Dell PowerEdge R6615 - Buyers Guide
Dense, Scalable Performance for Modern Data Center Workloads
-
Dense Virtualization — Consolidate more VMs per 1U chassis with up to 128 EPYC Genoa cores and hardware-enforced AMD SEV per-VM memory isolation.
-
Hyper-Converged Infrastructure (HCI) — Combine compute, storage, and networking in a single 1U node; up to 16 EDSFF E3.S Gen5 NVMe drives enable all-flash HCI density.
-
Network Functions Virtualization (NFV) — AMD EPYC 9004’s high core count and PCIe Gen5 bandwidth support telco-grade NFV with OpenStack for low-latency packet processing.
-
High Performance Computing (HPC) — Wide DDR5 memory bandwidth, large per-core L3 cache (up to 32 MB/core), and PCIe Gen5 interconnects accelerate scientific and simulation workloads.
-
Software-Defined Storage (SDS) — Pair up to 16 EDSFF E3.S Gen5 NVMe drives with direct-attach NVMe (no RAID overhead) for distributed block or object storage.
-
Virtual Desktop Infrastructure (VDI) — High core count enables more user sessions per server; optional GPU support handles graphics-intensive desktop workloads.
-
Large-Scale Databases — 3 TB DDR5 memory capacity and fast NVMe I/O support in-memory database operations and data warehousing at scale.
4th Gen AMD EPYC™ 9004 — Up to 128 Genoa Cores per Socket
-
Single-socket platform — One AMD EPYC™ 9004 series processor; eliminates second-socket software licensing costs and simplifies thermal design in a 1U chassis.
-
Up to 128 cores (Zen4c) — Or up to 96 Zen4 cores depending on SKU; high thread density enables maximum VM consolidation per rack unit.
-
5nm process technology — AMD’s Genoa generation leverages 5nm TSMC manufacturing for improved performance-per-watt versus prior generations.
-
Up to 4.1 GHz boost clock — High single-thread frequency benefits latency-sensitive applications, transactional databases, and VDI user density.
-
Up to 400W TDP — Maximum TDP configuration requires Direct Liquid Cooling (DLC); standard air cooling supports configurations up to 300W.
-
32 MB L3 cache per core — Largest available x86 L3 cache reduces memory latency and improves hit rates for data-intensive workloads.
-
AMD Infinity Guard security — Hardware-embedded AMD SEV (Secure Encrypted Virtualization) and AMD SME (Secure Memory Encryption) protect workloads at silicon level.
-
PCIe Gen5 I/O — Integrated I/O controller supports up to 128 PCIe Gen5 lanes for high-throughput NVMe and networking expansion.
12 DDR5 DIMM Slots — Up to 3 TB at 4800 MT/s
-
12 DDR5 DIMM slots — One DIMM per channel (1 DPC); all 12 channels feed directly off the EPYC 9004 memory controllers for maximum bandwidth.
-
Up to 3 TB capacity — Supports 256 GB 3DS RDIMM Octa-rank modules per slot; maximum configuration enables large in-memory databases and HPC datasets.
-
4800 MT/s speed — DDR5 at full rated speed when populated at 1 DPC; Infinity Fabric and memory clock synchronized for optimal latency.
-
RDIMM support only — Registered ECC DDR5 DIMMs required; UDIMMs not supported; ensures error correction for enterprise and mission-critical workloads.
-
DIMM capacities: 16 GB, 32 GB, 64 GB, 128 GB, 256 GB — Single, dual, and quad-rank RDIMMs supported; 256 GB 3DS RDIMM (Octa-rank) reaches full 3 TB maximum.
-
1.1 V operating voltage — DDR5’s lower voltage versus DDR4 (1.2 V) reduces per-DIMM power draw and data center cooling overhead.
-
Dual-channel per DIMM slot — DDR5 dual sub-channel architecture doubles channel count effectively, improving memory bandwidth for parallel workloads.
PERC H965i Through H355 — Full RAID Coverage Across All Tiers
-
PERC H965i (Premium Performance) — PERC 12 generation; top-tier RAID controller with the highest IOPs performance and advanced SSD optimization for NVMe-intensive workloads.
-
PERC H755N & H755 (Premium Performance) — PERC 11 generation; hardware RAID 0/1/5/6/10/50/60 with NVMe optimizations; H755N variant adds NVMe-direct support.
-
PERC H355 & HBA355i (Value) — Entry-level hardware RAID and HBA options for cost-sensitive deployments; H355 supports RAID 0/1/5/10, HBA355i provides pass-through HBA mode.
-
Software RAID S160 — CPU-based RAID via Windows Storage Spaces or Linux MD-RAID; zero-cost option for software-defined environments.
-
External adapters: HBA355e, HBA465e, H965e — Enable connection to external SAS/NVMe JBODs and storage arrays; up to 3 external cards depending on riser configuration.
-
Front PERC form factor (fPERC) — All internal controllers mount in a dedicated front PERC slot without consuming a PCIe expansion slot, preserving full PCIe capacity for NICs and GPUs.
-
BOSS-N1 internal boot — Hardware RAID 1 mirrored pair of M.2 NVMe SSDs for OS boot; all front and rear data bays remain available for user data.
Seven Chassis Configurations — Up to 16 EDSFF E3.S Gen5 NVMe Drives
-
4 x 3.5-inch SAS/SATA front bays — Up to 80 TB raw capacity; supports 7.2K and 10K HDD plus SATA SSD; ideal for high-capacity archival and nearline storage.
-
8 x 2.5-inch NVMe front bays — Up to 122.88 TB all-NVMe; Gen4 NVMe SSDs in capacities from 400 GB to 15.63 TB per drive for performance-dense deployments.
-
10 x 2.5-inch SAS/SATA/NVMe front bays — Up to 153.6 TB mixed-mode; universal backplane accepts SAS, SATA, or NVMe drives for maximum configuration flexibility.
-
14 x EDSFF E3.S Gen5 NVMe front bays — Up to 107.52 TB; E3.S form factor provides higher drive density than 2.5-inch U.2, enabling NVMe-only configurations in the same 1U space.
-
16 x EDSFF E3.S Gen5 NVMe front bays — Up to 122.88 TB; maximum storage density option with direct-attach NVMe Gen5 at full PCIe Gen5 bandwidth per drive.
-
Rear bay options — 2 x 2.5-inch SAS/SATA (up to 30.72 TB) or 2 x E3.S Gen5 NVMe (up to 15.36 TB); rear module adds capacity without using front bay slots.
-
Hot-swap across all configurations — Front and rear drive carriers are hot-swappable for live replacement without system downtime.
BOSS-N1 — Dedicated NVMe OS Boot, All Data Bays Free
-
BOSS-N1 module — Boot Optimized Storage Subsystem (NVMe generation); dedicated M.2 NVMe SSD carrier mounts in a rear slot separate from all front data bays.
-
Hardware RAID 1 mirrored pair — Two M.2 NVMe SSDs in hardware-mirrored configuration provide OS boot redundancy; one drive failure does not interrupt operation.
-
No PCIe slot consumed — BOSS-N1 connects via a dedicated internal connector, leaving all expansion card slots free for NICs, HBAs, and GPUs.
-
All data bays remain available — With BOSS-N1 handling OS boot, every front and rear drive bay is dedicated to user data, maximizing usable storage capacity.
-
Optional — blank installed by default — Systems ship with a BOSS-N1 blank; the BOSS-N1 module kit is an add-on upgrade for deployments requiring dedicated boot drives.
-
NVMe performance for OS — M.2 NVMe interface provides faster OS boot and hypervisor load times compared to legacy SATA BOSS modules in prior generations.
Up to 2 × 75W Single-Width GPU — PCIe Gen5 Acceleration in 1U
-
Up to 2 single-wide GPUs — R1Q + R4P riser configuration supports two full-height, half-length GPU cards; ideal for VDI, AI inference, and media transcoding workloads.
-
75W per GPU maximum — Passive, single-width cards within 1U thermal envelope; NVIDIA A2 16 GB is a validated GPU for graphics-accelerated workloads.
-
PCIe Gen5 x16 GPU slots — Full Gen5 bandwidth to each GPU eliminates the PCIe bottleneck for data-heavy AI inference pipelines.
-
No impact on storage bays — GPU risers occupy rear PCIe slots only; front and rear drive bays remain fully populated regardless of GPU configuration.
-
Optional DPU kit — Data Processing Unit upgrade kit available for network offload and SmartNIC functionality, compatible with R6615 upgrade kits (Chapter 8 of service manual).
-
Thermal envelope maintained — GPU configurations restrict to lower-TDP CPU SKUs (240W cTDP maximum) to stay within 1U air-cooling thermal limits with HPR Gold fans.
Up to 3 PCIe Gen5 Slots — Eight Riser Configurations
-
Up to 3 PCIe expansion slots — Slot 1, Slot 2, and Slot 3 across eight riser configurations; maximum of three low-profile or two full-height cards depending on riser selection.
-
PCIe Gen5 x16 slots available — Risers R2T, R2U, R3P, R3S, R4P support Gen5 x16 lanes; full 64 GB/s per-slot bandwidth for high-throughput NICs and NVMe HBAs.
-
PCIe Gen4 x8 and x16 slots — Risers R2A and R3A provide Gen4 options; backwards-compatible with Gen3 cards for legacy adapter reuse.
-
Eight riser configurations — R2A, R2T, R2U, R3A, R3P, R3S, R4P, and R1Q+R4P support diverse slot needs from pure NIC configurations to GPU and rear-storage combinations.
-
Full-height support (R1Q + R4P) — Two full-height, half-length Gen5 x16 slots support larger cards including full-height GPUs and high-port-count FC HBAs.
-
OCP 3.0 integrated slot — Dedicated OCP NIC slot independent of expansion risers; supports x8 PCIe bandwidth and keeps mezzanine NIC from consuming a PCIe expansion slot.
-
Slots not hot-swappable — System must be powered down for PCIe card installation or removal; plan card configurations before deployment to avoid maintenance windows.
OCP 3.0 Networking — Up to 25 GbE per Slot in 1U
-
OCP 3.0 NIC slot (optional) — Dedicated mezzanine slot supports interchangeable OCP 3.0 cards without consuming a PCIe expansion slot; connected via x8 PCIe bandwidth from the processor.
-
OCP card speeds: 1 GbE to 25 GbE — Options include 1 GbE x4, 10 GbE x2, 10 GbE x4, 25 GbE x2, and 25 GbE x4; vendor options from Intel, Mellanox, and Broadcom.
-
LOM card: 2 x 1 GbE (optional) — BCM5720-based dual GbE LAN-on-Motherboard provides base network connectivity; LOM and OCP can both be installed simultaneously.
-
Dedicated iDRAC9 Ethernet port — Rear-panel out-of-band management port; isolated from production traffic for secure remote access regardless of OS or NIC state.
-
PCIe expansion NICs — Additional NICs in PCIe slots support up to 400 GbE (Mellanox NDR200) or 100 GbE multi-port cards for high-throughput north-south traffic.
-
SmartNIC support — Full-height riser configuration (R1Q+R4P) supports Pensando, Intel, and NVIDIA Mellanox 25/100 GbE SmartNIC cards for DPU-class network offload.
-
iDRAC Quick Sync 2 (optional) — Wireless NFC/Bluetooth module on front control panel enables contactless configuration and inventory via OpenManage Mobile.
700W to 1800W — Platinum & Titanium Hot-Swap Redundant PSUs
-
1800W Titanium (200–240 V AC / 240 V HVDC) — Highest-efficiency PSU for 400W TDP processor configurations; required for maximum system power loads.
-
1400W Platinum (100–240 V AC / 240 V HVDC) — High-capacity option; derates to 1050W at low-line 100–120 V AC; supports most standard processor TDP configurations.
-
1100W Titanium (100–240 V AC / 240 V HVDC) — Balanced efficiency and capacity for mid-range TDP configurations; 240 V DC variant also available.
-
1100W LVDC (-48 to -60 VDC) — Telecom-grade DC input for NFV and carrier-grade deployments using −48 V DC power infrastructure.
-
800W Platinum (100–240 V AC / 240 V HVDC) — Entry-level high-power option for lower-TDP CPU SKUs; broad voltage range supports both universal AC and HVDC data centers.
-
700W Titanium (200–240 V AC / 240 V HVDC) — Highest efficiency at lowest wattage; optimized for energy-efficient deployments with 240W cTDP processor SKUs.
-
Hot-swap redundant — Both PSUs are field-replaceable under live load; system continues full operation during PSU swap, eliminating planned maintenance windows.
-
60 mm form factor — Same PSU dimensions as 15G PowerEdge; compatible PSU form factor shared with prior generation simplifies spare parts management.
SmartCooling — Air Cooling + Optional Direct Liquid Cooling
-
Air cooling (standard) — Up to four high-performance Gold (HPR Gold) dual fan modules support most processor configurations including 300W TDP SKUs at up to 35°C ambient.
-
Standard (STD) fans — Available for lower-TDP processor SKUs (240W cTDP and below); four hot-plug dual fan modules provide redundant cooling with N+1 fan failure tolerance.
-
Direct Liquid Cooling (DLC) optional — Required for 400W TDP processor SKUs; rear coolant tubes connect to rack manifold and cooling distribution unit (CDU) in a closed-loop system.
-
ASHRAE A2 standard support — Air-cooled configurations operate up to 35°C ambient (most SKUs); A3/A4 support with HPR Gold fans and restricted configurations at 40°C and 45°C.
-
Thermal restriction matrix — CPU TDP, DIMM capacity, front storage, and rear storage all factor into allowed ambient temperature; 400W TDP SKUs require DLC in most configurations.
-
Hot-plug fan modules — All four fan assemblies are hot-swappable under full load; fan failure triggers an alert in iDRAC9 and remaining fans increase speed automatically.
-
LCD bezel & security bezel options — Optional LCD bezel displays system status; security bezel locks the front panel to prevent unauthorized drive removal or button access.
Front, Rear & Internal I/O — USB 3.0, VGA, iDRAC Direct
-
Front panel: USB 2.0 + iDRAC Direct (Micro-AB USB) — One USB 2.0 port for local access; iDRAC Direct Micro-AB port enables direct laptop-to-iDRAC connection for configuration without a network connection.
-
Front panel: VGA — DB-15 VGA output for local console access; iDRAC Direct status LED indicates connection state.
-
Rear panel: USB 3.0 + USB 2.0 — One USB 3.0 port for high-speed peripheral access; one USB 2.0 for management dongles and boot media.
-
Rear panel: VGA — Second DB-15 VGA output; note VGA is omitted on Direct Liquid Cooling configurations where the rear cooling manifold occupies that space.
-
Rear panel: Dedicated iDRAC9 Ethernet — 1 GbE management port isolated from production LOM/OCP traffic; enables always-on out-of-band access.
-
Internal: USB 3.0 (optional) — Optional internal USB card supports a USB 3.0 port for internal boot drives or management tools without using external rear ports.
-
Optional serial COM port — 9-pin DTE serial connector card available for legacy console redirects and out-of-band serial access; installs like an expansion card bracket.
-
Matrox G200 integrated video — 16 MB frame buffer; supports up to 1920×1200 at 60 Hz for full HD console access without a discrete GPU.
Cyber Resilient Architecture — Silicon Root of Trust & AMD Infinity Guard
-
Silicon Root of Trust — Cryptographically anchored boot chain validated from silicon up through firmware and OS; any firmware tampering is detected and blocked before execution.
-
AMD Secure Encrypted Virtualization (SEV) — Per-VM memory encryption in hardware; each virtual machine’s memory is encrypted with a unique key invisible to the hypervisor, OS, and other VMs.
-
AMD Secure Memory Encryption (SME) — Full-system memory encryption without software overhead; protects against physical memory attacks and cold-boot exploits.
-
TPM 2.0 — FIPS and CC-TCG certified; China NationZ TPM 2.0 variant available; hardware-backed key storage for BitLocker, Measured Boot, and zero-trust attestation.
-
Secure Boot + Secure Erase — UEFI Secure Boot validates OS loader signatures; Secure Erase cryptographically erases drives at end-of-life, preventing data recovery.
-
Secured Component Verification — Hardware integrity check validates that installed components match factory-certified configurations; detects unauthorized component swaps.
-
System Lockdown — Requires iDRAC9 Enterprise or Datacenter license; prevents unauthorized firmware and configuration changes during production operation.
-
Data at Rest Encryption (SEDs) — Self-Encrypting Drives with local or external key management (KMIP) integrate with the security architecture for full-stack data protection.
iDRAC9 — Out-of-Band Lifecycle Management & Redfish API
-
iDRAC9 (Integrated Dell Remote Access Controller) — Gen 16’s embedded lifecycle controller provides always-on out-of-band management independent of server power state or OS availability.
-
iDRAC RESTful API with Redfish — Industry-standard DMTF Redfish API enables automation, scripting, and integration with Ansible, Terraform, VMware vCenter, and other orchestration platforms.
-
iDRAC Direct — Micro-AB USB front-panel port enables laptop-to-iDRAC direct configuration during initial racking without a network connection.
-
iDRAC Service Module — Host-side agent extends iDRAC visibility into OS-level metrics, logs, and crash events; bridges in-band and out-of-band management data.
-
OpenManage Enterprise — Centralized console manages thousands of PowerEdge servers; plugins for Power Manager, Update Manager, and Service extend capabilities without additional tools.
-
OpenManage integrations — Microsoft System Center, Red Hat Ansible Modules, VMware vCenter and vRealize Operations Manager, ServiceNow, BMC TrueSight, and Terraform Providers.
-
iDRAC Quick Sync 2 (optional) — NFC/Bluetooth wireless module on left control panel enables contactless configuration and server status check via OpenManage Mobile app.
-
Dell Lifecycle Controller — Embedded firmware enables bare-metal OS deployment, driver updates, and BIOS configuration without external media or OS.
Windows Server, RHEL, Ubuntu, SLES & VMware ESXi
-
Microsoft Windows Server with Hyper-V — Full Hyper-V hypervisor support on Windows Server 2019/2022; AMD SEV provides hardware VM isolation for multi-tenant environments.
-
Red Hat Enterprise Linux (RHEL) — Certified on current RHEL versions; Ansible playbooks and RHEL integration via OpenManage enable automated provisioning and lifecycle management.
-
Canonical Ubuntu Server LTS — Long-term support Ubuntu versions certified; popular for Kubernetes, OpenStack, and cloud-native workloads running on EPYC 9004 hardware.
-
SUSE Linux Enterprise Server (SLES) — Certified for SAP HANA and enterprise Linux deployments; SLES 15 SP versions supported with full driver certification from Dell.
-
VMware ESXi — VMware ESXi certified; OpenManage Enterprise Integration for VMware vCenter enables VM-to-hardware visibility and BIOS/firmware lifecycle management from vCenter.
-
Dell Lifecycle Controller OS deployment — Bare-metal OS install directly from iDRAC; no separate boot media required; drivers pre-bundled for accelerated provisioning.
-
OEM-ready version available — From bezel to BIOS to packaging, OEM-ready R6615 variants can be branded for white-label cloud and service provider deployments.
Universal Rail Kit — 1U 19-Inch Rack Ready with CMA Option
-
1U rack form factor — 42.8 mm height (1.685 in); 482 mm width (18.97 in); standard 19-inch EIA rack compatible with all major rack vendors and data center standards.
-
Depth: up to 823 mm (32.39 in) with bezel — Longer than previous gen (R6515 was 728 mm); verify rack depth before installation, especially in older 900 mm depth racks.
-
Rail sizing compatibility matrix — Dell publishes a rack compatibility matrix covering rail types, adjustability ranges, and rack mounting flange types; consult before ordering rails.
-
Cable Management Arm (CMA) option — Optional CMA organizes rear cables and allows the server to be extended from the rack on rails without cable disconnection during servicing.
-
Tool-less rail design — Standard PowerEdge rails support tool-free installation in square-hole racks; adapter kits available for round-hole and threaded-hole racks.
-
Maximum weight: 20.2 kg (44.5 lbs) fully loaded — Plan rack load capacity accordingly; 17.4 kg without drives and PSU for initial rackmount before component population.
-
Quick Resource Locator (QRL) — QR code on system information tag links to setup videos, service manual, and configuration-specific support resources for fast field servicing.
R6615 vs. R6515 — What’s New in 16th Generation
-
4th Gen AMD EPYC 9004 Genoa (up to 128 cores) — Versus 3rd Gen EPYC Rome (SP3); the Genoa platform delivers significantly more cores per socket, new Zen4 core architecture, and PCIe Gen5 I/O.
-
DDR5 at 4800 MT/s vs. DDR4 at 3200 MT/s — Generation-over-generation memory bandwidth improvement; DDR5 dual sub-channel architecture effectively doubles channel width over DDR4 at the same slot count.
-
EDSFF E3.S NVMe Gen5 vs. no EDSFF support — The R6615 adds up to 16 E3.S NVMe Gen5 front bays and 2 E3.S rear bays; the R6515 had no EDSFF configuration option.
-
Up to 3 PCIe Gen5 slots vs. up to 2 Gen4/Gen3 slots — PCIe Gen5 doubles bandwidth per lane over Gen4; additional third slot unlocks more simultaneous NIC and storage adapter configurations.
-
Optional DLC vs. air cooling only — The R6615 adds Direct Liquid Cooling support for 400W TDP processor SKUs; the R6515 was limited to air cooling.
-
BOSS-N1 (NVMe M.2) vs. BOSS S1 (SATA M.2) — The R6615’s BOSS-N1 uses faster NVMe M.2 for OS boot; the R6515’s BOSS S1 used SATA M.2 with lower sequential throughput.
| Feature | R6615 (Gen 16) | R6515 (Gen 15) |
|---|---|---|
| Processor | AMD EPYC 9004 Genoa, up to 128 cores | AMD EPYC Rome (SP3) |
| DDR5 DIMM Slots | 12 DDR5 | 16 DDR4 |
| Memory Speed | Up to 4800 MT/s | Up to 3200 MT/s |
| Max Memory | 3 TB DDR5 | 1 TB RDIMM / 2 TB LRDIMM |
| EDSFF NVMe | Up to 16 E3.S Gen5 | Not supported |
| PCIe Slots | Up to 3× Gen5 capable | Up to 2× Gen4/Gen3 |
| Management | iDRAC9 | iDRAC9 |
| GPU Support | Up to 2× 75W SW | Up to 2× 70W SW |
| Direct Liquid Cooling | Optional | Not Available |
| Boot Storage | BOSS-N1 (NVMe M.2) | BOSS S1 (SATA M.2) |
| Form Factor | 1U rack server | 1U rack server |
ProSupport & ProDeploy — End-to-End Service Coverage
-
Dell ProSupport — 24×7 access to Dell hardware and software engineers; automated case creation via SupportAssist monitors the server and opens support tickets before issues escalate.
-
Dell ProDeploy — Factory integration, on-site deployment, and configuration services; servers arrive rack-ready with OS and drivers pre-installed to accelerate go-live timelines.
-
Dell SupportAssist — Automated monitoring agent detects hardware anomalies, sends diagnostic data to Dell, and triggers proactive outreach from support engineers.
-
Dell Residency Services — Dell-certified engineers embedded on-site for knowledge transfer, architecture guidance, and hands-on operational support during production ramp-up.
-
Dell Data Migration Services — Structured migration planning and execution for workload moves from legacy hardware to R6615 platforms with minimal downtime.
-
Dell TechDirect — Self-service portal for warranty lookup, parts dispatch, remote diagnostics, and service request management; available 170+ countries.
-
APEX Flex on Demand — Consumption-based payment model for R6615 infrastructure; capacity scales with usage, converting capital expenditure to operational expenditure.
Titanium PSUs & EPYC Genoa — Performance-per-Watt at Scale
-
Titanium-efficiency PSUs — 700W and 1100W and 1800W Titanium-rated PSUs exceed 96% efficiency at peak; minimizes waste heat and lowers data center cooling costs at the rack level.
-
AMD EPYC 9004 5nm efficiency — Genoa’s 5nm process node delivers more instructions per watt than prior-generation 7nm Rome; same throughput at lower absolute power draw for equivalent workloads.
-
cTDP configurability — Many EPYC 9004 SKUs support configurable TDP (cTDP) settings; operators can reduce processor TDP below rated maximum to fit lower-power facility constraints.
-
Direct Liquid Cooling for high-TDP SKUs — DLC removes heat more efficiently than air, enabling 400W TDP processors in 1U without proportionally increasing air handler capacity in the data center.
-
ASHRAE A4 support (45°C) — With DLC and restricted configurations, the R6615 operates in hotter aisles, reducing chiller set points and mechanical cooling energy expenditure.
-
Dell OpenManage Power Manager — Monitors per-server power consumption in real time; supports capping policies to keep racks within PDU and facility power budgets.
-
Recycled materials & end-of-life program — Dell Technologies’ Product Environmental Datasheet documents material composition; take-back and recycling available in select countries via Dell.com/recyclingworldwide.
Frequently Asked Questions — Dell PowerEdge R6615
The R6615 supports one AMD EPYC 4th Generation 9004 Series processor (Genoa). SKUs range from 16-core entry configurations up to 128 cores (Zen4c) or 96 cores (Zen4) at the high end, with clock speeds up to 4.1 GHz and TDPs from 240W to 400W. The 5nm process technology and up to 32 MB of L3 cache per core deliver a significant performance-per-watt improvement over the EPYC Rome generation used in the R6515.
The R6615 offers seven front-bay chassis configurations: 4 x 3.5-inch SAS/SATA, 8 x 2.5-inch NVMe, 10 x 2.5-inch SAS/SATA/NVMe, 14 x EDSFF E3.S Gen5 NVMe, or 16 x EDSFF E3.S Gen5 NVMe. Rear bay options add 2 x 2.5-inch SAS/SATA or 2 x E3.S Gen5 drives. The 16 x EDSFF configuration reaches up to 122.88 TB raw NVMe capacity, and the 10 x 2.5-inch universal backplane tops out at 153.6 TB with mixed drive types.
Yes, the R6615 supports optional Direct Liquid Cooling (DLC). DLC is required for 400W TDP processor SKUs (such as the EPYC 9754 128-core or EPYC 9554P 64-core) that exceed the air-cooling thermal envelope. DLC is a rack-level solution requiring rack manifolds and a Cooling Distribution Unit (CDU) — it does not operate standalone. Air cooling supports most configurations up to 300W TDP without DLC. Note that DLC configurations omit the rear VGA port to accommodate coolant tube routing.
Yes. Express Computer Systems stocks professionally reconditioned refurbished Dell PowerEdge R6615 servers tested, cleaned, and configured to your specifications — ready to deploy at significant cost savings versus new. Shop refurbished Dell R6615 servers at ECS.
The R6615 (Gen 16) upgrades the R6515 (Gen 15) on every key dimension: AMD EPYC 9004 Genoa with up to 128 cores replaces the older Rome SP3 architecture; DDR5 at 4800 MT/s replaces DDR4 at 3200 MT/s; PCIe Gen5 slots replace Gen4/Gen3; EDSFF E3.S Gen5 NVMe support is entirely new (R6515 had no EDSFF option); optional Direct Liquid Cooling is available where the R6515 was air-only; and the BOSS-N1 NVMe boot module replaces the SATA-based BOSS S1.
Ready to Deploy the Dell PowerEdge R6615?
Express Computer Systems specializes in professionally reconditioned, off-lease Dell PowerEdge servers tested and configured to your exact specifications. Get enterprise AMD EPYC 9004 Genoa performance at a fraction of the cost of new — backed by our expert configuration team.
Start building your custom server today