2.3 GHz / 20 Cores / 40 Threads / 27.5 MB Cache / 125 W / SST-BF
| Socket | LGA3647 |
|---|---|
| Series | Xeon Gold |
| Generation | 2nd Gen (Cascade Lake) |
| Cores | 20 |
| Memory Type | DDR4 RDIMM · DDR4 LRDIMM · DDR4 3DS RDIMM |
| Max Memory Speed | 2933 MHz |
| PCIe Version | PCIe 3.0 |
| Lithography | 14 nm |
| Memory Channels | 6 channels |
| Condition | New · Used |
| Threads | 40 |
| Base Frequency | 2.3 GHz |
| Max Turbo Frequency | 3.5 GHz |
| Cache | 27.5 |
| TDP | 125 W |
Intel Xeon Gold 6230N — 20 cores and 40 threads at 2.3 GHz base / 3.5 GHz turbo, 27.5 MB cache, 125 W TDP, FCLGA3647, with three UPI links, four-socket scalability and Speed Select Technology – Base Frequency. Supplied new or as tested pulls, with a 1-year warranty.
The N suffix points at a different load profile, not a different chip. Against the 6230 the core count, the cache and the power are identical at 20 cores, 27.5 MB and 125 W; what moves is the clock shape. This part trades 400 MHz of peak turbo for 200 MHz of base — 2.3/3.5 against 2.1/3.9. For a workload that runs at or near all-core load most of the time, the base clock is the number that shows up on the bill, and this is the higher one.
The thermal side runs the other way and is worth reading carefully: TCASE 78 °C at 125 W leaves about 0.34 °C/W for the heatsink from a 35 °C inlet, against 0.42 for the plain 6230. TCASE is the maximum permitted IHS temperature rather than spare headroom, so the lower figure is the stricter requirement — the N part asks more of the cooler, because it is designed for enclosures that push inlet temperatures higher than a data-centre row does. If you are fitting one where a plain 6230 ran, check the heatsink and the inlet, not just the TDP.
Where it is genuinely unique in this family is SST - Base Frequency. Intel's Speed Select has two variants: Performance Profile rebalances cores and frequency across a fixed envelope, and Base Frequency raises the guaranteed base clock on a chosen subset of cores. This SKU supports the base-frequency variant; the plain 6230 and the 6230T do not. In practice that means firmware can guarantee a higher sustained clock to the cores running the latency-critical thread while the rest of the socket works at the standard base — useful for network functions and control-plane work, where a small number of threads decide the outcome.
Twenty cores at 1.375 MB of cache each is a throughput arrangement, and the clock shape tells you what kind of throughput. A sustained all-core load settles at 2.3 GHz here, which is 9.5 % above the plain 6230's 2.1 GHz; a burst on one or two threads tops out at 3.5 GHz, which is 10 % below the plain 6230's 3.9 GHz. So the part is faster where the load is steady and slower where it is spiky. That is the correct trade for a packet-processing or session-handling node and the wrong one for a build server.
The rest of the platform is the standard Gold set: six channels of DDR4-2933 to 1 TB per socket, PCIe 3.0 with 48 lanes, Optane persistent memory, AVX-512 with two FMA units, DL Boost (VNNI), Intel RDT, Intel VMD and Run Sure. Embedded options are available on this SKU, which is what makes it orderable for a platform with a fixed bill of materials and a long qualification cycle.
| Product collection | 2nd Gen Intel Xeon Scalable (Cascade Lake) |
|---|---|
| Processor number / sSpec | 6230N / SRFPR |
| Ordering code | CD8069504283604 (tray, B1 stepping; no boxed SKU) |
| Cores / Threads | 20 / 40 (Hyper-Threading) |
| Base / Max turbo frequency | 2.3 GHz / 3.5 GHz |
| Cache | 27.5 MB |
| TDP | 125 W (TCASE 78 °C) |
| Socket | FCLGA3647 (FC-LGA14B, 76.0 mm × 56.5 mm) |
| Memory | 6 channels DDR4-2933, up to 1 TB per socket, ECC RDIMM / LRDIMM / 3DS RDIMM, Optane persistent memory supported |
| Scalability | 4S (up to four sockets), 3 UPI links at 10.4 GT/s |
| PCIe | 3.0, 48 lanes |
| Speed Select | SST – Base Frequency: yes; SST – Performance Profile: no |
| Lithography | 14 nm |
| Launch | Q2 2019 |
| Status | Launched; End of Servicing Updates 30 June 2025 |
| Embedded options | Available |
| Accelerators | AVX-512 with 2 FMA units, Intel DL Boost (VNNI), Intel RDT, Intel VMD, Run Sure |
The 6230N fits FCLGA3647 boards whose BIOS supports 2nd Gen Xeon Scalable Cascade Lake parts, in single-, two- and four-socket configurations. Because the N parts are usually designed into a fixed platform rather than swapped into a general-purpose server, confirm the platform's approved processor list before buying:
Two checks matter more here than on a plain desktop-grade swap: the inlet temperature the enclosure actually runs at, because the 78 °C case limit is tighter than the 6230's 87 °C; and whether the firmware exposes the SST - Base Frequency table, because without platform support the feature is simply absent and the part behaves as a standard 20-core SKU.
| Processor | Generation | Cores / Threads | Base / Max turbo | Cache | TDP | Memory support |
|---|---|---|---|---|---|---|
| Xeon Gold 6230N | 2nd Gen (Cascade Lake) | 20 / 40 | 2.3 / 3.5 GHz | 27.5 MB | 125 W | 6ch DDR4-2933 · 1 TB · 4S |
| Xeon Gold 6230 | 2nd Gen (Cascade Lake) | 20 / 40 | 2.1 / 3.9 GHz | 27.5 MB | 125 W | 6ch DDR4-2933 · 1 TB · 4S |
| Xeon Gold 6230T | 2nd Gen (Cascade Lake) | 20 / 40 | 2.1 / 3.9 GHz | 27.5 MB | 125 W | 6ch DDR4-2933 · 1 TB · 4S |
| Xeon Gold 6230R | 2nd Gen Refresh | 26 / 52 | 2.1 / 4.0 GHz | 35.75 MB | 150 W | 6ch DDR4-2933 · 1 TB · 2S |
| Xeon Gold 6248 | 2nd Gen (Cascade Lake) | 20 / 40 | 2.5 / 3.9 GHz | 27.5 MB | 150 W | 6ch DDR4-2933 · 1 TB · 4S |
| Xeon Gold 6242 | 2nd Gen (Cascade Lake) | 16 / 32 | 2.8 / 3.9 GHz | 22 MB | 150 W | 6ch DDR4-2933 · 1 TB · 4S |
Against the plain 6230 it is a straight swap of peak for floor: 200 MHz more base, 400 MHz less turbo, the same 125 W, and a case limit that is 9 °C tighter. Steady-state workloads gain, bursty ones lose, and the cooling requirement goes up rather than down.
Against the 6230R the Refresh part offers six more cores and 8.25 MB for 25 W, but drops to two sockets. Where the node is a two-socket box and the firmware supports Refresh parts, that is more machine; where the platform is a fixed four-socket design, this SKU is the one that fits it.
Against the 6248 you give up 200 MHz of base and 400 MHz of turbo, and save 25 W per socket. For a twelve-month deployment in a power-capped rack that trade is usually right; for a build node it is usually not.
Against the 6242 four cores and 5.5 MB cost 35 MHz of base and 25 W. Core count is the wrong thing to economise on in a throughput node, so the 6242 only makes sense when the licence tier is set at 16 cores.
Every processor carries a 1-year warranty against defects in workmanship and function. Verified faulty units are replaced or refunded and we cover return shipping. Stock is held in both new and used condition; the condition you order is confirmed before dispatch, and used units are pull-tested stock rather than untested lots. Stock is held in our own warehouse and normally dispatches within three to seven working days, with tracking issued at dispatch; volume orders are quoted individually with lead time confirmed at the time of enquiry.
Pricing is quoted per order because condition and quantity both move it. Send the quantity, the condition you need and the platform the part is going into, and you will get a quote with lead time the same working day. If the design depends on the SST - Base Frequency table, say so at enquiry and we will confirm the sSpec and stepping against your platform's approved list before quoting.
Tell us who you are — we reply within one working day with a firm quote, stock status and lead time. Volume discounts available.
2.1 GHz / 8 Cores / 16 Threads / 11 MB Cache / 85 W / 78 °C TCASE
2.2 GHz / 8 Cores / 16 Threads / 11 MB Cache / 70 W / 91 °C TCASE
2.3 GHz / 10 Cores / 20 Threads / 13.75 MB Cache / 95 W / 91 °C TCASE
2.4 GHz / 10 Cores / 20 Threads / 13.75 MB Cache / 100 W / 2-Socket