581 Search Results for "NVIDIA ConnectX-6"
-
Vendor:Cisco
Cisco UCSB-NVMEM6-W3800 | 2.5in NVMe SSD | 3.8TB | PCIe NVMe | M6-optimized
UCSB-NVMEM6-W3800
Stock Status: Enquire
Vendor:CiscoCisco UCSB-NVMEM6-W3800 | 2.5in NVMe SSD | 3.8TB | PCIe NVMe | M6-optimized
High-capacity enterprise NVMe SSD for UCS platforms, delivering strong throughput and low latency for mixed workloads and consolidated virtualization hosts. Designed to integrate with M6-generation NVMe backplanes. Requires appropriate UCS tray/caddy. -
Vendor:Cisco
Cisco WS-X6524-100FX-MM | 24x 100BASE-FX MM MT-RJ | CEF256 fabric
WS-X6524-100FX-MM
Stock Status: Enquire
Vendor:CiscoCisco WS-X6524-100FX-MM | 24x 100BASE-FX MM MT-RJ | CEF256 fabric
24-port 100BASE-FX multimode fiber line card with MT-RJ connectors for Catalyst 6500. Fabric-enabled (CEF256) for high-performance campus aggregation of legacy FE fiber links. Fixed MT-RJ interfaces—no SFPs required. -
Vendor:Cisco
Cisco WS-X6816-10T-2T2TXL | 16x 10GBASE-T (RJ-45, 100 m) | DFC4/DFC4XL | 2T/2TXL fabric
WS-X6816-10T-2T2TXL
Stock Status: Enquire
Vendor:CiscoCisco WS-X6816-10T-2T2TXL | 16x 10GBASE-T (RJ-45, 100 m) | DFC4/DFC4XL | 2T/2TXL fabric
High-density 16-port 10GBASE-T copper line card delivering 10GbE over RJ-45 with integrated DFC4 or DFC4XL for distributed forwarding. Designed for 6800 Series and compatible 6500-E chassis operating with Supervisor 2T fabric for campus core/distribution upgrades using existing copper cabling. -
Vendor:Cisco
Cisco UCSX-NVB6T4O1PM6 | 6.4TB 2.5in U.2 NVMe, PCIe 4.0 x4, 15mm, data center SSD, high endurance (~3 DWPD)
UCSX-NVB6T4O1PM6
Stock Status: Enquire
Vendor:CiscoCisco UCSX-NVB6T4O1PM6 | 6.4TB 2.5in U.2 NVMe, PCIe 4.0 x4, 15mm, data center SSD, high endurance (~3 DWPD)
6.4 TB enterprise NVMe SSD in a 2.5-inch U.2, 15 mm form factor for PCIe 4.0 x4 backplanes. Delivers steady latency and throughput for mixed workloads, with power-loss protection and end-to-end data integrity to support demanding UCS application stacks. -
Vendor:NVIDIA
NVIDIA 980-9IA0Q-00N01A | 800Gb/s twin-port OSFP DAC 1.5m, InfiniBand NDR, OSFP-OSFP
980-9IA0Q-00N01A
Stock Status: Enquire
Vendor:NVIDIANVIDIA 980-9IA0Q-00N01A | 800Gb/s twin-port OSFP DAC 1.5m, InfiniBand NDR, OSFP-OSFP
Short-reach twin-port passive copper DAC for InfiniBand NDR fabrics, providing up to 800 Gb/s aggregate bandwidth (2x 400G) with minimal latency. Optimized for dense HPC and AI clusters where cable routing and airflow matter. -
Vendor:Juniper
Juniper QFX10000-60S-6Q | 60x1/10G SFP/SFP+ + 6x40G QSFP+ + 2x100G QSFP28 line card | for QFX10000 chassis | PoE: None | Uplinks: 2x100G
QFX10000-60S-6Q
Stock Status: Enquire
Vendor:JuniperJuniper QFX10000-60S-6Q | 60x1/10G SFP/SFP+ + 6x40G QSFP+ + 2x100G QSFP28 line card | for QFX10000 chassis | PoE: None | Uplinks: 2x100G
Flexible-density modular line card offering 60 x 1/10GbE SFP/SFP+, 6 x 40GbE QSFP+, and 2 x 100GbE QSFP28 interfaces for QFX10000 modular chassis. Ideal for leaf aggregation with 100G uplinks and mixed 10/40/100G access. -
Vendor:NVIDIA
NVIDIA 900-82888-0050-000 | Jetson AGX Xavier 64GB AI Compute Module, 32 TOPS, Volta GPU
900-82888-0050-000
Stock Status: Enquire
Vendor:NVIDIANVIDIA 900-82888-0050-000 | Jetson AGX Xavier 64GB AI Compute Module, 32 TOPS, Volta GPU
The NVIDIA Jetson AGX Xavier 64GB module is a high-performance system-on-module designed for edge AI and autonomous machine applications. Delivering up to 32 TOPS of AI performance while consuming as little as 10W, this compact module features a 512-core Volta GPU with 64 Tensor cores, 8-core Carmel ARMv8.2 CPU, 64GB LPDDR4x memory, and 32GB eMMC storage. Ideal for robotics, industrial automation, smart cities, and AI-powered applications requiring substantial memory for multiple neural network models. -
Vendor:Cisco
Cisco CAI-GPU-H200-NVL | H200 NVL Graphics Card, 141GB HBM3e, 600W, PCIe 5.0
CAI-GPU-H200-NVL
Stock Status: Enquire
Vendor:CiscoCisco CAI-GPU-H200-NVL | H200 NVL Graphics Card, 141GB HBM3e, 600W, PCIe 5.0
The Cisco CAI-GPU-H200-NVL is an NVIDIA H200 NVL Tensor Core GPU built on the Hopper architecture, delivering exceptional performance for AI training, LLM inference, and high-performance computing workloads. Featuring 141GB of HBM3e memory with 4.8TB/s bandwidth and 16,896 CUDA cores, this PCIe 5.0 add-in card is optimized for enterprise data centers requiring powerful AI acceleration in air-cooled rack environments. With NVLink support for 2-way or 4-way GPU interconnect and multi-instance GPU capabilities, it provides scalable performance for generative AI, deep learning, and scientific computing applications. -
Vendor:Cisco
Cisco HX-GPU-V100-32 | Tesla V100 GPU, 32GB HBM2, 250W, PCIe Gen3 x16
HX-GPU-V100-32
Stock Status: Enquire
Vendor:CiscoCisco HX-GPU-V100-32 | Tesla V100 GPU, 32GB HBM2, 250W, PCIe Gen3 x16
The Cisco HX-GPU-V100-32 is an NVIDIA Tesla V100 GPU accelerator designed for AI, deep learning, and high-performance computing workloads in data center environments. Featuring 32GB of HBM2 memory, 5,120 CUDA cores, and 640 Tensor Cores, this PCIe Gen3 x16 card delivers up to 125 TFLOPS of deep learning performance. The passively-cooled design with 250W TDP makes it ideal for deployment in Cisco UCS and HyperFlex servers running compute-intensive AI training, inference, and scientific simulation applications. -
Vendor:NVIDIA
NVIDIA 980-9I80P-00N003 | Optics: 800G→2x400G, OSFP→2xQSFP112, 800GBASE-CR8/400GBASE-CR4, passive copper DAC, 3 m, connectors: OSFP to 2x QSFP112
980-9I80P-00N003
Stock Status: Enquire
Vendor:NVIDIANVIDIA 980-9I80P-00N003 | Optics: 800G→2x400G, OSFP→2xQSFP112, 800GBASE-CR8/400GBASE-CR4, passive copper DAC, 3 m, connectors: OSFP to 2x QSFP112
Passive copper splitter cable for breaking out one 800G OSFP port into two independent 400G QSFP112 links. Suited for leaf/spine aggregation and GPU/accelerated computing clusters that need zero‑power, ultra‑low‑latency connectivity over 3 meters across Ethernet or InfiniBand NDR‑class networks. -
Vendor:NVIDIA
NVIDIA 980-9IAM3-00X02A | MCA4K50 | 1600G OSFP Active Copper Cable | Reach 2.5m | 1.6T Ethernet | Connector: OSFP‑to‑OSFP
980-9IAM3-00X02A
Stock Status: Enquire
Vendor:NVIDIANVIDIA 980-9IAM3-00X02A | MCA4K50 | 1600G OSFP Active Copper Cable | Reach 2.5m | 1.6T Ethernet | Connector: OSFP‑to‑OSFP
1.6T OSFP active copper assembly with integrated heat sinks on both ends for improved thermal performance. The 2.5m reach provides flexible short‑distance wiring while preserving the ultra‑low latency and power advantages of copper for high‑bandwidth fabrics. -
Vendor:Cisco
Cisco C885L-P-N7S400GFO | 1x400GbE QSFP112 | PCIe Gen5 x16 | ConnectX‑7 class features: RoCEv2, RDMA, NVMe/TCP, PTP
C885L-P-N7S400GFO
Stock Status: Enquire
Vendor:CiscoCisco C885L-P-N7S400GFO | 1x400GbE QSFP112 | PCIe Gen5 x16 | ConnectX‑7 class features: RoCEv2, RDMA, NVMe/TCP, PTP
Single‑port 400GbE PCIe Gen5 adapter with QSFP112 cage for ultra‑low‑latency fabrics and storage offloads. Ideal for AI/ML clusters, hyper‑converged nodes, and high‑throughput data pipelines in modern PCIe servers and UCS C885L M8 configurations. -
Vendor:Cisco
Cisco 88-LC0-36FH | 36x400G, QSFP56-DD, chassis line card
88-LC0-36FH
Stock Status: Enquire
Vendor:CiscoCisco 88-LC0-36FH | 36x400G, QSFP56-DD, chassis line card
Ultra‑dense 400G line card for 8800-series modular chassis, providing 36x400G QSFP56‑DD ports for high-capacity spine and core aggregation. Supports 4x100G breakout on qualified optics to ease migration from 100G to 400G. Ideal for large-scale fabrics requiring maximum throughput per slot with operational simplicity. -
Vendor:Cisco
Cisco E-SSD-U2N-2TB | U.2 NVMe SSD 2TB 2.5in | PCIe NVMe | for UCS-E M6
E-SSD-U2N-2TB
Stock Status: Enquire
Vendor:CiscoCisco E-SSD-U2N-2TB | U.2 NVMe SSD 2TB 2.5in | PCIe NVMe | for UCS-E M6
High-performance U.2 NVMe 2.5-inch SSD for UCS-E M6 compute modules, delivering low-latency storage ideal for high IOPS workloads, caching, and primary application volumes via PCIe NVMe. -
Vendor:Cisco
Cisco HCI-P-V5Q50G-M6 | 4x10/25/50GbE PCIe CNA | C-Series (M6) | Secure Boot | SR-IOV | Flexible Port Speeds
HCI-P-V5Q50G-M6
Stock Status: Enquire
Vendor:CiscoCisco HCI-P-V5Q50G-M6 | 4x10/25/50GbE PCIe CNA | C-Series (M6) | Secure Boot | SR-IOV | Flexible Port Speeds
4-port converged network adapter engineered for UCS C-Series (M6 generation) with flexible 10/25/50GbE per-port speeds and secure boot. Ideal for mixed-speed environments and virtualization-heavy deployments. -
Vendor:Cisco
Cisco HCIX-440P | PCIe expansion node | GPU/accelerator ready | connects to modular compute nodes | requires 8-slot chassis
HCIX-440P
Stock Status: Enquire
Vendor:CiscoCisco HCIX-440P | PCIe expansion node | GPU/accelerator ready | connects to modular compute nodes | requires 8-slot chassis
High-density PCIe expansion node designed to add GPUs, accelerators, and high-speed adapters to modular compute environments. It integrates through the chassis fabric to extend I/O capability without adding standalone servers. -
Vendor:Cisco
Cisco HX-GPU-A100-80 | A100 80GB HBM2e Graphics Card, 6912 CUDA, 300W
HX-GPU-A100-80
Stock Status: Enquire
Vendor:CiscoCisco HX-GPU-A100-80 | A100 80GB HBM2e Graphics Card, 6912 CUDA, 300W
The Cisco HX-GPU-A100-80 is a PCIe Gen4 compute accelerator based on the NVIDIA A100 Tensor Core GPU with 80GB HBM2e memory and 300W TDP. Designed for AI training, HPC workloads, and large-scale data analytics in Cisco UCS server environments, it features passive cooling and double-width form factor. This GPU delivers exceptional performance for deep learning, scientific computing, and inference tasks requiring massive memory capacity. -
Vendor:Cisco
Cisco NCS1K14-2.4T-X-K9 | 2.4Tbps line card | 24x DWDM ports (pluggable coherent) | X-mode cross-connect | ITU-T G.694.1 C-band, tunable | Long-haul/metro reach (design-dependent) | LC optical interface (pluggable-dependent)
2.4 Tbps high-capacity line card with enhanced cross-connect modes, enabling flexible, colorless service grooming and ROADM integration in the 1014 chassis. Provides up to 24 coherent DWDM interfaces (as enabled by pluggables), with tunable C-band operation for metro, DCI, and long-haul builds. X-mode cross-connect capabilities simplify any-to-any service mapping within the shelf. -
Vendor:Cisco
Cisco UCS-FI-6664-U | 2RU fabric interconnect, up to 64 unified ports: 10/25/40/100GbE + 8/16/32/64G FC; uplinks: 100GbE QSFP28; license: UCS Manager
UCS-FI-6664-U
Stock Status: Enquire
Vendor:CiscoCisco UCS-FI-6664-U | 2RU fabric interconnect, up to 64 unified ports: 10/25/40/100GbE + 8/16/32/64G FC; uplinks: 100GbE QSFP28; license: UCS Manager
High-density, 2RU unified fabric interconnect for UCS domains, combining 100GbE switching with up to 64G Fibre Channel on unified ports. This UCS Manager SKU enables on-premises policy control, service profiles, and lifecycle automation. Deploy as a redundant pair for continuous, non-disruptive operations across blade and rack compute. -
Vendor:Cisco
Cisco UCS-M6-CPU-CAR | M6 CPU carrier | Intel Xeon Scalable (Ice Lake) | LGA4189 | tool-less install
UCS-M6-CPU-CAR
Stock Status: Enquire
Vendor:CiscoCisco UCS-M6-CPU-CAR | M6 CPU carrier | Intel Xeon Scalable (Ice Lake) | LGA4189 | tool-less install
Replacement CPU retention carrier for M6-generation servers using 3rd Gen Intel Xeon Scalable (Ice Lake) processors. Ensures correct CPU alignment, even pressure with the heatsink, and protects socket pins during installation. Compatible with supported M6 rack and blade platforms using LGA4189. Not cross-compatible with later M7/M8 LGA4677 platforms. -
Vendor:Cisco
Cisco UCSC-ADGPU-240M6 | GPU air duct | 2U SFF/NVMe | C240 M6 | double‑wide/full‑length only
UCSC-ADGPU-240M6
Stock Status: Enquire
Vendor:CiscoCisco UCSC-ADGPU-240M6 | GPU air duct | 2U SFF/NVMe | C240 M6 | double‑wide/full‑length only
Dedicated GPU airflow duct for the C240 M6 2U SFF/NVMe chassis, intended for double‑wide and/or full‑length accelerator configurations to preserve thermal margins and performance. -
Vendor:Cisco
Cisco UCSC-HSLP-M6 | 1U/2U GPU heatsink | LFF/SFF chassis support | passive thermal block for GPU configurations
UCSC-HSLP-M6
Stock Status: Enquire
Vendor:CiscoCisco UCSC-HSLP-M6 | 1U/2U GPU heatsink | LFF/SFF chassis support | passive thermal block for GPU configurations
Purpose-built passive heatsink for GPU-enabled 1U/2U LFF/SFF server configurations, ensuring proper thermal coupling and airflow compliance in dense chassis layouts. Ideal as a field replacement or for GPU retrofit builds where dedicated thermal mass is required. -
Vendor:Cisco
Cisco UCSC-P-Q7D64GF | Dual 64GFC Gen7 HBA | 2x 64G Fibre Channel, PCIe 4.0 x8, NVMe/FC, SFP56 optics, backward 32G/16G
UCSC-P-Q7D64GF
Stock Status: Enquire
Vendor:CiscoCisco UCSC-P-Q7D64GF | Dual 64GFC Gen7 HBA | 2x 64G Fibre Channel, PCIe 4.0 x8, NVMe/FC, SFP56 optics, backward 32G/16G
High-performance dual-port 64G Fibre Channel host bus adapter for low-latency, high-throughput SAN connectivity. Supports NVMe/FC and traditional FCP with backward compatibility to 32GFC and 16GFC for smooth migrations. Includes secure, modern PCIe 4.0 performance for demanding virtualized and database workloads. -
Vendor:Cisco
Cisco UCSX-NVMEM6-W6400 | 2.5in U.2 NVMe SSD | 6.4TB | PCIe | enterprise, hot-swap capable
UCSX-NVMEM6-W6400
Stock Status: Enquire
Vendor:CiscoCisco UCSX-NVMEM6-W6400 | 2.5in U.2 NVMe SSD | 6.4TB | PCIe | enterprise, hot-swap capable
Enterprise-grade 2.5-inch U.2 NVMe solid-state drive delivering 6.4 TB of high-performance flash on a PCIe interface for latency-sensitive workloads in modern servers. Hot-swap capable in supported bays and designed for consistent throughput and reliability in data center deployments.
