Loading...

DiGiCOR | Digicor | NVIDIA Virtual GPU Solutions

Virtual GPU Solutions

SuperServers Optimized For NVIDIA® Tesla® GPUs

Most powerful solution for Deep Learning Training with the Fastest GPU to GPU Interconnections

PERFORMANCE: Highest Parallel peak performance with NVIDIA Tesla V100 GPUs
THROUGHPUT: Best in class GPU-to-GPU bandwidth with a maximum speed of 300GB/s
SCALABILITY: Designed for direct interconections between multiple GPU nodes
FLEXIBILITY: PCI-E 3.0 x16 for low latency I/O expansion capacity & GPU Direct RDMA support
DESIGN: Optimized GPU cooling for highest sustained parallel computing performance
EFFICIENCY: Redundant Titanium Level power supplies & intelligent cooling control


High Performance GPU Accelerated Virtual Desktop Infrastructure

Supermicro 1U and 2U Systems + NVIDIA GPUs

Ultra 1U 1029U VDI with High Graphic Performance

Technical Specifications
Processor Support
  • Dual Intel® Xeon® Scalable processors
  • Up to 165W CPU TDP with one or more GPUs
  • Memory Capacity
  • 24 DIMM slots, up to 6TB DDR4-2933 ECC RDIMM/LRDIMM
  • PCI-E Expansion Slots
  • 2 PCI-E Gen 3.0 x16 (2 full height 10.5” length)
  • 2 PCI-E Gen 3.0 x8 (1 low profile, 1 internal low profile)
  • Networking
  • TR4: 4x 1GbE ports
  • -TR4T: 4x 10Gbase-T ports
  • -TRT: 2x 10Gbase-T ports
  • -TRTP: 2x 10G SFP+ ports
  • -TRTP2: 2x 1GbE and 2x 10G SFP+ ports
  • -TR25M: 2x 25GbE ports
  • Drive Bays
  • 10 hot-swap 2.5” drives bays (Default 10 SATA3 ports)
  • Optional 2 NVMe/SATA3 Hybrid ports
  • Optional 8 SAS3 ports support via Add-on Card
  • Power Supply
  • Redundant 750W Platinum Level digital power supplies
  • VDIs
    Standard VDI
    Install 1x Tesla® M10
    32 Users with 1B Profile
    Entry-Mid End VDI
    Install 2x Tesla® T4
    8 ~ 4 Users with 2Q ~ 4Q Profile
    High End VDI
    Install 1x Tesla® V100
    4 Users with 8Q Profile

    Ultra 2U 2029U VDI with High Graphic Performance

    Technical Specifications
    Processor Support
  • Dual Intel® Xeon® Scalable processors
  • Up to 165W CPU TDP with one or more GPUs
  • Memory Capacity
  • 24 DIMM slots, up to 6TB DDR4-2933 ECC RDIMM/LRDIMM
  • PCI-E Expansion Slots
  • 8 PCI-E: 1 PCI-E 3.0 x16 (FH 10.5” L), 7 PCI-E 3.0 x8 (5 FH 10.5”L, 1 LP, 1 LP Internal)
  • -TR4T: 7 PCI-E: 2 PCI-E 3.0 x16 (FH 10.5” L), 5 PCI-E 3.0 x8 (4 FH 10.5”L, 1 LP)
  • Networking
  • 4x 1GbE ports *
  • 4x 10Gbase-T ports *
  • 2x 10Gbase-T ports *
  • 2x 10G SFP+ ports *
  • 2x 25GbE ports
  • Drive Bays
  • 24 hot-swap 2.5” drives bays (default 14 SATA3 ports)
  • Optional 24 SAS3 ports support via Add-on Card(s)
  • Optional 4 NVMe ports (2 NVMe from CPU1 + 2 NVMe from CPU2)
  • Optional 2 rear hot-swap 2.5” drive bay kit
  • Power Supply
  • Redundant 1000W Titanium Level digital power supplies
  • VDIs
    Standard VDI
    Install 2x Tesla® M10
    32 Users with 2TB Profile

    Entry-Mid End VDI
    Install 4x Tesla® T4
    32 ~ 16 Users with 2Q ~ 4Q Profile

    High End VDI
    Install 2x Tesla® V100
    8 ~ 4 Users with 8Q ~ 16Q Profile

    2U 2029U All-Flash NVMe with 2x NVIDIA® Tesla® GPUs VDI

    Technical Specifications

    Processor Support
  • Dual Intel® Xeon® Scalable processors
  • Up to 165W CPU TDP with one or more GPUs
  • Memory Capacity
  • 24 DIMM slots, up to 6TB DDR4-2933 ECC RDIMM/LRDIMM
  • PCI-E Expansion Slots
  • 2 PCI-E Gen 3.0 x16 (1 full height 10.5” length)
  • 1 PCI-E Gen 3.0 x8 (low profile
  • Networking
  • 4x 10GBase-T Ethernet ports
  • 24 hot-swap 2.5” drives bays
  • 12 NVMe ports (NVMe from CPU1)
  • 8 NVMe ports + 4 NVMe/SATA3 Hybrid ports (NVMe from CPU2)
  • 2 rear 2.5” hot-swap bays (default 2 SATA3 ports)
  • Power Supply
  • Redundant 1600W Titanium Level digital power supplies
  • VDIs
    Standard VDI
    Install 2x Tesla® M10
    64 Users with 1B Profile

    Entry-Mid End VDI
    Install 4x Tesla® T4
    32 ~ 16 Users with 2Q ~ 4Q Profile

    High End VDI
    Install 2x Tesla® V100
    8 ~ 4 Users with 8Q ~ 16Q Profile

    Notes

  • The number of users supported varies depending on specific applications and workloads.
  • Q-series virtual GPU types require a Quadro vDWS license.
  • B-series virtual GPU types require a GRID Virtual PC license.
  • For more information about NVIDIA vGPU technology, see
  • Share