You add memory, press the power button, and the server beeps, hangs or reports less memory than you installed. Almost every case traces back to one of five causes, and each leaves a clue in the server log. This guide walks through them in the order worth checking, with what Dell, HPE, Lenovo and Supermicro document for each.
The short answer
- Read the log before changing anything. The management controller (iDRAC on Dell, iLO on HPE, XClarity Controller on Lenovo, the BMC on Supermicro) records what the memory check found. The error text points to the cause.
- The most common cause is the wrong module type. Registered ECC memory in a board that takes unbuffered memory, or the other way round, will not start.
- The next most common is a module the processor does not support. Dell documents cases where the DIMM looks right but the processor generation or family rules it out.
- Then population and mixing rules. Mixed module types, too many capacities, unbalanced channels and wrong slot order can stop a boot or leave memory unused.
- Last, memory mode and firmware. A protection mode can demand a specific layout, and some module capacities need a minimum BIOS version.
If you would rather have the match checked before you buy, send us the model and processor from our quote form. NEMIX RAM supplies matched memory for Dell PowerEdge, HPE ProLiant, Lenovo ThinkSystem and Supermicro servers.
Symptom to cause
| What you see | Most likely cause | Check first |
|---|---|---|
| Beeps or no video, with only the new modules installed | Wrong module type for the board, for example registered memory in an unbuffered-only board | Module type and DDR generation against the board or server specification |
| Log says no memory is detected after adding larger modules | Modules not supported by the processor generation or DRAM density | Processor model and the vendor memory support list |
| Server starts but shows less memory than installed | Population or mixing rules not met | Slot order, matching channels, and the number of capacities in use |
| Server starts and the log warns about an unsupported DIMM configuration | Memory protection mode needs a different module layout | The configured memory mode in the BIOS or system utilities |
| Fails only after a processor swap | The new processor generation does not support the installed modules | Processor family against the module part number |
Step 1: read what the server says
Dell logs the result in the System Event Log, which you can read in iDRAC. After new DIMMs on a 14th generation PowerEdge, Dell documents the message "No memory is detected" and traces it to a mismatch between the processor and the memory specification. HPE records memory problems in the Integrated Management Log. Supermicro boards can signal missing or unusable memory with a beep pattern. Write down the exact message and the slot it names, because the sections below key off it.
Cause 1: the wrong type of module
A board accepts one family of module. Supermicro answers one such case in its support FAQ: a SuperServer would not finish its first boot and beeped, because it takes unbuffered DDR3 modules and the owner had installed registered ECC modules. The fix was to fit the correct type. The same trap runs in both directions, and across generations. DDR4 and DDR5 do not fit each other's slots, entry servers take unbuffered ECC modules (UDIMM) only, and mainstream servers take registered modules (RDIMM) or load-reduced modules (LRDIMM).
Before ordering, confirm three things from the server specification: the DDR generation, the buffering (UDIMM, RDIMM or LRDIMM), and whether ECC is required. Our memory types guide and ECC vs non-ECC guide cover the differences, and the vendor guides cover each platform: Dell PowerEdge, HPE, Lenovo ThinkSystem and Supermicro.
Cause 2: the processor does not support the module
A module can be the right type and still be refused, because the processor sets which memory it can use. Dell documents two cases.
- 14th generation with first generation Xeon Scalable (Skylake). Dell's support article for new DIMMs that go undetected says Skylake processors require 8 GB density memory and that 16 GB density modules are not expected to work. Density here describes the memory chips on the module rather than its total size, so a high-capacity module built from denser chips can fail even though its capacity looks normal. The article lists many 14th generation models, including the R440, R640, R740 and T640.
- R860 and R960 with 96GB DDR5 modules. Dell reports "No memory found" during POST when 96GB DDR5 RDIMMs go into servers with medium core count processors. Per Dell, Intel's specification allows only extreme core count processors (the 84xxH family) to use them, not the 64xxH family, and the BIOS must be 1.6.6 or later.
The lesson is to check the processor model before the capacity. This matters most after a processor swap or on a server bought second hand, where the installed processor may not be the one the memory list assumed. Intel ARK lists the memory types, speeds and maximum capacity for each processor, and the vendor's memory support list narrows it to specific modules.
Cause 3: population and mixing rules
Each vendor publishes rules for what can share a channel and where modules go. They differ by model, so use the manual for your exact server. The two examples below show the shape of the rules.
| Rule | Dell PowerEdge R750 manual | Lenovo ThinkSystem SR665 documentation |
|---|---|---|
| Mixing module types | Do not mix RDIMM, LRDIMM and 3DS LRDIMM in a channel, across a memory controller or across sockets | RDIMM and Performance+ RDIMM do not coexist, and 3DS RDIMM does not coexist with other DIMM types |
| Mixing capacities and densities | Capacities can be mixed if the other rules are met, but more than two capacities in one system is not supported | Different DRAM densities can share a channel and a system, with the higher density installed first |
| Mixing ranks | Not covered on the pages I read | Different ranks can share a channel and a system, with the higher rank installed first |
| Mixing speeds | All modules run at the speed of the slowest module | Not covered on the page I read |
| DDR generation, ECC and voltage | All modules must be DDR4 | ECC modules only, 1.2V only, and different voltages do not coexist |
| Slot order | Fill the white release-tab sockets first, and keep channels identical | A single module goes in slot 1 of its channel, and diagrams show the order for each count |
Dell adds that mismatched configurations may prevent the system from booting, stop it during memory configuration, or leave some memory unused, and that unbalanced channels lose performance and may not be recognized. The practical rules that follow:
- Keep one module type in the system. Do not mix RDIMM and LRDIMM unless the manual for your model says it is allowed.
- Keep capacities to what the manual allows. On the Dell R750 that is two capacities at most.
- Fill slots in the documented order, and keep both processors populated the same way.
- Expect the slowest module to set the speed when speeds are mixed.
Cause 4: the memory mode demands a specific layout
Some servers boot but log an error because the configured memory protection mode cannot work with the modules fitted. HPE documents this on Gen10 ProLiant servers. From ROM revision 1.50, Fast Fault Tolerance is enabled by default for every workload profile except Low Latency. It needs at least two ranks on each populated channel and HPE SmartMemory in x4 organization. When the modules do not meet that, the log shows "Unsupported DIMM Configuration Detected - Configuration Does NOT Support Configured AMP Mode". HPE gives two fixes: install modules that meet the requirement, or change the memory mode to Advanced ECC in the ROM-based setup utility (RBSU).
Cause 5: firmware that predates the module
Large or new modules can need a newer BIOS. The Dell R860 and R960 case above is an example: 96GB DDR5 RDIMMs are supported from BIOS 1.6.6. Update the BIOS and management firmware to the vendor's current release before a large upgrade, while the server still starts with its original memory.
A recovery routine that works on any server
- Power down and remove the new modules. Confirm the server starts with the original memory. If it does not, the problem is not the new memory.
- Read the log from the last attempt and note the message and slot.
- Check type, generation and density of the new module against the specification and the processor.
- Test one module in the first slot, which is slot 1 on a Lenovo channel and the first white-tab socket on a Dell. A module that works alone and fails in a group points to population or mixing rules.
- Add modules in the documented order and in the counts the manual lists, checking the installed capacity at each step.
- Update the BIOS and firmware, then repeat the full population.
- Send us the details. The server model, the processor, the part number of the installed modules and the log message are enough for us to say which module fits.
Frequently asked questions
Why won't my server boot after I added RAM?
Usually the module is the wrong type or generation for the board, the processor does not support it, or the population breaks the vendor's rules. Read the management controller log for the exact message, then check each cause in that order.
What does "No memory is detected" mean after a Dell 14G upgrade?
Dell links it to a mismatch between the processor and the memory specification. On first generation Xeon Scalable processors Dell says 8 GB density memory is required and 16 GB density modules are not expected to work.
Can I mix RDIMM and LRDIMM in a server?
Dell's R750 manual says not to mix RDIMM, LRDIMM and 3DS LRDIMM in a channel, across a memory controller or across sockets. Check the manual for your model before mixing anything.
Can I mix different capacity modules?
Sometimes. Dell's R750 allows it with a limit of two capacities in one system, and Lenovo's SR665 allows different densities with the higher density installed first. Rules vary by model, so confirm in the manual.
Should I update the BIOS before installing larger modules?
Yes. Some capacities need a minimum BIOS version, as Dell documents for 96GB DDR5 RDIMMs on the R860 and R960. Update first, while the server still starts with its original memory.
Do DDR4 and DDR5 modules fit the same slot?
No. They are keyed differently and do not interchange, so the server's DDR generation decides which you can fit.
Sources
Read on October 28, 2026: Dell support article on 14th generation PowerEdge servers that do not detect new DIMMs, Dell support article on the R860 and R960 No memory found error with 96GB DDR5 modules, the Dell PowerEdge R750 Installation and Service Manual, the HPE support advisory on the unsupported DIMM configuration message in the IML, Lenovo's ThinkSystem SR665 memory population rules and a Supermicro support FAQ on a server that beeps with the wrong memory type. Rules differ by model and generation, so confirm your exact server in its own manual.
Memory for servers, matched to the model
NEMIX RAM is a US-based memory specialist, in business since 1993. Every server module is researched, matched, and compatibility-tested by NEMIX RAM, is JEDEC-compliant, and is backed by a lifetime warranty. Purchase orders, net terms and volume pricing are available for IT teams, resellers, and government and education buyers. Request a quote with your server models, processors and quantities, or contact our sales team. You can also browse DDR4 RDIMM, DDR4 LRDIMM, DDR5 RDIMM, ECC UDIMM, Dell PowerEdge memory and HPE memory, or the data center memory refresh guide.
NEMIX RAM is an independent memory supplier and is not affiliated with Dell Technologies, Hewlett Packard Enterprise, Lenovo or Super Micro Computer. Dell, PowerEdge, iDRAC, HPE, ProLiant, iLO, Lenovo, ThinkSystem, XClarity, Supermicro, Intel, Xeon and AMD are trademarks of their respective owners, used here only to identify compatible systems.