46K2590: IBM 64GB DDR3 1066MHz ECC Registered CDIMM 1.5V 276p Memory
MPN: 46K2590|By: IBM
Price
Quoted on request
Current pricing within 48 business hours.
In Stockships worldwide
- Condition:
- New (authorized channel)
- Warranty:
- Full manufacturer warranty
Prefer your own inbox? Email us about this part
Bulk order? Request volume pricing
Authorized & warrantied
Sourced through authorized distribution, with the full manufacturer warranty.
Compatibility checked
Validated against your server platform before dispatch.
Delivery & lead time
Confirmed with your quote, including expedited freight where available.
Business & public sector
POs, net terms and volume pricing for qualified organizations, government and edu.
Product overview
This is the IBM 64GB DDR3 1066MHz ECC registered CDIMM 1.5V 276p Memory, order code 46K2590. A channel runs at the speed of its slowest module, so a faster part fitted alongside an older one buys nothing. Tell us what is already installed and we will keep the channel consistent.
Ask for the datasheet and firmware notes with the quote; we would rather you check our claim than take it.
Specifications
General
- Manufacturer part number
- 46K2590
- Product Type
- Server Memory
- Manufacturer
- IBM
Technical specifications
- Capacity
- 64GB
- Memory Type
- DDR3
- Module Type
- ECC
Published specifications for 46K2590. Anything not listed here we will confirm against the manufacturer's documentation before quoting, because we would rather answer the question than let you assume.
Where the IBM 46K2590 is deployed
- Like-for-like replacement of 46K2590 in an existing system
- Raising host memory to fit more virtual machines per node
- Filling the remaining channels on a part-populated board
Who this part is for
- Virtualisation teams raising host memory to fit more guests per node
- Procurement buying memory kits against a documented configuration
- Teams retiring mixed-speed memory in favour of one consistent part
What you get ordering this from us
Ordered by part number
Quoted and supplied against 46K2590 specifically, so what arrives is what your build sheet asks for.
Volume pricing
The listed price is per unit; tiered pricing applies at quantity and purchase orders are accepted.
Answered by a human
A quote comes back with the reasoning attached, including the cases where we think you should buy something else.
Questions buyers ask about this part
Is 46K2590 genuine and is the warranty intact?
Yes, new and sealed, with the IBM warranty and the firmware update path intact. We do not sell refurbished stock as new, and we do not source grey-market units.
Will 46K2590 work in my server?
Send the server model and a list of the DIMMs already fitted. Rank, voltage and population order all have to agree, and a module that is electrically compatible can still be unsupported in a given slot. We map the population before you order.
Do you accept purchase orders and quote at volume?
Yes, POs from qualified organisations, net terms on approval, and banded pricing once the quantity is worth banding. Send the number and we will price it properly.
Compare other Server Memory we stock
| Model | Brand | Key spec | Availability | From |
|---|---|---|---|---|
| SK hynix 32GB DDR5-5600 PC5-44800 ECC Registered CL46 288-Pin RDIMM Server Memory Module | SK hynix | Capacity: 32GB | In Stock | $215 |
| Samsung 16GB DDR4-3200 PC4-25600 ECC Registered CL22 288-Pin RDIMM Server Memory Module | Samsung | Capacity: 16GB | In Stock | $55 |
| Micron 128GB DDR4-3200 PC4-25600 ECC Load-Reduced CL22 288-Pin LRDIMM Server Memory Module | Micron | Capacity: 128GB | In Stock | $500 |
| Dell Technologies 16GB DDR5-4800 PC5-38400 ECC RDIMM CL40 288-Pin Server Memory Module | Dell Technologies | Capacity: 16GB | In Stock | $95 |
| HPE 16GB DDR5-4800 PC5-38400 ECC RDIMM CL40 288-Pin Server Memory Module | HPE | Capacity: 16GB | In Stock | $95 |
| Lenovo 16GB DDR5-4800 PC5-38400 ECC RDIMM CL40 288-Pin Server Memory Module | Lenovo | Capacity: 16GB | In Stock | $90 |
Browse all Server Memory or everything we stock from IBM.
Specifying a system, not just the part?
Host memory on an AI node is normally sized against accelerator memory, so the model and its working set both fit.