Where are the DDR5 ECC UDIMMs?
- Thread starter thebluehue
- Start date
Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.
Ah, bummer. This page tricked me: M323R4GA3BB0-CQK - 1x 32GB DDR5-4800 UDIMM PC5-38400U Dual Rank x8 Replacement
The table on the right says ECC.
i doubt its a very low profile like shown in the picture.
low profile is the normal height we have on all most all DIMMs.
I'm considering giving these a shot on the Asus Pro WS W680. The report of them working on the supermicro X13SAE w680 board is encouraging but holy crap is it an expensive roll of the dice if it they don't work with the return policy being go pound sand.
I have done the same with my Gigabyte AM5 B650 board. I have tested DDR5 Micron ECC modules with the B650 board and ran commands if they are running in ECC mode. I ran the commands "dmidecode" and "wmic". Both commands report that the memory should run in "multi-bit ECC" mode. Unfortunately, it is probably not the case, according to this Gigabyte comment:
Gigabyte said the ECC memory is compatible but are only run in non-ECC mode.
So I don't think there is really a way to test if ECC works unless you want to overclock the RAM, blowing a hair dryer on the RAM, short-circuiting the RAM contacts, buying defective RAM and testing it, etc.
Rumors say that once AGESA 1.0.0.5 is released, AM5 motherboards will support ECC. However, I am not holding my breath. ECC was removed in the spec pages of Gigabyte and ASRock and they might not be bringing ECC back into their specs. Even if ECC becomes officially supported, the other issue is RAM compatibly itself. For example, there are already memory issues on my other Gigabyte W480M VISION W board. I am fairly certain my Gigabyte W480M board is defective, damaged or incompatible with the models of ECC memory I tested. It is not a defect of the CPU or memory, as I have tested and four different memory modules from two brands, and two different Xeon CPUs, and both output errors in Memtest86. However, when I sent the W480M board to Gigabyte for inspection and they found no errors with a Core i5 CPU with a different brand of memory.
Last edited:
I do not know how to test, but since ECC is supported by the CPU, my guess (hope?) is that if the OS sees ECC, then ECC is active and errors are being corrected/reported. But honestly it would mostly be hope unless and until someone more knowledgeable than me can actually confirm that is the case.
FWIW, the page for the ASUS Prime board I'm planning to use does mention ECC, but it's pretty lukewarm (and there are no ECC UDIMMs on the memory QVL list):
I just had my first experience with Wired Zone and it didn't go well. Their website was showing the Hynix ECC RAM in stock. I ordered 2. Email order confirmation that came 5 mins later said Drop Ship. About an hour later, their website then shows the RAM as not in stock - drop ship.
I emailed them and they basically say "sorry - mix up - SuperMicro's fault". So i give them the benefit of the doubt. 3 days later, the WZ website shows the RAM in stock - but I still don't have a shipping confirmation.
5 days later the SM web store shows them in stock, yet I still don't have a shipping tracking number. I email again asking for the tracking # since SM clearly had them in stock, otherwise cancel and I'll just go order directly from SM. They cancel the order. Then I go over to the SM web store and now out of stock.
/frustrating
I emailed them and they basically say "sorry - mix up - SuperMicro's fault". So i give them the benefit of the doubt. 3 days later, the WZ website shows the RAM in stock - but I still don't have a shipping confirmation.
5 days later the SM web store shows them in stock, yet I still don't have a shipping tracking number. I email again asking for the tracking # since SM clearly had them in stock, otherwise cancel and I'll just go order directly from SM. They cancel the order. Then I go over to the SM web store and now out of stock.
/frustrating
Last edited:
That definitely sounds like a frustrating experience 
.
My experience was different with Wiredzone.com. Here is how it went:
.My experience was different with Wiredzone.com. Here is how it went:
- On 2023-02-18, I ordered 4 of the Hynix HMCG88MEBEA Memory 32GB DDR5 4800MHz UDIMM MEM-DR532L-HL01-EU48 (it showed MFG Drop Ship), with FedEx 2 Day Shipping option, as the Micron MTC20C2085S1EC48BA1 Memory 32GB DDR5 4800MHz UDIMM MEM-DR532L-CL01-EU48 (showed out of stock at the time, but currently at the time of this forum post are in stock and lower in price than the HL01 variant).
- Got order number email almost immediately there after placing the order.
- It wasn't until 2023-02-21 that I got an update about the order being processed and the ETA for manufacturer drop-ship would be Early March. Not the most ideal as all my other parts were slated to have arrived in the week starting 2023-02-19, but not the end of the world.
- On 2023-02-22 I got another update saying the order is ready to ship and shipment tracking information would be provided soon.
- On 2023-02-23 I got an email stating the order is on its way and link to FedEx tracking, indicating scheduled delivery date of 2023-02-24.
Dihuni has the 32 GB Micron ECC memory in stock for $164.00. No sales tax but they do charge for shipping.
Last edited:
AGESA 1.0.0.5-C BIOS update has been released for Gigabyte AM5 boards. Might test sometime on my B650 board.
Specs still mention only non-ECC memory as supported though.
Might contact Gigabyte again to check whether ECC will be supported in this update, or if it only will run in non-ECC mode.
Probably not, since the description only mentions "1. Update AMD AGESA 1.0.0.5 C for Ryzen 9 7950X3D & Ryzen 9 7900X3D support".
Don't know if the ECC issue is fixed in the 1.0.0.5-C or if it would be fixed another patch such as 1.0.0.5-D, 1.0.0.5-E, etc. Don't know if ECC will ever be fixed for this board. Don't know if it is already fixed in 1.0.0.4.
Specs still mention only non-ECC memory as supported though.
Might contact Gigabyte again to check whether ECC will be supported in this update, or if it only will run in non-ECC mode.
Probably not, since the description only mentions "1. Update AMD AGESA 1.0.0.5 C for Ryzen 9 7950X3D & Ryzen 9 7900X3D support".
Don't know if the ECC issue is fixed in the 1.0.0.5-C or if it would be fixed another patch such as 1.0.0.5-D, 1.0.0.5-E, etc. Don't know if ECC will ever be fixed for this board. Don't know if it is already fixed in 1.0.0.4.
Last edited:
I have updated BIOS to F5b which has AGESA 1.0.0.5-C. Memtest86 10.2 displays the same message: "ECC Enabled: N/A (Unknown)".
My CPU is defective but it is not sent to AMD for replacement yet. Even with CPB set to Disabled, it still crashed three times during several days of use. I just tried booting into Linux and it froze.
My CPU is defective but it is not sent to AMD for replacement yet. Even with CPB set to Disabled, it still crashed three times during several days of use. I just tried booting into Linux and it froze.
Last edited:
Oof. That's rough. I'm reconsidering Zen4 for my uses for reasons unrelated to ECC support. Idle power is just too high for me for something that is going to be running 24/7. Too bad because the power-per-watt efficiency is off the charts for these chips, but I'm typically running 5-10% total CPU load most of the time, with only occasional spikes of full usage. I'm thinking a 12th/13th gen i5 on a W680 board is a better fit for my usage pattern.