Intel Data Center GPU Flex 170 OCP Accelerator Mezzanine 16GB GDDR6

Intel Data Center GPU Flex 170 OCP Accelerator Mezzanine 16GB GDDR6

Brand: Intel | Category: GPUs

SKU: XPGEDCGPUF170O | Part #: XPGEDCGPUF170O | MPN: XPGEDCGPUF170O

Contact for Pricing — Request a Quote

Request a Quote Contact Us

About the Intel Data Center GPU Flex 170 OCP Accelerator Mezzanine 16GB GDDR6

AI inference and real-time media processing workloads benefit from the Intel Data Center GPU Flex 170's Intel Xe-HP architecture, which delivers dedicated hardware acceleration across a broad range of enterprise applications. The XPGEDCGPUF170O is Intel's purpose-built accelerator for organizations requiring efficient, low-power GPU compute in density-constrained environments. With 32 Xe-Cores distributed across 2 compute tiles and 16GB of GDDR6 memory paired to a 256-bit interface, this mezzanine card moves data at 448 GB/s—critical for latency-sensitive inference pipelines and media transcode workflows that demand sustained throughput without excessive power consumption.

Data center architects and AI infrastructure teams select the Intel Data Center GPU Flex 170 for its combination of thermal efficiency and open standards support. The card operates within a 75W thermal design power envelope using passive cooling managed by the host chassis, eliminating the need for dedicated thermal infrastructure. Its OCP Accelerator Mezzanine 2.0 form factor and PCIe 4.0 x16 host interface integrate seamlessly into existing Intel-based server architectures. The Xe Matrix Extensions (XMX) compute engines handle INT8 and FP16 precision, while the dedicated Xe Media Engine accelerates AV1, HEVC (H.265), AVC (H.264), and VP9 codec operations. Full SR-IOV virtualization support enables multi-tenant isolation, and comprehensive software stacks including Intel oneAPI Toolkit, Intel oneVPL, and Intel Media SDK ensure rapid application deployment across Linux operating systems—RHEL, Ubuntu, and CentOS. Additional flexibility comes through OpenCL, Level Zero, SYCL, and oneAPI programming interfaces.

Whether you are evaluating accelerators for inference scaling, real-time transcoding, or hybrid workload consolidation, the Intel Data Center GPU Flex 170 delivers measurable performance within strict power budgets. For specifications, availability, and volume pricing on part number XPGEDCGPUF170O, contact Omnixon Global to request a detailed quotation.

Technical Specifications

BrandIntel
CategoryGPUs
SKUXPGEDCGPUF170O
Part NumberXPGEDCGPUF170O
ConditionNew
Manufacturer Part NumberXPGEDCGPUF170O
Product FamilyIntel Data Center GPU Flex Series
GPU ModelIntel Data Center GPU Flex 170
ArchitectureIntel Xe-HP
Xe-Cores32 (across 2 compute tiles)
Memory Capacity16GB GDDR6
Memory Interface Width256-bit
Memory Bandwidth448 GB/s
Form FactorOCP Accelerator Mezzanine 2.0
TDP (Thermal Design Power)75W
CoolingPassive (chassis-managed airflow)
Host InterfacePCIe 4.0 x16
API SupportOpenCL, Level Zero, SYCL, oneAPI
Media EngineIntel Xe Media Engine (dedicated hardware encode/decode)
Supported CodecsAV1, HEVC (H.265), AVC (H.264), VP9
Virtualization SupportSR-IOV
AI Compute EnginesXe Matrix Extensions (XMX) — INT8, FP16
Software FrameworkIntel oneVPL, Intel Media SDK, oneAPI Toolkit
Operating System SupportLinux (RHEL, Ubuntu, CentOS)

Frequently Asked Questions about Intel Data Center GPU Flex 170 OCP Accelerator Mezzanine 16GB GDDR6

What server platforms accept the Intel Data Center GPU Flex 170 OCP Accelerator Mezzanine 16GB GDDR6?

Reference servers include Dell PowerEdge XE9680 / XE9712, HPE Cray XD670, Lenovo ThinkSystem SR685a / SR675 V3, Supermicro AS-A21GE / SYS-821GE, Gigabyte G593 / G894, ASUS ESC. Share your target platform in the RFQ and we will confirm chassis-to-GPU compatibility and recommended NIC pairing.

How long is the lead time on AI GPUs?

Highly model-dependent. L40S / RTX-class: typically 3-6 weeks. H100/H200/B200 in SXM form factor: 12-16 weeks for whole-platform allocations. We quote genuine-channel ETAs only — no grey-market promises.

Do you supply matched networking (Quantum InfiniBand / Spectrum-X)?

Yes — Omnixon stocks the full NVIDIA networking lineup (Quantum-2 / Quantum-X InfiniBand, Spectrum-X Ethernet, ConnectX NICs, BlueField DPUs) so we can quote a complete training-cluster BOM, not just the GPUs.

Can you help with NVIDIA AI Enterprise licensing?

Yes. We hold genuine-channels for NVIDIA AI Enterprise software subscriptions. Add it to your RFQ and we quote node-aligned licensing along with the hardware.