Really bizarre LAN chip/port issue

[Aquinus]

We, as forum regulars, must not assume everyone reading has the same level of knowledge. So what may be obvious to you, may
No it isn't!

I know there is confusing data out there but you cannot believe everything you read in Wikipedia. CMOS memory "devices" have been around and used in electronics for many years longer than the PC.

CMOS memory "devices" are volatile! If you remove the power from the device, the stored data is lost. That's the definition of volatile memory. That is exactly why CMOS "devices" were selected by IBM engineers in the original IBM PCs to store changes to the defaults set in the BIOS - so the BIOS could easily be reset.

CMOS "circuits", on the other hand, are non-volatile because, and as long as, they have a battery to keep the CMOS module powered, they will retain their data when the entire computer is powered off.

Computers are electronics. This is basic electronics any student of electronics knows. Sadly, many so called technical computer articles are not written by formally educated EEs or technicians and the authors of these articles don't know the difference between a CMOS device and a CMOS circuit.

If the CMOS was non-volatile as you proclaim, then by the definition of non-volatile, you should be able to unplug the computer and remove the battery and the CMOS data will still be retained - just as it is on a hard drive, SSD, or flash drive. But it's retained. In fact, it gets dumped almost instantly.

He had it backwards. NVRAM is a type memory done using a CMOS process. Flash memory usually is on a CMOS process. Therefore CMOS can describe both non-volatile and volatile memory. It can also describe your CPU, your GPU, RAID controllers, and so much more. It simply describes a process for building circuits by using both P and N type MOSFETs. When people in the past said "CMOS" they meant S(tatic)RAM which was built on a CMOS process. Stop trying to start an argument and get us off topic. Simple point is that the bios memory should have been reset on the OP's machine. That's done and doesn't need to be revisited.

For more info, see my previous post.

 

[Bill_Bright]

Flash memory usually is on a CMOS process

No it isn't! Google is your friend. What is flash memory?

Flash memory is a type of electrically erasable programmable read-only memory (EEPROM)

It can also describe your CPU, your GPU

Oh come on! PLEASE, do your homework. Use Google. CPUs and GPUs are NOT memory devices - CMOS, Flash or otherwise. IN NO WAY can CMOS be used to describe a "processor".

 

[Aquinus]

No it isn't! Google is your friend. What is flash memory?
Oh come on! PLEASE, do your homework. Use Google. CPUs and GPUs are NOT memory devices - CMOS, Flash or otherwise. IN NO WAY can CMOS be used to describe a "processor".

O rly?

Complementary metal–oxide–semiconductor (CMOS) /ˈsiːmɒs/ is a technology for constructing integrated circuits. CMOS technology is used in microprocessors, microcontrollers, static RAM, and other digital logic circuits. CMOS technology is also used for several analog circuits such as image sensors (CMOS sensor), data converters, and highly integrated transceivers for many types of communication. In 1963, while working for Fairchild Semiconductor, Frank Wanlass patented CMOS (US patent 3,356,858).

CMOS is also sometimes referred to as complementary-symmetry metal–oxide–semiconductor (or COS-MOS).[1] The words "complementary-symmetry" refer to the fact that the typical design style with CMOS uses complementary and symmetrical pairs of p-type and n-type metal oxide semiconductor field effect transistors (MOSFETs) for logic functions.[2]

BIOS memory was simply stored on SRAM because it was cheap and the battery already was there to drive the RTC so it could stay powered. Nothing more, nothing less.

Static random-access memory (SRAM or static RAM) is a type of semiconductor memory that uses bistable latching circuitry (Flip-flop) to store each bit. The term static differentiates it from dynamic RAM (Dynamic random-access memory) which must be periodically refreshed. SRAM exhibits data remanence,[1] but it is still volatile in the conventional sense that data is eventually lost when the memory is not powered.

...and...

Nonvolatile BIOS memory refers to a small memory on PC motherboards that is used to store BIOS settings. It was traditionally called CMOS RAM because it used a volatile, low-power complementary metal-oxide-semiconductor (CMOS) SRAM (such as the Motorola MC146818 or similar) powered by a small battery when system power was off (called the CMOS battery).

The term remains in wide use but it has grown into a misnomer: nonvolatile storage in contemporary computers is often in EEPROM or flash memory (like the BIOS code itself); the remaining usage for the battery is then to keep the real-time clock (RTC) going. The typical NVRAM capacity is 512 bytes, which is generally sufficient for all BIOS settings. The CMOS RAM and the real-time clock have been integrated as a part of the southbridge chipset and it may not be a standalone chip on modern motherboards.

However, this is all still off topic and as I said before, I will re-iterate again:

Stop trying to start an argument and get us off topic. Simple point is that the bios memory should have been reset on the OP's machine. That's done and doesn't need to be revisited.

Make a new thread if you want to start an argument, it doesn't belong here.

 

[Bill_Bright]

Make a new thread if you want to start an argument, it doesn't belong here.

Not trying to start any argument - just want the OP to have the 'true" facts and thanks now to your "...and..." above quote that shows:

(1) CMOS ram is "volatile" and,
(2) CMOS is incorrectly called "non-volatile" and,
(3) nonvolatile storage in contemporary computers is often in EEPROM or flash memory (like the BIOS code itself) - that is "NOT CPUs and GPUs as you previously claimed.​

So you have confirmed what I have been saying all along. Therefore, thank you! We are in agreement now and I agree that if YOU want to continue arguing that YOU start a new thread!

 

[Aquinus]

(1) CMOS ram is "volatile" and,

SRAM is on a CMOS process, it is volatile memory. NVRAM is on a CMOS process and is not volatile. When you say "CMOS ram" you're not being specific as it could refer to any number of types of circuits. You're right that SRAM is volatile.

(2) CMOS is incorrectly called "non-volatile" and,

No, CMOS is incorrectly used instead of the term SRAM. CMOS is not a thing, it's a process, a way of making circuits. People have improperly used the term "CMOS" when they really mean "SRAM".

(3) nonvolatile storage in contemporary computers is often in EEPROM or flash memory (like the BIOS code itself) - that is "NOT CPUs and GPUs as you previously claimed.

Read my last post.

Not trying to start any argument - just want the OP to have the 'true" facts

Except you are.

 

[Bill_Bright]

Except you are.

And you're not?

You said CMOS RAM in your reference. I quoted you now you accuse me of saying CMOS RAM as though that is wrong.

I don't care what you call it. The fact of the matter is, the CMOS, as it is used on an computer motherboard today is volatile. Why? Because if you remove all power from the device, regardless its compensation/architecture/construction/integration - it loses its data. And that, by definition, is volatile memory.

Now you can have the last word and continue your "non"-arguing . I will waste no more time on this and I am done here.

 

[OneMoar]

Aquarius don't they aren't worth the breath
CMOS is a process my post was pointing out that fact as he kept incorrectly referring to it as a type of storage
https://en.wikipedia.org/wiki/Nonvolatile_BIOS_memory
the battery just keeps the clock running on modern boards it doesn't supply power to anything other then the clock crystal,ever wonder why pulling the batter doesn't always clear settings its because the battery doesn't do anything the clr jmpr shorts the data pin to ground which cause its to erase ... tho someboards contain bios code to automatically re-set to default when the clock loses power(legacy trouble shooting support I guess lol)

either way the op doesn't have a software/firmware problem
what he has is a dead nic

 

[RejZoR]

It's not dead. It works. When it decides to work. And then works absolutely flawlessly until I shut down the system and leave it that way for few hours. But when it works, there is no packet or connection problems.

 

[OneMoar]

It's not dead. It works. When it decides to work. And then works absolutely flawlessly until I shut down the system and leave it that way for few hours. But when it works, there is no packet or connection problems.

that means exactly nothing anything from a bad ground to loss of capacitance to cold joints will cause that
my money is on a cold joint between two parts of the chip causing a spike in resistance and making things all wonky when its cold
you can check that by getting a hair dryer on low and pointing it at the lan-chip
if it gets better when warm then theres your problem
Further if by chance or whatever intermittent fault the chip has causes it to function long enough to get warmed up to a state where the fault mitigates then bing bang boom and bobs your uncle

 

[RejZoR]

The thing is, it will do this even if it was running for 3 days and you shut it down, cut power, wait 30 seconds and boot. It doesn't have to be shut off for hours.

 

[Mussels]

might be late to the party, but does the bios have any power saving/"green" options for the LAN? i've ran into boards in the past that these dubious power saving features would make LAN ports invisible to windows if it didnt detect a connection at POST, and some hardware combinations simply didn't work with that annoying 'feature'.

 

[RejZoR]

Nope, no such thing in BIOS at least. It's only LAN Disable/Enable control and ROM boot. Or do you mean the power saving settings for LAN in Windows Device Manager?

 

[Bill_Bright]

ever wonder why pulling the batter doesn't always clear settings its because the battery doesn't do anything the clr jmpr shorts the data pin to ground which cause its to erase ...

Removing the battery removes the voltage, right?

Well, the "clr jmpr" shorts the "power pin" on the same circuit to ground - thus removing the power from the device! In other words, it's the same thing! So to say the battery doesn't do anything is inaccurate.

Google any article for resetting the BIOS and they will all show the same 2, some times 3 ways. (1) Pull the battery, (2) use the CLR_CMOS (names vary depending on board maker) jumper and (3) with the reset option in the BIOS Setup Menu.

If you are right and the battery doesn't do anything, they why do all the tutorials say otherwise?

Can you show us link to even 1 article that says pulling the battery does not do anything? Or that pulling the battery and moving the CMOS reset jumper serve different functions?

Some, not all but some motherboards also have a Password Reset jumper. This is not the same as the CLR_CMOS function and should not be confused with that.

And note just for clarification for others still reading, the RTC is not a "clock" in the sense it keeps track of the time and date. It is a "counter". As you noted RTCs use a crystal oscillator but again, it does not keep time, it simply counts at a precise rate and the BIOS uses that "count" and translates and interprets that count into a date and time. This is why when you reset the BIOS, the date and time is set back to something else (often the date and timestamp of the last BIOS version).

 

[Bill_Bright]

might be late to the party, but does the bios have any power saving/"green" options for the LAN?

Not sure you can call them power saving/green options, but typically there are several "Wake on..." options that keep various features alive during standby. Typically these include Wake on Keyboard and Wake on Mouse so you can press a key, wiggle the mouse or press a mouse button to wake the computer. But some BIOS/chipsets allow you to enable Wake on LAN so the system will keep monitoring the network. This is a nice feature when unattended remote access is needed.

Hmmm, I thought it would be easy to find the power consumption of a NIC but it is not easy. I did find the specs for a professional 2-port 10Gb card was just 3.8W so I would imagine a PC NIC or integrated 1 port 1Gb card would be even less than that, especially when idle. So even if there was some "green" option, I don't think it would amount to much "green" left in your pocket after paying the power bill.

 

[OneMoar]

might be late to the party, but does the bios have any power saving/"green" options for the LAN? i've ran into boards in the past that these dubious power saving features would make LAN ports invisible to windows if it didnt detect a connection at POST, and some hardware combinations simply didn't work with that annoying 'feature'.

was one of the first things I suggested
also the PCI-E ASPM is known to cause such issues on some boards either way when a problem like this suddenly starts out of the blue its usually a hardware issue

 

[Mussels]

Nope, no such thing in BIOS at least. It's only LAN Disable/Enable control and ROM boot. Or do you mean the power saving settings for LAN in Windows Device Manager?

i meant in BIOS. If theres no options there, thats not what i was thinking of.

 

[RejZoR]

I think I'm onto something. Might be a random observation, but I hope I found the reason.

It's the "Plug & Play OS" option in BIOS (with YES and NO selection). I had this enabled ever since I can remember because lets face it, Windows 7 and up is a Plug & Play OS, however after reading the description of this feature in BIOS today in the morning, it got me thinking.

The setting says it gives the OS control over device detection on boot when set to YES and if you set it to NO, it gives BIOS control over device detection on system boot. Apparently Windows 8.1 at some point received an update that somehow breaks this. Which would explain why the problematic device works perfectly fine when it is detected but when it doesn't work, it appears like it doesn't even exist. Because OS failed to allocate its resources entirely on system boot (and that's why doing a cold boot often fixed it, because it re-initialized everything). It's purely theoretic observation that will require further testing, but it's a start.

I've also had situations where Sound Blaster Z would be just gone on one boot but after few restarts it would just magically appear without physically fiddling with the soudncard inside the case.

What I've also noticed is that after setting it to NO is that my CPU temperature dropped significantly. Before, testing stability with OCCT would result in almost instant 95°C at 4,2GHz with various stability issues even in AC cooled room to 25°C. Now, the temperatures were fluctuating at around 88°C (Linpack with 8 threads) at 29°C room temperature and I haven't changed anything related to cooling! Welcome to my bizarro land

Anyway, I'll be testing this and if I'm lucky, the disappearing LAN problem should be gone, meaning I won't have to replace my system and I'll be able to wait for Skylake-E and AMD Zen. Hurray!

 

[Mussels]

I think I'm onto something. Might be a random observation, but I hope I found the reason.

It's the "Plug & Play OS" option in BIOS (with YES and NO selection). I had this enabled ever since I can remember because lets face it, Windows 7 and up is a Plug & Play OS, however after reading the description of this feature in BIOS today in the morning, it got me thinking.

The setting says it gives the OS control over device detection on boot when set to YES and if you set it to NO, it gives BIOS control over device detection on system boot. Apparently Windows 8.1 at some point received an update that somehow breaks this. Which would explain why the problematic device works perfectly fine when it is detected but when it doesn't work, it appears like it doesn't even exist. Because OS failed to allocate its resources entirely on system boot (and that's why doing a cold boot often fixed it, because it re-initialized everything). It's purely theoretic observation that will require further testing, but it's a start.

I've also had situations where Sound Blaster Z would be just gone on one boot but after few restarts it would just magically appear without physically fiddling with the soudncard inside the case.

What I've also noticed is that after setting it to NO is that my CPU temperature dropped significantly. Before, testing stability with OCCT would result in almost instant 95°C at 4,2GHz with various stability issues even in AC cooled room to 25°C. Now, the temperatures were fluctuating at around 88°C (Linpack with 8 threads) at 29°C room temperature and I haven't changed anything related to cooling! Welcome to my bizarro land

Anyway, I'll be testing this and if I'm lucky, the disappearing LAN problem should be gone, meaning I won't have to replace my system and I'll be able to wait for Skylake-E and AMD Zen. Hurray!

that is utterly bizarre, although i've always wondered what disabling that option actually did.

 

[RejZoR]

This is the kind of crap that's turning my hair grey on daily basis lol. XD

 

[Bill_Bright]

lets face it, Windows 7 and up is a Plug & Play OS

XP and up, actually but took several years for the less than mainstream hardware industry to finally, and completely catch on.

FTR, I never change that setting from the default - which is typically "No". IIRC, this setting was added to the BIOS to allow, in part, for USB mice and keyboards when PS/2 devices were phasing out so you could access the BIOS Setup Menu during boot. The setting really has no bearing once the OS boots.

Apparently Windows 8.1 at some point received an update that somehow breaks this.

Pretty sure there was no Windows Update that did that. Perhaps a driver update changed or became corrupted. But if was a Windows update, there would be millions of users having similar issues.

Were you constantly getting "Found new hardware" whenever you booted?

http://www.tldp.org/HOWTO/Plug-and-Play-HOWTO-3.html

Before, testing stability with OCCT would result in almost instant 95°C at 4,2GHz with various stability issues even in AC cooled room to 25°C. Now, the temperatures were fluctuating at around 88°C (Linpack with 8 threads) at 29°C room temperature and I haven't changed anything related to cooling! Welcome to my bizarro land

That is bizarre! Is this the first time you checked these temps since replacing the battery/resetting the BIOS. If you manually changed the settings in the ASUS Fan Xpert control panel sometime in the past, they would have reverted to the defaults after a reset. And in that case, it would suggest default settings should be left as is.

That said, OCCT and LINPACK are different programs using different algorithms. You cannot compare between the two - only before and after with each. That is, before with OCCT and after with OCCT. Before with LINPACK and after with LINPACK.

 

[RejZoR]

Millions of users don't own Rampage II Gene motherboards...

OCCT has own torture and also includes Linpack...

 

[Aquinus]

that is utterly bizarre, although i've always wondered what disabling that option actually did.

Doesn't it default to disabled? I thought that was for booting OSes like OS/2...

 

[JunkBear]

If you dont mind about losing the gigabits just dump an older 100m/bits card with a realtek chip on it. Those with realtek chip on a pci slot are already having windows drivers. 5$ used

 

[RejZoR]

I have one, but having bunch of expanson cards in a microATX board with a super massive graphic card becomes an issue...

 

[Bill_Bright]

Millions of users don't own Rampage II Gene motherboards...

That's true. But there are millions of motherboards that use that same chipset and support the same CPUs so while each motherboard will have a unique BIOS, the differences, even among motherboards brands are slight. But to your point, if an update for W8.1 broke something with that particular motherboard, ASUS is very good at putting out updates. But still, Microsoft does not code to specific motherboards. It codes to standards so if something broke, then it would be because the board did not comply with standards. And if that were the case, there would be a whole lot of Rampage II Gene owners complaining about this problem. And I don't see it in the ASUS forums.

OCCT has own torture and also includes Linpack...

Even if in the same suite of tools, they are still different tests.

I thought that was for booting OSes like OS/2...

It really has nothing to do with operating systems today. That setting is all about what happens before the boot drive is even touched so the hardware (including legacy hardware) can communicate at a really basic level. Once the OS boots - whether Linux or Windows, their own protocols take over. At least that's my understanding.

It does not help that there are different meanings for PnP.

How PnP Works (simplified) - though I think simplified is a relative term here.