über Anwendungen in der Industrie und dem Handel, bis hin zum selbstfahrenden Auto.

 

devCUBE

Sie sind neu auf dem Gebiet des Deep Learnings, oder haben schlicht nicht den Platz für ein Server-System? Dann ist unser devCUBE als Deskside-Solution genau das Richtige. Bis zu vier 4 NVIDIA® GeForce® GTX Titan X GPUs (insgesamt 12.288 CUDA® Kerne), hohe Speicherdichte und einen modernen CORE™ i7 Prozessor in einem einzigen Gehäuse, das bequem unter Ihren Schreibtisch passt, und trotzdem kühl und leise bleibt kombiniert mit der bewährten sysGen-Qualität - der devCUBE.

 

Auch NVIDIA sieht das riesige Potential des Deep Learnings und bietet mit der DGX-1™ nun den ersten komplett auf Deep Learning optimierten Supercomputer auf Basis der neuen Pascal™ Architektur an.


DGX-1

8 GPUs der neuen Tesla® P100 Technologie liefern weit über 28.000 CUDA® Kerne, 128GB HBM2 Speicher und damit beeindruckende 170 TFLOPS(FP16). Zusätzlich sind in dem 3U Gehäuse 512GB DDR4 Speicher, 4x 1.92TB SSD Speicher im Raid0 und zwei Intel® Xeon® CPUs verbaut. Diese Power kann man mit ca. 250 herkömmlichen CPU-Servern gleichsetzen, und bleibt mit etwa 3200W dann doch vergleichsweise sparsam. Als NVIDIA® Preferred Solution Provider sind wir froh, Ihnen diesen Supercomputer anbieten zu können.

 
 

Natürlich ist Rechenpower alleine nicht alles. So sind die von uns angebotenen Deep Learning Solutions mit einem umfangreichen Softwarepaket ausgestattet. Grundlegende Software wie NVIDIA® DIGITS, cuDNN, NCCL und ein GPU optimiertes Linux sind genauso an Bord wie auch viele der wichtigsten Frameworks wie etwa Caffe, Torch, Theano und BIDMach.

DGX1 Frameworks
 

Neben dem bekannt vielseitigen Support durch unsere sysGen-Mitarbeitet ist im Rahmen der DGX-1™ auch ein umfangreiches Service und Support Paket von NVIDIA® möglich, sowie auf die DGX-1™ optimierte Software-Lösungen.




 

Sie können sich nicht entscheiden, was genau Sie brauchen?

Kontaktieren Sie uns gerne.


 

 

 

 

€ 3.067,09 (€ 3.649,83 inkl. MwSt.)

Kurzbeschreibung:

  1. Dual socket R3 (LGA 2011) supports Intel® Xeon® processor E5-2600 v3/v4 family; QPI up to 9.6GT/s
  2. Up to 3TB ECC, up to DDR4 2400MHz; 24x DIMM slots
  3. Expansion slots: 8 PCI-E 3.0 x16 (double-width) slots 2 PCI-E 3.0 x8 (in x16) slots 1 PCI-E 2.0 x4 (in x16) slot
  4. Dual 10GBase-T LAN with Intel® X540
  5. 24x 2.5" Hot-swap drive bays
  6. 8x 92mm RPM Hot-Swap cooling fans
  7. 1600W Redundant (2+2) Power Supplies;Platinum Level (94%+)
zum Hersteller-Datenblatt
 

€ 2.503,58 (€ 2.979,26 inkl. MwSt.)

Kurzbeschreibung:

  • Complete System, completely configured, incl. all required Software, Ready to Go
  • NVIDIA DIGITS software providing powerful design, training, and visualization of deep neural networks for image classification

System Expansion:
  1. Desk-Side Workstation enclosure with 1500W power supply in Titanium Quality
  2. Mainboard X99-E WS with 4x PCIe x16
  3. Intel Core i7 Broadwell-E CPU
  4. 64GB DDR4 Memory (8x 16GB DDR4 2400 MHz UDIMM, 288-Pin)
  5. Up to four NVIDIA Pascal Titan X or GeForce GTX 1080 GPUs
  6. NVIDIA DIGITS software: Interactive Deep Learning GPU Training System
  7. Pre-installed Ubuntu 14.04. with Caffe, Torch, Theano, BIDMach, cuDNN and CUDA *
  8. Storage-bays:
    1. Internal (swap-Frame): 3x 3.5" HDD and 6x 2.5" HDD / SSD drives
    2. External (hot hwap): 8x 2.5" HDD / SSD drives (pre selected) or 3x 3.5" HDD drives

*Pre-requisite: Before sysGen can install cuDNN v2, you have to successfully register on the NVIDIA website as a system developer for "Deep Learning" and sign the NVIDIA "CUDA Software Distribution Agreement for Toolkit cuDNN".
 

€ 2.980,35 (€ 3.546,61 inkl. MwSt.)
Kurzbeschreibung:
  1. Dual socket R3 (LGA 2011) supports Intel® Xeon® processor E5-2600 v4; /v3 family; QPI up to 9.6GT/s
  2. Up to3TB ECC 3DS LRDIMM, up to DDR4- 2400 MHz; 24x DIMM slots
  3. Expansion slots: 8 PCI-E 3.0 x16 (double-width) slots 2 PCI-E 3.0 x8 (in x16) slots 1 PCI-E 2.0 x4 (in x16) slot
  4. Dual GbE LAN with Intel i350
  5. 24x 2.5" Hot-swap drive bays
  6. 8x 92mm RPM Hot-Swap cooling fans
  7. 1600W Redundant (2+2) Power Supplies;Platinum Level (94%+)
zum Hersteller-Datenblatt
 

€ 3.863,34 (€ 4.597,37 inkl. MwSt.)

sysGen/SUPERMICRO Server SYS-4028GR-TRT2, bestehend aus:
Mainboard MBD-X10DRG-OT+ / 4U Rackmount CSE-418GTS-R3200B

  • Up to 10 Double-Width GPUs
  • GPUs under Single-Root complex
  • Supports RDMA
  1. Dual socket R3 (LGA 2011) supports Intel® Xeon® processor E5-2600 v4 family; QPI up to 9.6GT/s
  2. Up to 3TB ECC, up to DDR4 2400MHz; 24x DIMM slots
  3. Expansion slots: 11 PCI-E 3.0 x16 (FH,FL), 1 PCI-E 3.0 x8 (in x16)
  4. Dual 10GBase-T LAN with Intel® X540
  5. 24x 2.5" Hot-swap drive bays
  6. 8x 92mm RPM Hot-Swap cooling fans
  7. 2000W Redundant (2+2) Power Supplies;Titanium Level (96%+)
zum Hersteller-Datenblatt

 

€ 3.775,63 (€ 4.493,00 inkl. MwSt.)

sysGen/SUPERMICRO Server SYS-4028GR-TR2 , bestehend aus:
Mainboard MBD-X10DRG-O+ / 4U Rackmount CSE-418GTS-R3200B

  • Up to 10 Double-Width GPUs
  • GPUs under Single-Root complex
  • Supports RDMA
  1. Dual socket R3 (LGA 2011) supports Intel® Xeon® processor E5-2600 v4 family; QPI up to 9.6GT/s
  2. Up to 3TB ECC, up to DDR4 2400MHz; 24x DIMM slots
  3. Expansion slots: 11 PCI-E 3.0 x16 (FH,FL), 1 PCI-E 3.0 x8 (in x16)
  4. Dual GbE LAN with Intel® i350
  5. 24x 2.5" Hot-swap drive bays
  6. 8x 92mm RPM Hot-Swap cooling fans
  7. 2000W Redundant (2+2) Power Supplies;Titanium Level (96%+)
zum Hersteller-Datenblatt

 

€ 5.896,37 (€ 7.016,68 inkl. MwSt.)

Kurzbeschreibung:

  • Up to 8 Tesla P100 SXM2
  • Up to 80 GB/s GPU-to-GPU NVLINK
  • Supports GPUDirect RDMA
  1. Dual socket R3 (LGA 2011) supports Intel® Xeon® processor E5-2600 v4† /v3 family; QPI up to 9.6GT/s
  2. Up to3TB†ECC 3DS LRDIMM, up to DDR4- 2400† MHz; 24x DIMM slots
  3. 4 PCI-E 3.0 x16 (double-width) slots 2 PCI-E 3.0 x8 slots
  4. Dual 10GBase-T LAN with Intel® X540
  5. 16 Hot-swap 2.5" SATA/SAS drives; Supports 8 NVMe drives
  6. 8x 92mm cooling fans
  7. 2000W Redundant (2+2) Power Supplies;Titanium Level (96%+)
zum Hersteller-Datenblatt

 

€ 5.815,48 (€ 6.920,42 inkl. MwSt.)

sysGen/SUPERMICRO Server SYS-4028GR-TXR, bestehend aus:
Mainboard MBD-X10DGO / 4U Rackmount CSE-R422BG

  • Up to 8 Tesla P100 SXM2
  • Up to 80 GB/s GPU-to-GPU NVLINK
  • Supports GPU Direct RDMA
  1. Dual socket R3 (LGA 2011) supports Intel® Xeon® processor E5-2600 v4† /v3 family; QPI up to 9.6GT/s
  2. Up to3TB†ECC 3DS LRDIMM, up to DDR4- 2400† MHz; 24x DIMM slots
  3. 4 PCI-E 3.0 x16 (double-width) slots 2 PCI-E 3.0 x8 slots
  4. Dual 1GbE LAN with Intel® i350
  5. 16 Hot-swap 2.5" SATA/SAS drives; Supports 8 NVMe drives
  6. 8x 92mm cooling fans
  7. 2000W Redundant (2+2) Power Supplies;Titanium Level (96%+)
zum Hersteller-Datenblatt

 

€ 2.383,76 (€ 2.836,68 inkl. MwSt.)
Z10-D8 WS Mainboard
  • Complete watercooling System - silent high-end performance
  • Ready to Go System - completely configured, incl. all required Software
  • NVIDIA DIGITS software providing powerful design, training, and visualization of deep neural networks for image classification

  1. Up to 4x NVIDIA GTX 1080 with a total of 32GB GDDR5X Memory and 10240 CUDA Cores
  2. Dual Intel Xeon E5-2687W v4 CPUs with up to 3.5 GHz each
  3. Highest level of silent computing with a system wide, low maintenance water cooling system
  4. Pre-Installed and Pre-configured software package for deep learning
  5. Pre-installed Ubuntu with Caffe, Torch, Theano, BIDMach, cuDNN and CUDA
    Pre-requisite: Before sysGen can install cuDNN , you have to successfully register on the NVIDIA website as a system developer for "Deep Learning" and sign the NVIDIA "CUDA Software Distribution Agreement for Toolkit cuDNN".
  6. Up to 512GB ECC DDR4 LR-DIMM / RegDIMM Memory
  7. Space for 6x SSDs / HDDs
 

€ 2.383,76 (€ 2.836,68 inkl. MwSt.)
Z10-D8 WS Mainboard

sysGen HPC/DL Power Workstation SYS-devCUBE-H₂O-B1

High end computing (HPC) and deep learning (DL) power with up to 4x water cooled NVIDIA® GPUs and 2 Intel® Xeon® CPUs, Cool and Quiet - Completely water-cooled.

The devCUBE is one of the best deskside solution for HPC and AI/DL and we are proud to present its even cooler powerfull brothers: The devCUBE H2O group – all sysGen Originals


  • Complete watercooling System - silent high-end performance
  • Ready to Go System - completely configured, incl. all required Software
  • NVIDIA DIGITS software providing powerful design, training, and visualization of deep neural networks for image classification

  1. Up to 4x NVIDIA GTX 1080ti or Tesla Pascal GPUs
  2. Dual Intel Xeon E5-2687W v4 CPUs with up to 3.5 GHz each
  3. Highest level of silent computing with a system wide, low maintenance water cooling system
  4. Pre-Installed and Pre-configured software package for deep learning
  5. Pre-installed Ubuntu with Caffe, Torch, Theano, BIDMach, cuDNN and CUDA
    Pre-requisite: Before sysGen can install cuDNN , you have to successfully register on the NVIDIA website as a system developer for "Deep Learning" and sign the NVIDIA "CUDA Software Distribution Agreement for Toolkit cuDNN".
  6. Up to 512GB ECC DDR4 LR-DIMM / RegDIMM Memory
  7. Space for 6x SSDs / HDDs

Datasheet/Datenblatt / Overview Deep-Learning
 

€ 2.560,35 (€ 3.046,83 inkl. MwSt.)

Kurzbeschreibung:

  • Supports up to 8x double slot GPU/MIC cards
  • Supporting active and passive Computing Cards, including P100
  • Highest Computing Density in its Category
  • GPU Server, Mission-critical application, enterprise server, oil & gas, financial, 3D rendering, chemistry, HPC, AI - Deep Learning
  1. Dual socket R (LGA 2011-3) Intel® Xeon® processor E5-2600 v3/v4 product family
  2. Up to 3TB ECC 3DS LRDIMM, up to DDR4-2400MHz; 24x DIMM slots
  3. Supports up to 8x double slot GPU/MIC cards (NVIDIA Tesla, AMD Fire Pro, Intel Xeon Phi)
  4. 2x GbE LAN ports (Intel® I350-AM2)
  5. 10x SATA III 6Gb/s ports
  6. 8x Hot-swap 2.5" SATA Drive Bays
  7. 80 PLUS Platinum 2000W redundant PSU
  8. Aspeed® AST2400 management controller
  9. Avocent® MergePoint IPMI 2.0 web interface

zum Hersteller-Datenblatt

 

€ 2.622,50 (€ 3.120,77 inkl. MwSt.)

Kurzbeschreibung:

  • Supports up to 8x double slot GPU/MIC cards
  • Supporting active and passive Computing Cards, including P100
  • Highest Computing Density in its Category
  • GPU Server, Mission-critical application, enterprise server, oil & gas, financial, 3D rendering, chemistry, HPC, AI - Deep Learning
  1. Dual socket R (LGA 2011-3) Intel® Xeon® processor E5-2600 v3/v4 product family
  2. Up to 3TB ECC 3DS LRDIMM, up to DDR4-2400MHz; 24x DIMM slots
  3. Supports up to 8x double slot GPU/MIC cards (NVIDIA Tesla, AMD Fire Pro, Intel Xeon Phi)
  4. 2x 10GbE SFP+ ports (Intel® 82599ES)
  5. 10x SATA III 6Gb/s ports
  6. 8x Hot-swap 2.5" SATA Drive Bays
  7. 80 PLUS Platinum 2000W redundant PSU
  8. Aspeed® AST2400 management controller
  9. Avocent® MergePoint IPMI 2.0 web interface

zum Hersteller-Datenblatt

 

€ 2.985,22 (€ 3.552,41 inkl. MwSt.)

Kurzbeschreibung:

  • Up to 4x Tesla P100 SXM2
  • Up to 80 GB/s GPU-to-GPU NVLINK
  • Supports GPUDirect RDMA
  1. Dual socket R3 (LGA 2011) supports Intel® Xeon® processor E5-2600 v4/v3 family; QPI up to 9.6GT/s
  2. Up to 2TB ECC 3DS LRDIMM, up to DDR4-2400MHz; 16x DIMM slots
  3. 3x PCI-E 3.0 x16 slots, 1 PCI-E 3.0 x8 Low-profile slot
  4. 2x 10GBase-T LAN via Intel X540
  5. 2x 2.5" Hot-swap drive bays, 2x 2.5" internal drive bays
  6. 8x 4cm heavy duty counter-rotating fans with air shroud & optimal fan speed control
  7. 2000W Redundant Power Supplies Titanium Level (96%)
zum Hersteller-Datenblatt

 

€ 2.903,30 (€ 3.454,93 inkl. MwSt.)

Kurzbeschreibung:

  • Up to 4x Tesla P100 SXM2 (NVLink)
  • Up to 80 GB/s GPU-to-GPU NVLINK
  • Supports GPUDirect RDMA
  1. Dual socket R3 (LGA 2011) supports Intel® Xeon® processor E5-2600 v4/v3 family; QPI up to 9.6GT/s
  2. Up to 2TB ECC 3DS LRDIMM, up to DDR4-2400MHz; 16x DIMM slots
  3. 3x PCI-E 3.0 x16 slots, 1x PCI-E 3.0 x8 Low-profile slot
  4. Intel i350 Dual port GbE LAN
  5. 2x 2.5" Hot-swap drive bays, 2x 2.5" internal drive bays
  6. 8x 4cm heavy duty counter-rotating fans with air shroud & optimal fan speed control
  7. 2000W Redundant Power Supplies Titanium Level (96%)
zum Hersteller-Datenblatt

 

€ 1.946,16 (€ 2.315,93 inkl. MwSt.)
Kurzbeschreibung:

  • Newest 1U HPC Server Supporting 4 Computing Cards
  • Supporting active and passive Computing Cards, including P100 and TITAN X
  • Highest Computing Density in its Category
  • GPU Server, Mission-critical application, enterprise server, oil & gas, financial, 3D rendering, chemistry, HPC, AI - Deep Learning
  1. Dual socket R3 (LGA 2011) supports Intel® Xeon® processor E5-2600 v3/v4 family; QPI up to 9.6GT/s
  2. Up to 2TB ECC, up to DDR4 2400MHz; 16x DIMM slots
  3. 4x PCI-E 3.0 x16 (4 GPU cards opt.) and 2x PCI-E 3.0 x8 (in x16) LP slots
  4. Dual port 10GbE LAN
  5. System management: Built-in Server management tool (IPMI 2.0, KVM/media over LAN) with dedicated LAN port
  6. 2x 2.5" Hot-swap drive bays, 2x 2.5" internal drive bays
  7. 9x 4cm heavy duty counter-rotating fans w/ optimal fan speed control
  8. 2000W Redundant Power Supplies Platinum Level (94%+)
zum Hersteller-Datenblatt

 

€ 1.866,79 (€ 2.221,49 inkl. MwSt.)

Kurzbeschreibung:

  1. Dual socket R3 (LGA 2011) supports Intel® Xeon® processor E5-2600 v3/v4 family; QPI up to 9.6GT/s
  2. Up to 512GB ECC 3DS LRDIMM, up to DDR4-2400MHz; 16x DIMM slots
  3. 4x PCI-E 3.0 x16 slots (4x GPU/Xeon Phi cards opt.), 2x PCI-E 3.0 x8 (in x16) LP slot
  4. Intel i350 Dual port GbE LAN
  5. 2x 2.5" Hot-swap drive bays, 2x 2.5" internal drive bays
  6. 9x 4cm heavy duty counter-rotating fans with air shroud & optimal fan speed control
  7. 2000W Redundant Power Supplies Platinum Level (94%+)

zum Hersteller-Datenblatt

 

€ 120.943,83 (€ 143.923,16 inkl. MwSt.)

The World's First Deep Learning Supercomputer in a box


The NVIDIA® DGX-1™ is the world’s first purpose-built system for deep learning with fully integrated hardware and software that can be deployed quickly and easily. Its revolutionary performance significantly accelerates training time, making the NVIDIA DGX-1 the world’s first deep learning supercomputer in a box.


Key Features

  1. 170 TFLOPS(FP16)
  2. 8x Tesla GP100
  3. 16GB Memory per GPU
  4. 28672 NVIDIA CUDA® Cores
  5. Dual Intel® Xeon® E5-2698 v4
  6. 512GB 2133MHz DDR4 LRDIMM System Memory
  7. 4x 1.92TB SSD in RAID 0
  8. 2x 10GbE, 4x Infiniband EDR
  9. 3200W
  10. incl. 1 Year NVIDIA Support Service

Download the Datasheet here

+

€ 147.540,47 (€ 175.573,16 inkl. MwSt.)

The World's First Deep Learning Supercomputer in a box


The NVIDIA® DGX-1™ is the world’s first purpose-built system for deep learning with fully integrated hardware and software that can be deployed quickly and easily. Its revolutionary performance significantly accelerates training time, making the NVIDIA DGX-1 the world’s first deep learning supercomputer in a box.


Key Features

  1. 170 TFLOPS(FP16)
  2. 8x Tesla GP100
  3. 16GB Memory per GPU
  4. 28672 NVIDIA CUDA® Cores
  5. Dual Intel® Xeon® E5-2698 v4
  6. 512GB 2133MHz DDR4 LRDIMM System Memory
  7. 4x 1.92TB SSD in RAID 0
  8. 2x 10GbE, 4x Infiniband EDR
  9. 3200W
  10. incl. 3 Years NVIDIA Support Service

Download the Datasheet here

+

€ 132.058,86 (€ 157.150,04 inkl. MwSt.)

The World's First Deep Learning Supercomputer in a box


The NVIDIA® DGX-1™ is the world’s first purpose-built system for deep learning with fully integrated hardware and software that can be deployed quickly and easily. Its revolutionary performance significantly accelerates training time, making the NVIDIA DGX-1 the world’s first deep learning supercomputer in a box.


Key Features

  1. 170 TFLOPS(FP16)
  2. 8x Tesla GP100
  3. 16GB Memory per GPU
  4. 28672 NVIDIA CUDA® Cores
  5. Dual Intel® Xeon® E5-2698 v4
  6. 512GB 2133MHz DDR4 LRDIMM System Memory
  7. 4x 1.92TB SSD in RAID 0
  8. 2x 10GbE, 4x Infiniband EDR
  9. 3200W
  10. incl. 1 Year NVIDIA Support Service

Download the Datasheet here

+

€ 163.932,77 (€ 195.080,00 inkl. MwSt.)

The World's First Deep Learning Supercomputer in a box


The NVIDIA® DGX-1™ is the world’s first purpose-built system for deep learning with fully integrated hardware and software that can be deployed quickly and easily. Its revolutionary performance significantly accelerates training time, making the NVIDIA DGX-1 the world’s first deep learning supercomputer in a box.


Key Features

  1. 170 TFLOPS(FP16)
  2. 8x Tesla GP100
  3. 16GB Memory per GPU
  4. 28672 NVIDIA CUDA® Cores
  5. Dual Intel® Xeon® E5-2698 v4
  6. 512GB 2133MHz DDR4 LRDIMM System Memory
  7. 4x 1.92TB SSD in RAID 0
  8. 2x 10GbE, 4x Infiniband EDR
  9. 3200W
  10. incl. 3 Years NVIDIA Support Service

Download the Datasheet here

+