• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

Unbuf ECC regular desktop driver pc

Joined
Dec 26, 2006
Messages
4,161 (0.62/day)
Location
Northern Ontario Canada
Processor Ryzen 5700x
Motherboard Gigabyte X570S Aero G R1.1 BiosF5g
Cooling Noctua NH-C12P SE14 w/ NF-A15 HS-PWM Fan 1500rpm
Memory Micron DDR4-3200 2x32GB D.S. D.R. (CT2K32G4DFD832A)
Video Card(s) AMD RX 6800 - Asus Tuf
Storage Kingston KC3000 1TB & 2TB & 4TB Corsair MP600 Pro LPX
Display(s) LG 27UL550-W (27" 4k)
Case Be Quiet Pure Base 600 (no window)
Audio Device(s) Realtek ALC1220-VB
Power Supply SuperFlower Leadex V Gold Pro 850W ATX Ver2.52
Mouse Mionix Naos Pro
Keyboard Corsair Strafe with browns
Software W10 22H2 Pro x64
Good day

I always found this an interesting option since amd offered the capability way back. I have seen posts saying even if price was at par with regular ram there is no reason or point of this on the regular ordinary desktop anyway and just use regular ram and carry on with life.

thoughts?
 
Unless your doing sensitive work on your PC then I can't see a cost benefit to using ECC ram, all it will do is add latency
 
Good day

I always found this an interesting option since amd offered the capability way back. I have seen posts saying even if price was at par with regular ram there is no reason or point of this on the regular ordinary desktop anyway and just use regular ram and carry on with life.

thoughts?
I have used ECC RAM on my machines since the Phenom II days. If you don't keep your computer on 24/7, regular RAM will be faster, but you can also tighten the timings of ECC RAM and rest easy knowing that any errors will be caught.
 
My mistake I should have said DDR4
Not a mistake. DDR5 ECC is only for ensuring functional memory as without on-die ECC, die yield is no longer acceptable. This isn't equivalent to ECC for previous generations of DDR.
 
Regular DDR5 ECC is "resting" only. I dont believe it will correct errors being sent wrong, only check is for the bits in the cells not getting flipped. Its been a while since I read up on it. Someone else probably has a better explanation.
 
Good day

I always found this an interesting option since amd offered the capability way back. I have seen posts saying even if price was at par with regular ram there is no reason or point of this on the regular ordinary desktop anyway and just use regular ram and carry on with life.

thoughts?
I've been using ECC in my rigs since 2006-ish. Since I run VM's ECC gives me some extra piece of mind that ram won't be a problem. Never really had any instability problems daily (many years without BSOD) with it so that is why I stick with it. There are opportunities to OC or tighten the timings within reason. For example I had been running DDR4-2666 at DDR4-3200 before I was able to get an actual DDR4-3200 kit. For many years I didn't even realize my ASUS board was OC-ing my DDR2-667 - no problems. AM4 supports it with the right motherboards and CPU's (non-Pro APU's don't support it) and I had been able to confirm it's working and Win10 will log corrected errors which can be useful when overclocking. If your doing any important work and favor stability over benchmarking performance or working with high capacity of ram I would recommend ECC. I would never trust putting files on a ramdisk without ECC. The price can be comparable to DDR4 kits you can find today approx $80-$100 per 16GB module.
 
Last edited:
Thanks for the feedback. I don’t do anything critical but cpu will be a 5600 and motherboard will support it. I’m looking at a jedec 2x16GB dual channel kit and the difference between regular and ecc is about $30. When I game with friends I’m the one that hosts so I’m thinking maybe $30 extra may as well?
 
Thanks for the feedback. I don’t do anything critical but cpu will be a 5600 and motherboard will support it. I’m looking at a jedec 2x16GB dual channel kit and the difference between regular and ecc is about $30. When I game with friends I’m the one that hosts so I’m thinking maybe $30 extra may as well?
If you do decide to use ECC might want to check your motherboard QVL list. Back in the early days of Zen+ I used Micron 18ASF2G72AZ-2G6D1 (I think the model was CT16G4WFD8266) unbuffered DDR4-2666 initially but Asrock needed to give me a beta bios to fix compatibility issues. This same ram also worked fine for Zen2 (3800x, 3950x) and Zen3 (5950x). If you are running a game server you could make the argument to just get non-ECC faster DDR4-3600/DDR4-3800 ram instead.
 
Last edited:
All DDR5 kits have an ecc chip. It is the norm.

This is not accurate.
The data is stored with ECC but (unless you have a real ECC DDR5) the ECC bits are not transferred between the memory and the CPU. Especially if you are overclocking a bitflip can occur there...

Unless your doing sensitive work on your PC then I can't see a cost benefit to using ECC ram, all it will do is add latency

I wouldn't do anything on my PC, even writing text documents without using an ECC ram. The possibility of a silent data corruption is just not worth it.

Of course, if you are just playing games and not doing any work (in the broadest sense) than you probably don't need an ECC. But I wouldn't buy and use any PC without ECC ram if it is possible...
 
This is not accurate.
The data is stored with ECC but (unless you have a real ECC DDR5) the ECC bits are not transferred between the memory and the CPU. Especially if you are overclocking a bitflip can occur there...
It's accurate alright.

But maybe in your head you misread...

All DDR5 has an ecc chip.

I did not say if you had to enable it or not.
 
Sorry but no

Unlike DDR4, all DDR5 chips have on-die ECC, where errors are detected and corrected before sending data to the CPU. This, however, is not the same as true ECC memory with an extra data correction chip on the memory module. DDR5's on-die error correction is to improve reliability and to allow denser RAM chips which lowers the per-chip defect rate. There still exist non-ECC and ECC DDR5 DIMM variants; the ECC variants have extra data lines to the CPU to send error-detection data, letting the CPU detect and correct errors that occurred in transit.[23]
 
It's accurate alright.

Wrong.

But maybe in your head you misread...

All DDR5 has an ecc chip.

So what? The CPU knows nothing about it, the OS cannot log it and warn you and if a bit-flip occurs on the bus (between the memory and the CPU) then you are screwed. This is just a fake ECC, not a real one.
 
Wrong.



So what? The CPU knows nothing about it, the OS cannot log it and warn you and if a bit-flip occurs on the bus (between the memory and the CPU) then you are screwed. This is just a fake ECC, not a real one.
ECC is and always has been handled on the hardware level. You enable in bios, not windows.

Windows indicates memory errors to the user through WHEA.

However you can enable TLS ECC Curve in windows.


Correct A♤
 
ECC is and always has been handled on the hardware level. You enable in bios, not windows.

Who speaks about some idiotic windows? Never had that shit on my PC.

My Linux logs any ECC error through syslogd, so that I can see if that happens... Irrelevant where you enable it... Of course, this applies only to the real ECC, not to the fake DDR5 ECC which is internal to the DDR5 memory module - which is exactly the problem...
 
ECC is and always has been handled on the hardware level. You enable in bios, not windows.

Windows indicates memory errors to the user through WHEA.

However you can enable TLS ECC Curve in windows.


Correct A♤
You are referencing an article about Transport Layer Security, used to sign and verify SSL tickets over a network.

Elliptic Curve Cryptography is a replacement for RSA and not the same thing as Error Correction Code… It’s not even in the same ballpark.

Just stop this is embarrassing

The presence of on-die ECC on DDR5 memory has been the subject of many discussions and a lot of confusion among consumers and the press alike. Unlike standard ECC, on-die ECC primarily aims to improve yields at advanced process nodes, thereby allowing for cheaper DRAM chips. On-die ECC only detects errors if they take place within a cell or row during refreshes. When the data is moved from the cell to the cache or the CPU, if there’s a bit-flip or data corruption, it won’t be corrected by on-die ECC. Standard ECC corrects data corruption within the cell and as it is moved to another device or an ECC-supported SoC. (Thanks to Ian Cuttress for his explanation)


On-Die ECC (Error Correction Code) is a new feature designed to correct bit errors within the DRAM chip. As DRAM chips increase in density through shrinking wafer lithography, potential for data leakage increases. On-Die ECC mitigates this risk by correcting errors within the chip, increasing reliability and reducing defect rates. This technology cannot correct errors outside of the chip or that occur on the bus between the module and memory controller housed within the CPU. ECC enabled processors for servers and workstations feature the coding that can correct single or multi-bit errors on the fly. Extra DRAM bits must be available to allow this correction to occur, featured on ECC class module types like ECC unbuffered, Registered, and Load Reduced.


 
Last edited:
You have file systems that with ECC RAM can guarantee that you will never have data corruption or bit rot on a file. If you are going to use ECC RAM on FreeBSD or TrueNAS then you will have the most reliable and robust system to store any data. You can also use ZFS without ECC but then you have no guarantee that data corruption will not happen.

Unfortunately, ECC on Windows NTFS is not of much use.
Metadata checksum/ ECC NO

NTFS is 35 years behind ZFS in most technical areas so it's a bit odd that Microsoft isn't simply eliminating NTFS from their operating system and replacing it with one of the many superior options that have been around for decades.

The price of ECC wouldn't be a factor at all for me as I actually only have 117MB of active RAM on my daily driver which I've been using for over 3 years without reinstallation.
pekwm_screenshot-20221103T154346-1920x1080.png

I would have enough with one ECC RAM stick of 4GB.
 
This is so off-topic that I’m reluctant to post, but actually no, 4GB is not enough RAM for a ZFS server, unless you’re running less than 3TB, which would defeat the purpose of using ZFS.

Maybe I just don’t understand your posts honestly confused rn
 
In simple terms,
non-ECC DDR4 has no ECC,
non-ECC DDR5 has weak ECC.
 
Nah, ddr-5 has on-die ECC that does not communicate with the CPU and does not correct errors across the I/O, which really isn’t ECC at all.

sorry I’m bored
 
Nah, ddr-5 has on-die ECC that does not communicate with the CPU and does not correct errors across the I/O, which really isn’t ECC at all.

sorry I’m bored
I think of on-die ecc as chip defect mitigation.
 
Back
Top