Intel 1226-V 2.5GbE Ethernet Chipset Showing Connection Drop Issues (Chipset Used on most Raptor Lake Motherboards)

[Zarathustra[H]]

im at the point of looking at 25g switches myself. id rather go that route then getting a 40g switch. But they are still expensive but id also like to learn ONIE.

Yeah. 40g is a dead end. This is an expensive hobby so I try to take small bits at a time. Since the NIC's are supposed to be backwards compatible, and I can benefit from a 25gig direct link instantly using fiber I've already run, getting two of them, one of the server and one for the workstation seems like the way to go.

Then when I upgrade the switch down the line, it will bump up the switched traffic to 25gig as well!

Mikrotik doesn't have SFP28 models yet though, and the other ones are a little bit too rich for my blood at this point.

 

[lopoetve]

I've considered picking up a pair of Intel XXV710-DA2's one for my server and one for my worstation.

One of the two ports on each would still have to work in 10gig mode though as I don't have a 25gbe switch, and likely won't for some time. The other port would be a direct link between the two at 25gig for some sweet sweet storage speed

I’ve got 16 ports of 25G (or 4x100), but the breakout cables even for those are absurdly priced. 😂😂. Use them for a backplane and 10G front-end storage out to the rest of the environment. Intel tends to downclock well. It’s the switch ports that are picky. And Broadcom nics

 

[LOCO LAPTOP]

How many of you are experiencing this issue?

To be back on topic...

Yes, in fact. I got my machine running last night. This morning I was transferring files and what would you know:

The description for Event ID 27 from source e2fexpress cannot be found. Either the component that raises this event is not installed on your local computer or the installation is corrupted. You can install or repair the component on the local computer.

If the event originated on another computer, the display information had to be saved with the event.

The following information was included with the event:

Intel(R) Ethernet Controller I226-V

The message resource is present but the message was not found in the message table

Sigh. Wifi, worked but even when reporting 1.6Gbps, it tops out at 500/600 going to the Asus AX11000 with a LAN cable to my other machine with 2.5G. (on the i225-V) Both drivers are from Asus website and I'm running Windows 10.

ill mess with it for a bit but it seems like ill dig out my old Intel PRO/1000 CT.

 

[Meeho]

People should start returning the motherboards and demanding Realtek ones.

 

[HockeyJon]

You know, between this and the constant failure with that Puma nonsense, maybe it’s time Intel stayed away from anything related to networking and the internet.

 

[jobert]

This was a nightmare to deal on a new build with troubleshooting and googling for days thinking it was just my mobo. Then the pain in the ass of removing the mobo and sending it back for another one to find out it was just the shitty intel port that comes on most mobos. Fuck Intel for wasting my time and causing so much stress and aggravation while trying to build a new pc. At least in the end I can just use Wi-Fi and get the same speeds since the router is right next to the pc.

 

[Lakados]

This was a nightmare to deal on a new build with troubleshooting and googling for days thinking it was just my mobo. Then the pain in the ass of removing the mobo and sending it back for another one to find out it was just the shitty intel port that comes on most mobos. Fuck Intel for wasting my time and causing so much stress and aggravation while trying to build a new pc. At least in the end I can just use Wi-Fi and get the same speeds since the router is right next to the pc.

Fortunately if you manually set the port speed to 1000 full duplex it will work fine. The signaling difference won’t let it auto negotiate.

 

[jobert]

Fortunately if you manually set the port speed to 1000 full duplex it will work fine. The signaling difference won’t let it auto negotiate.

I saw that mentioned at first but it has proven not to work at all by plenty of people including the guy running techpowerup. In fact not even two days later after trying that no web pages would connect at all and ethernet just dropped out. I am only going to use wifi now as that seems to work with no issues.

 

[German Muscle]

You know, between this and the constant failure with that Puma nonsense, maybe it’s time Intel stayed away from anything related to networking and the internet.

except that outside of that they have some of the best network cards.

 

[Mega6]

except that outside of that they have some of the best network cards.

Too bad these bad Intel nic chips have proliferated like cancer on the latest generation of MBs. They are literally everywhere.

 

[Lakados]

I saw that mentioned at first but it has proven not to work at all by plenty of people including the guy running techpowerup. In fact not even two days later after trying that no web pages would connect at all and ethernet just dropped out. I am only going to use wifi now as that seems to work with no issues.

That blows... I can say it worked for us here thinking back on it I wonder if I also had to set the advertised speeds on the switching ports? I might have. Doesn't matter now but it's sad that those chips and others like it will be plaguing MB's for a few years as they are given away by Intel and their partners to get them out of stock.
I wonder if this affects the Realtek network chips as well, those are also following the NBase-T spec but they look to have better 802.3bz support.
I wonder if going forward OEM's would start putting the ethernet controllers on M.2 keyed addon cards like they have wireless? would allow us to just replace them should we need.

 

[HockeyJon]

except that outside of that they have some of the best network cards.

Ok, well maybe they should do more of that and less of this?

 

[Lakados]

Ok, well maybe they should do more of that and less of this?

The problem is this is only a thing if you are connecting these Intel chips to Broadcom, specifically "cheaper" Broadcom-powered switches. Sadly, Broadcom is now in the majority of consumer enthusiast products. Intel is the one here getting done dirty because they used the better more open and compliant NBase-T standard while Broadcom used the more limited less popular MGBase-T standard. Broadcom just put their chips out cheap and in everything. Once you move up into the commercial and enterprise-class switches the Intel chips play much nicer as there you get proper 802.3bz support, but these MBs aren't being sold to people who are going to be running that sort of networking hardware, this is 100% an AIB problem as they are using the wrong hardware for the job, and they are making it our problem as a result. This is why you generally don't want enterprise parts on consumer equipment, they don't play the same game.
I 100% call out Asus, MSI, and Gigabyte as stupid for putting these chips on these boards, and Broadcom is just playing games leaving out proper 802.3bz on their consumer equipment.

 

[D-EJ915]

but these MBs aren't being sold to people who are going to be running that sort of networking hardware, this is 100% an AIB problem as they are using the wrong hardware for the job

This is the integrated mac low cost option from them so unless you mean use 1GB instead then yes it is the correct hardware. It just still comes down to that Intel needs to have better support and figure it out. We can talk about what is "better" all day long but in the end it doesn't matter if it doesn't work.

Since it seems nobody uses their enterprise switching hardware too they're discontinuing that, seems like their networking division can use some help.

2.87 dollars https://ark.intel.com/content/www/us/en/ark/products/210599/intel-ethernet-controller-i226v.html
4.86 dollars https://ark.intel.com/content/www/us/en/ark/products/210595/-intel-ethernet-controller-i226lm.html
7.02 dollars https://ark.intel.com/content/www/us/en/ark/products/210591/-intel-ethernet-controller-i226it.html

The prices are the same for the 225 chips.

 

[German Muscle]

This is the integrated mac low cost option from them so unless you mean use 1GB instead then yes it is the correct hardware. It just still comes down to that Intel needs to have better support and figure it out. We can talk about what is "better" all day long but in the end it doesn't matter if it doesn't work.

Since it seems nobody uses their enterprise switching hardware too they're discontinuing that, seems like their networking division can use some help.

2.87 dollars https://ark.intel.com/content/www/us/en/ark/products/210599/intel-ethernet-controller-i226v.html
4.86 dollars https://ark.intel.com/content/www/us/en/ark/products/210595/-intel-ethernet-controller-i226lm.html
7.02 dollars https://ark.intel.com/content/www/us/en/ark/products/210591/-intel-ethernet-controller-i226it.html

The prices are the same for the 225 chips.

Better support for what? They are supporting the appropriate 802.3bz standard. Its other vendors not using the correct standard.

 

[Lakados]

It just still comes down to that Intel needs to have better support and figure it out. We can talk about what is "better" all day long but in the end it doesn't matter if it doesn't work. To me it's wild this is now the 4th version of their 2.5GB chip and it still has issues.

Since it seems nobody uses their enterprise switching hardware too they're discontinuing that, seems like their networking division can use some help.

But it's not "technically" an Intel problem. NBase-T is supported by all major vendors it is the standard by which all (except Broadcom) 2.5, 5, and 10 Gb connections signal, MG Base-T is exclusive to Broadcom for their 2.5G and 5G connections.
The IEEE standard 802.3bz is fully backward compatible with the NBase-T signaling standard and any hardware that supports 802.3bz will work with the Intel chips.
Broadcom does not include 802.3bz support on their consumer or enthusiast class chips and only uses the MGBase-T support.
IEEE has gone and corrected this for everybody with the 1588v2 standard that supports N and MG, in addition to full 802.3bz making this a problem of the past as it will allow NICs and Switches to negotiate signaling rates, in addition to network speeds and gives far better tolerances for network latency. But you aren't going to find any controller capable of that unless you start looking at the 10Gb NICs.
This isn't an Intel problem to fix, this is caused by Broadcom playing proprietary games. The AIBs know full well that Broadcom is the absolute dominant provider of consumer networking equipment and they also know that Broadcom 2.5G and 5G does not play nice with anybody else.
Knowing that this is a thing the OEMs should never have included those Intel networking chips on their consumer product stack.
The OEMs should fix this by putting out a V2 of those boards with Broadcom chips and accepting RMA requests from affected users. An MG-based NIC will negotiate with an NBase switch, but an NBase NIC can't negotiate with an MG-based switch.

 

[RanceJustice]

Seems we've got some users who are real knowledgeable 10Gb+ networking hardware; I'll have to ask more about it on the networking subforum as I'll need some suitable upgrades.

That said, what do you think the chance is that Intel will be able to sort this out with a driver or firmware update? Just dug into the Asus X670E Crosshair Extreme's driver section on the website and apparently the Intel 2.5g is using a "225/226" driver so its likely to be one of those probably 226 given its newer; not sure which variant. I mean, I'll be able to use the AQuantia/Marvell 10Gb ethernet instead on that board at least, but there are a ton of even high end mobos that don't come with dual NICs and/or the 2.5Gb Intel chipset IS their fastest NIC. My NAS/home server box with a Asus X570 Crosshair Dark Hero has an Intel I211-AT 1Gb (which I hope is okay, given Intel's 1Gb NICs have always been pretty solid) and I knew the faster of its dual NIC was a 2.5Gb, but checking its actually a Realtek RTL8125-CG which seems to be decently supported (even on Linux, if you have a 5.x kernel it seems?). If it comes down to it and I get some new networking hardware, I may go about buying a new 10Gb NIC for my server box but for now at least I don't need to worry about the 2.5Gb thanks to it being a Realtek...unless it has some issues of which I am not aware.

Really hope Intel can get it sorted out. If some of the earlier conversation was accurate, I too wonder if OEMs will start putting mobo NICs on a M.2 cards; as I recalled a generation or so ago this was the case but only for a few models usually with 10Gb since those were always add-ins not part of the core chipset be it X570 or Z690 etc. Would kind of be nice if they would just get everyone switched over to 10Gb (or maybe even 25Gb) for consumer hardware kinda like how we had 1Gb as the "reasonably high end consumer hardware, but not bank breaking" standard for around 2 decades. There was a time when a gigabit NIC was seen as so insanely fast no home user could even conceive of a reason to use anything faster for general use, so though 25Gb or even 10Gb may not be prevalent now, it would be backward compatible and accessible when more users upgraded as time moves on.

 

[Lakados]

Seems we've got some users who are real knowledgeable 10Gb+ networking hardware; I'll have to ask more about it on the networking subforum as I'll need some suitable upgrades.

That said, what do you think the chance is that Intel will be able to sort this out with a driver or firmware update? Just dug into the Asus X670E Crosshair Extreme's driver section on the website and apparently the Intel 2.5g is using a "225/226" driver so its likely to be one of those probably 226 given its newer; not sure which variant. I mean, I'll be able to use the AQuantia/Marvell 10Gb ethernet instead on that board at least, but there are a ton of even high end mobos that don't come with dual NICs and/or the 2.5Gb Intel chipset IS their fastest NIC. My NAS/home server box with a Asus X570 Crosshair Dark Hero has an Intel I211-AT 1Gb (which I hope is okay, given Intel's 1Gb NICs have always been pretty solid) and I knew the faster of its dual NIC was a 2.5Gb, but checking its actually a Realtek RTL8125-CG which seems to be decently supported (even on Linux, if you have a 5.x kernel it seems?). If it comes down to it and I get some new networking hardware, I may go about buying a new 10Gb NIC for my server box but for now at least I don't need to worry about the 2.5Gb thanks to it being a Realtek...unless it has some issues of which I am not aware.

Really hope Intel can get it sorted out. If some of the earlier conversation was accurate, I too wonder if OEMs will start putting mobo NICs on a M.2 cards; as I recalled a generation or so ago this was the case but only for a few models usually with 10Gb since those were always add-ins not part of the core chipset be it X570 or Z690 etc. Would kind of be nice if they would just get everyone switched over to 10Gb (or maybe even 25Gb) for consumer hardware kinda like how we had 1Gb as the "reasonably high end consumer hardware, but not bank breaking" standard for around 2 decades. There was a time when a gigabit NIC was seen as so insanely fast no home user could even conceive of a reason to use anything faster for general use, so though 25Gb or even 10Gb may not be prevalent now, it would be backward compatible and accessible when more users upgraded as time moves on.

Very small, MGBase-T is proprietary, it is Broadcoms. Intel would have to reach out to Broadcom and work out an agreement there to implement it. I don’t see that happening not for something that has almost no solutions in the wild.

 

[NamelessPFG]

The problem is this is only a thing if you are connecting these Intel chips to Broadcom, specifically "cheaper" Broadcom-powered switches. Sadly, Broadcom is now in the majority of consumer enthusiast products. Intel is the one here getting done dirty because they used the better more open and compliant NBase-T standard while Broadcom used the more limited less popular MGBase-T standard. Broadcom just put their chips out cheap and in everything. Once you move up into the commercial and enterprise-class switches the Intel chips play much nicer as there you get proper 802.3bz support, but these MBs aren't being sold to people who are going to be running that sort of networking hardware, this is 100% an AIB problem as they are using the wrong hardware for the job, and they are making it our problem as a result. This is why you generally don't want enterprise parts on consumer equipment, they don't play the same game.
I 100% call out Asus, MSI, and Gigabyte as stupid for putting these chips on these boards, and Broadcom is just playing games leaving out proper 802.3bz on their consumer equipment.

I figured that you couldn't really expect enterprise networking gear to really integrate with consumer equipment to begin with because high-end networking quickly adopted SFP+, then QSFP and so on, and most switches don't even offer any sort of SFP port until you're spending hundreds, if not literal thousands of dollars. This isn't even getting into separate standards like Fibre Channel and InfiniBand that don't even see consumer adoption at all, which means you're often stuck buying retired older enterprise equipment to homelab with any of it.

This sounds more like subtle dickery on Broadcom's part, though, much like how I've read about anything with SFP+/QSFP ports often having vendor lock-in issues with the transceivers to the point that you can't just plug and play your transceivers into any bit of kit even though it should physically just work. Makes shopping for higher-end gear more difficult than it needs to be.

Broadcom is certainly capable of more overt dickery, though - just ask anyone scrambling to find VMware alternatives now after their recent acquisition and licensing price hikes.

 

[Lakados]

I figured that you couldn't really expect enterprise networking gear to really integrate with consumer equipment to begin with because high-end networking quickly adopted SFP+, then QSFP and so on, and most switches don't even offer any sort of SFP port until you're spending hundreds, if not literal thousands of dollars. This isn't even getting into separate standards like Fibre Channel and InfiniBand that don't even see consumer adoption at all, which means you're often stuck buying retired older enterprise equipment to homelab with any of it.

This sounds more like subtle dickery on Broadcom's part, though, much like how I've read about anything with SFP+/QSFP ports often having vendor lock-in issues with the transceivers to the point that you can't just plug and play your transceivers into any bit of kit even though it should physically just work. Makes shopping for higher-end gear more difficult than it needs to be.

Broadcom is certainly capable of more overt dickery, though - just ask anyone scrambling to find VMware alternatives now after their recent acquisition and licensing price hikes.

I dont want to get started on the VMWare stuff... My licenses are good till 2026 so I have at least that long to figure it out.

 

[TheSlySyl]

I recently got a few usb3 to 2.5gb adapters and I'm glad that I got all realtek ones. Honestly, mostly because they were cheap and I didn't want to spend the money on 10gb stuff.

I haven't had a single problem with any of them. Hit the full 2.5gb. Using Cat6a thin cable if that matters.

 

[chameleoneel]

Is it known if this is a problem when connecting to a nice Asus router, such as the AX86U? (that's what I have)

 

[erek]

TechPowerUp posted an article today regarding random disconnect issues with Intel's latest Ethernet chipset (1226-V) that most motherboard manufacturers are integrating on their Raptor Lake (Z700 series) boards. Here's a quote:



I checked the Event Viewer on my PC using the Asus Rog Strix Z790-F motherboard and I'm seeing the problem. Details on the particular error messages can be found in the article.

Techpowerup Article

How many of you are experiencing this issue?

Patched

https://www.techpowerup.com/305356/...-workaround-and-patch-for-ethernet-stuttering

 

[Zarathustra[H]]

Patched

https://www.techpowerup.com/305356/...-workaround-and-patch-for-ethernet-stuttering

Nice. Temporary patch disables a power saving feature that is only supposed to kick in at idle, but was kicking in when it wasn't supposed to, until such time as they can figure out why it is kicking in when it isn't supposed to, when further fixes might come (but those might require hardware changes)

Sounds if even before this driver was released, one could manually just disable EEE and it would stop the problems. If people had only been told, they wouldn't have had to wait for the patch...

 

[Mega6]

Patched

https://www.techpowerup.com/305356/...-workaround-and-patch-for-ethernet-stuttering

But not really fixed, if ever.

"but as this isn't a final solution to the problem, it's unlikely that this is the last we'll hear about the issue"

Just stay away from Intel i225/i226 multigig nics.

 

[erek]

But not really fixed, if ever.

"but as this isn't a final solution to the problem, it's unlikely that this is the last we'll hear about the issue"

Just stay away from Intel i225/i226 multigig nics.

Wonder if they managed to avoid any performance impact with this update. Would like to see if it was worked around cleanly with some benchmarks or if there was unrelated collateral damages in achieving this stability

 

[Zarathustra[H]]

Just stay away from Intel i225/i226 multigig nics.

Sounds to me like as long as you don't care about power saving (which I know a lot of us don't) you are fine.

 

[Mega6]

Sounds to me like as long as you don't care about power saving (which I know a lot of us don't) you are fine.

Maybe, maybe not. I have better things to do than to validate intel nics with years of problems. I avoid hardware with a tainted history like the plague.

 

[Zarathustra[H]]

Maybe, maybe not. I have better things to do than to validate intel nics with years of problems. I avoid hardware with a tainted history like the plague.

That is fair, but the alternative, Realtek (in most cases, not sure if Broadcom has made a netXtreme version of multigig) doesn't exactly have a bbrilliant track record either. For years, I've avoided anything with a Realtek NIC like the plague.

I suspect that now that they have figured out the root cause, the Intel i226 and i225 chips will probably be the more reliable choice.

I've never met a Realtek chip that didn't have problems, in just about every category (network, sound, etc.)

 

[Mega6]

That is fair, but the alternative, Realtek (in most cases, not sure if Broadcom has made a netXtreme version of multigig) doesn't exactly have a bbrilliant track record either. For years, I've avoided anything with a Realtek NIC like the plague.

I suspect that now that they have figured out the root cause, the Intel i226 and i225 chips will probably be the more reliable choice.

I've never met a Realtek chip that didn't have problems, in just about every category (network, sound, etc.)

I understand, Intel USED to have a flawless rep in nics and Realtek - we all know that mediocre rep they have. Well as far as multigig nics go, from everything I've been seeing - Intel i225/6 are nothing but a problem and the realtek ones work fine out of the box. Mind you we are talking two generations of intel nics with issues, not just one. That points to inherent design team issues/ hardware flaws.

 

[Lakados]

Sounds to me like as long as you don't care about power saving (which I know a lot of us don't) you are fine.

Maybe, maybe not. I have better things to do than to validate intel nics with years of problems.

Broadcom has a long ass history of not supporting what they call "Green Ethernet" between themselves and other NIC providers
They have it on all their 1G units, and their 10G and faster but it is pretty much absent from their 2.5/5G lineups along with proper PoE support there.
Broadcom only properly supports EEE on their newer 16nm 2.5/5G chips and those only went into production in 2020.
For the majority of their hardware, it straight up tells you to not let windows/mac manage your ethernet power state and to disable it in your NIC's preferences, and until recently by default OEM drivers usually left that feature as off as well.
The MGBase-T alliance wanted to create their 2.5/5G protocols as open source, where the NBase-T stuff is closed and licensed to high hell, so anything built around MGBase-T has a very different way of signaling when things are in a low power state and also does PoE signaling a little different which if you aren't expecting is an even bigger PITA.
Intel hardware is fine, the 2.5/5G standard is the thing that is messed up and it has been since the getgo, I understand why 2.5/5G is gaining in popularity now, but as business and enterprise basically skipped it, so support there was never an issue.
Even here 2.5G is the "new" hot topic and only because the networking companies have finally started releasing PoE powered 2.5G WiFi AP's. But only because Broadcom finally got onboard with their 2020 chip refresh to properly support the stuff so there is now enough qualified hardware out there to make it worth while.

 

[Lakados]

That is fair, but the alternative, Realtek (in most cases, not sure if Broadcom has made a netXtreme version of multigig) doesn't exactly have a bbrilliant track record either. For years, I've avoided anything with a Realtek NIC like the plague.

I suspect that now that they have figured out the root cause, the Intel i226 and i225 chips will probably be the more reliable choice.

I've never met a Realtek chip that didn't have problems, in just about every category (network, sound, etc.)

Broadcom at least has drivers and support...
Realtek is garbage the second you stray from whatever OEM configuration it shipped in, and even then you aren't safe.
Dell and HP for a solid year were shipping laptops with Realtek wifi cards that didn't have functioning drivers for the Ubuntu distro it shipped with, so unless you were running an open 2.4ghz network they would just fail to connect.
So yeah that was a major PITA... Will never willingly touch a Realtek nic of any sort, and I simply "try" to avoid them for audio, because really between BT speakers and USB headsets they don't really get used there for any actual audio work.

 

[Zarathustra[H]]

Broadcom at least has drivers and support...
Realtek is garbage the second you stray from whatever OEM configuration it shipped in, and even then you aren't safe.
Dell and HP for a solid year were shipping laptops with Realtek wifi cards that didn't have functioning drivers for the Ubuntu distro it shipped with, so unless you were running an open 2.4ghz network they would just fail to connect.
So yeah that was a major PITA... Will never willingly touch a Realtek nic of any sort, and I simply "try" to avoid them for audio, because really between BT speakers and USB headsets they don't really get used there for any actual audio work.

Damn. That's pretty shitty.

I mean, I'm used to factory wifi not liking linux (at least years ago, I've heard it has been better since)

I used to always replace the mini wlan card in every laptop I bought with ones I knew would work (usually Intel 7260) but those didn't ship with linux installed.

Way to go for actually testing that what you sell works together guys... 🙄

I'd imagine this is more difficult these days when everyone is soldering everything to the mainboard.

But maybe it's not as bad as I suspect. I don't have much experience with newer laptops. My current ones are both a decade or older, so I haven't had to deal with new ones

15" Latitude E6540 on the left with a Haswell era 4C/8T i7-4810MQ and a Radeon HD 8790M
14" Latitude E6430s on the right with an Ivy Bridge era 2C/4T i5-3320m.

Both of them feel just as snappy as anything modern and are still reasonably user maintainable, and have decent keyboards unlike anything modern!

(Though admittedly they are a bit bulky compared to newer laptops, and battery life is a hell of a lot shorter, even with a brand new battery. That and the 1366x768 screen on the 14" is pretty cramped by modern standards)

 

[wandplus]

I have a Gigabyte mobo with the B560 chipset, does it apply to this one as well? Because I'm having some issues.

 

[Mega6]

I have a Gigabyte mobo with the B560 chipset, does it apply to this one as well? Because I'm having some issues.

The motherboard chipset is irrelevant. It's the nic chip on the motherboard.

 

[German Muscle]

tbh EEE should be disabled anyway. Always causes problems.

 

[Mega6]

Several sources are saying disabling EEE is a "mitigation" and not a fix. Also claim that this is a hardware level bug and will need a new hardware revision to solve. Intel is STILL investigating the root cause.

 

[Mega6]

From TechPowerUp:

At the end of January, we reported on issues with Intel's i226-V 2.5 Gbps Ethernet controllers, which are used on a wide range of motherboards with both Intel and AMD CPUs, where users were having issues with stuttering and connections dropping. Intel has now released a workaround and a patch for the issue, not only for the i226, but also the i225 and the Killer E3100 2.5 Gbps network controller. What the workaround does, is disable the Energy-Efficient Ethernet mode, or EEE as it's also known as. EEE is only supposed to kick in when an Ethernet connection is idle and it's said to reduce power consumption by up to 50 percent.

However, in this case, it seems like EEE kicks in even when the Ethernet connection is active, which is causing the stuttering and connection dropouts. The patch disables EEE for all speeds above 100 Mbps, but it's also possible to disable EEE manually in the device settings in Windows.

 

[erek]

From TechPowerUp:

At the end of January, we reported on issues with Intel's i226-V 2.5 Gbps Ethernet controllers, which are used on a wide range of motherboards with both Intel and AMD CPUs, where users were having issues with stuttering and connections dropping. Intel has now released a workaround and a patch for the issue, not only for the i226, but also the i225 and the Killer E3100 2.5 Gbps network controller. What the workaround does, is disable the Energy-Efficient Ethernet mode, or EEE as it's also known as. EEE is only supposed to kick in when an Ethernet connection is idle and it's said to reduce power consumption by up to 50 percent.

However, in this case, it seems like EEE kicks in even when the Ethernet connection is active, which is causing the stuttering and connection dropouts. The patch disables EEE for all speeds above 100 Mbps, but it's also possible to disable EEE manually in the device settings in Windows.

Pending Class Action litigation over this?

 

[Mega6]

Pending Class Action litigation over this?

Doubtful - 99 percent of the functionality is there and it's an easy work around. Just makes we wonder what else may be lurking in these nics. Easy fix or not, I wouldn't touch them with a 10 foot pole.