LGA 1700 Alder Lake "Servers"

Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.

unwind-protect

Active Member
Mar 7, 2016
415
156
43
Boston
So did the CPU just die or did the board kill the CPU? ASUS is currently in hot water for killing the Ryzen 7000 X3D chips due to excessive overvolting... I guess we'll see if the replacement CPU will live for more than the 5 weeks the original CPU stayed alive.
Requires turning on EXPO, though.
 

AlOX

New Member
Mar 30, 2023
13
1
3
Before updating the Asus w680 Pro Bios to 2305, Does anyone knows how can i save the older bios? There isn't any downloadable file on Asus website.
If anyone thinks this update is buggy, I would appreciate an heads up.
 
Last edited:

AlOX

New Member
Mar 30, 2023
13
1
3
Update: It was a dead 13700K cpu. I was lead astray by the board not booting with the Celeron G6900, but the replacement board did also not boot (= quickly turns off after powering on) with either the G69000 or the 13700K and a replacement 13700K booted in both the original board and the replacement board.

So did the CPU just die or did the board kill the CPU? ASUS is currently in hot water for killing the Ryzen 7000 X3D chips due to excessive overvolting... I guess we'll see if the replacement CPU will live for more than the 5 weeks the original CPU stayed alive.
Sorry to hear about the dead CPU.
Maybe try this time to monitor after the SOC voltage.
 

steve0

New Member
Jan 19, 2023
10
7
3
Are you talking about the KSM48E40BD8KM-32HM modules? Are you using the first release BIOS 0203 or have you updated to 2305? What CPU are you using?

I have upgraded today from 2x32GB Kingston to 4x32GB Kingston and was expecting memory speeds to drop to 4400MHz (the official 2DPC speed for Alder Lake) but was surprised when the BIOS remained at 4800MHz (I had at no point changed any settings regarding memory speeds, so it's not overclocked or manually set to 4800MHz). Initial boot with 4 modules took some time (memory training!) but cold booting afterwards was back to normal speeds.

System survived ~50 minutes of MemTest86 Free v10.4b1000 after which I aborted the test and started the server again. Keeping my fingers crossed it won't crash or misbehave... I continue to be VERY happy with the ASUS W680-ACE.
I apologise for the very late reply. Yes, I am referring to KSM48E40BD8KM-32HM, and I'm using a 13900k.

I was on the initial release but just updated it to the latest. I had a similar experience to you, but I found that the memory controller was running at 1.35v and the DRAM at 1.3v when leaving everything on auto!

I've got the memory controller down to 1.11875v and set VDDQ and VDD to 1.1v. All appears stable so far at 4800MHz.
 
Last edited:

steve0

New Member
Jan 19, 2023
10
7
3
I apologise for the very late reply. Yes, I am referring to KSM48E40BD8KM-32HM, and I'm using a 13900k.

I was on the initial release but just updated it to the latest. I had a similar experience to you, but I found that the memory controller was running at 1.35v and the DRAM at 1.3v when leaving everything on auto!

I've got the memory controller down to 1.11875v and set VDDQ and VDD to 1.1v. All appears stable so far at 4800MHz.
I stand corrected. I'm failing Test 9 (Modulo 20, random pattern). Are the sticks bad, or is this an expectation with DDR5?

Everything is on auto, and I've bumped VDD & VDDQ a tad over 1.1v, but it won't pass.

i9-13900k
4x KSM48E40BD8KM-32HM @ 4800MHz
W680 ACE
 
Last edited:

Ivan Dimitrov

Member
Jul 10, 2016
52
5
8
40
Hi, none of the boards I checked officially support more than 128GB of RAM but some of the customer LGA1700 already do.
Has anybody tried 48GB ECC or non-ECC modules on any of the boards?
Thanks.
 

steve0

New Member
Jan 19, 2023
10
7
3
I stand corrected. I'm failing Test 9 (Modulo 20, random pattern). Are the sticks bad, or is this an expectation with DDR5?

Everything is on auto, and I've bumped VDD & VDDQ a tad over 1.1v, but it won't pass.

i9-13900k
4x KSM48E40BD8KM-32HM @ 4800MHz
W680 ACE
False alarm! I was using Memtest86 6.0 from a Proxmox 7.3 image which failed instantly at Test 9.

I'm running v10 now, and so far, so good. I will run this for 24 hours or so to be sure and provide an update. Phew!
 
Last edited:

VioletGiraffe

New Member
Apr 19, 2020
3
0
1
no ECC, not a server. ;-)
I have two questions about this.
1. Has anyone here actually ever encountered memory errors? I can't say I have, despite never running ECC-enabled hardware. But I'm not running a hundred servers 24/7 either, my experience is limited and not statistically significant.
2. What about the built-in ECC that every DDR5 module has? Judging from the specs it should be good enough, but maybe I misunderstand something.
 
Last edited:

unwind-protect

Active Member
Mar 7, 2016
415
156
43
Boston
I have two questions about this.
1. Has anyone here actually ever encountered memory errors? I can't say I have, despite never running ECC-enabled hardware. But I'm not running a hundred servers 24/7 either, my experience is limited and not statistically significant.
I once had a chipset chip, located under the RAM, heat up the RAM so much that errors came piling in. Without the notifications from ECC it would have scrambled my filesystems before I added a chipset fan.

2. What about the built-in ECC that every DDR5 module has? Judging from the specs it should be good enough, but maybe I misunderstand something.
No, that almost makes it worse. You have no reporting, so for all you know you have broken cells right from the beginning, all through your memtest runs, and you are sitting on a statistical time bomb.
 
  • Like
Reactions: VioletGiraffe

Stephan

Well-Known Member
Apr 21, 2017
920
698
93
Germany
1. Many times. The usual server experiences a handful of bit flips per year. The cause afaik is decaying radioactive atoms upsetting a bit which were already present since memory left the factory. Most don't even have a direct effect, because it only concerns data, not program code. Flips in program code might also be of no consequence, if the path is unused. Some times a module will gradually develop more errors. You will never know without ECC and a service like rasdaemon which can read and report error counters. I have come to rely on beefy Xeons (82xx) even at home, because their memory controllers are just so much more resourceful than small Xeons derived from desktop chips.

2. DDR5 wouldn't even work without on-chip ECC, because the structures have gotten so small even a slight disruption (bit flip from cosmic rays, decay, high temperatures, etc.) will cause data errors.
 
  • Like
Reactions: VioletGiraffe

drampelt

New Member
May 8, 2023
7
3
3
Just set up my new build with Asus PRO WS W680-ACE, i5 13500, and 2x 32gb Hynix HMCG88MEBEA.

Using dmidecode I see 80 total width which is a good sign, however, memtest is just showing N/A for ECC support. Any ideas what I could be missing here or things I could try to get (or confirm) ECC working?

memtest overview.jpg memtest detailed.jpg
 

RolloZ170

Well-Known Member
Apr 24, 2016
5,158
1,549
113
memtest is just showing N/A for ECC support. Any ideas what I could be missing here or things I could try to get (or confirm) ECC working?
are you using actual MemTest86 version with Raptor Lake support ?
BIOS up to date ?
 

VioletGiraffe

New Member
Apr 19, 2020
3
0
1
Why do Raptor Lake CPUs only support 128 GB of RAM? What will happen if you install 4x48 GB modules, will it refuse to POST?
 
Last edited: