Internet causing lag in offline games?

[Jean-Francois]

Sup people!

 

I come here to ask about an issue I noticed.

Whenever I'll be playing a game, occasionally I'll get some random lag spikes, and it doesn't matter if it's online or not, of which game I'm playing.

 

And each time this happen, I get a spike in the Ethernet graph in Task Manager.

No spikes are seen on the GPU and CPU when this happens, it's only the Ethernet.

 

If any of you would help me with this that would be super appreciated.

My hardware is:

 

Nvidia Geforce GTX 1060 6GB

Ryzen 5 2600X

16 GB of Ram, G.SKILL Aegis DDR4 SDRAM 3000mhz

B450M GAMING PLUS Motherboard

Windows 10

 

I know that the first thing I should do is check out what's using the bandwidth, which I did. First it was Cortana which I have now completely wiped out of my system. 

It did improve the performance but I still get the same issue, just not as often. As a result, I assume the problem will persiste no matter what is using the internet.

[nextg3n]

could be something as simple as windows update... it is know for using maximum resources, id change the hours or disable it as much as i could...

[Jean-Francois]

12 hours ago, nextg3n said:

could be something as simple as windows update... it is know for using maximum resources, id change the hours or disable it as much as i could...

I just tried this, and the issue doesn't seem to be resolved unfortunately. 

[Nord]

You are most likely looking at a faulty driver or a faulty (or incorrectly wired) LAN cable.
 

I’d suggest that you:

1. Unplug your LAN cable from the PC and see if the issue persists.
   If it does not, try with a different cable.
    
2. Re-install (or update if installed from the mobo CD) your network drivers - or better yet all drivers while you are at it.
    
3. (Only if 1 & 2 did not fix the issue)
   Temporarily disable any antivirus software, windows defender, 3rd party firewall, etc. and see if there is a change in behavior.

 

 

FYI: you might want to ditch task manager for something along the lines of hwinfo, hwmonitor or afterburner as task manager is at most a vague generalization of what your computer actually does.

[Jean-Francois]

5 hours ago, Nord said:

You are most likely looking at a faulty driver or a faulty (or incorrectly wired) LAN cable.
 

I’d suggest that you:

1. Unplug your LAN cable from the PC and see if the issue persists.
   If it does not, try with a different cable.
    
2. Re-install (or update if installed from the mobo CD) your network drivers - or better yet all drivers while you are at it.
    
3. (Only if 1 & 2 did not fix the issue)
   Temporarily disable any antivirus software, windows defender, 3rd party firewall, etc. and see if there is a change in behavior.

 

 

FYI: you might want to ditch task manager for something along the lines of hwinfo, hwmonitor or afterburner as task manager is at most a vague generalization of what your computer actually does.

So I updated the network drivers. I tried playing a game with the Ethernet disconnected and turns out the problem is yet still present, so turns out it's not an internet issue.

After Installing MSI Afterburner and activating GPU and CPU display, I played The Witcher 3 max settings, and the CPU was around 40-60% usage, meanwhile the GPU was hitting 100% on some occasions and the lag spikes were happening at the same time. So I guess it's safe to assume that my GPU is just hitting it's maximum performances and that's what causing the lag.

 

As you said, Task manager seem to be extremely vague and wasn't giving me enough information. Afterburner gave me way more information and I know know it's my GPU. Thanks a bunch!

 

[nextg3n]

if im actually using userbenchmark right it should stay stable 45 fps with that setup at 1440p  max its kinda hard to believe that ur gpu cant handle it maybe gpu is brand new? or used(mining card perhaps?)

[Nord]

12 hours ago, Jean-Francois said:

So I updated the network drivers. I tried playing a game with the Ethernet disconnected and turns out the problem is yet still present, so turns out it's not an internet issue.

After Installing MSI Afterburner and activating GPU and CPU display, I played The Witcher 3 max settings, and the CPU was around 40-60% usage, meanwhile the GPU was hitting 100% on some occasions and the lag spikes were happening at the same time. So I guess it's safe to assume that my GPU is just hitting it's maximum performances and that's what causing the lag.

 

As you said, Task manager seem to be extremely vague and wasn't giving me enough information. Afterburner gave me way more information and I know know it's my GPU. Thanks a bunch!

 

Even at 100% GPU usage you should not get “lag spikes”.
We might not be exactly on the same page as to what a lag spike is exactly, but generally if the system hangs up shorty, freezes, generally stutters for more than half a second or acts in a similar manner, it's still not normal behavior. 

 

A 1060 should be able to run unmodded witcher 3 almost maxed (excluding shadows, hairworks off & motion blur off) on 1080p with near to no noticeable stuttering or other problems, while almost never dropping below 55~ FPS.
My 970 could do that already.

[Jean-Francois]

8 hours ago, Nord said:

Even at 100% GPU usage you should not get “lag spikes”.
We might not be exactly on the same page as to what a lag spike is exactly, but generally if the system hangs up shorty, freezes, generally stutters for more than half a second or acts in a similar manner, it's still not normal behavior. 

 

A 1060 should be able to run unmodded witcher 3 almost maxed (excluding shadows, hairworks off & motion blur off) on 1080p with near to no noticeable stuttering or other problems, while almost never dropping below 55~ FPS.
My 970 could do that already.

By lag spikes I meant stutter, sorry about that. I'm running it on ultra with hair works off. The games run smoothly at 60 fps, but stutter every 30 seconds or so.

Oddly enough, I don't recall having this issue with an old i7 950, before upgrading to Ryzen 5 2600X 

13 hours ago, nextg3n said:

if im actually using userbenchmark right it should stay stable 45 fps with that setup at 1440p  max its kinda hard to believe that ur gpu cant handle it maybe gpu is brand new? or used(mining card perhaps?)

My GPU isn't brand new. I bought it maybe 6 months ago from someone who had it for maybe a year. I find it weird as well, especially since I'm only running at 1080p 60hz

[Jean-Francois]

I also just Updated my BIOS to the latest drivers. There seems to be a bit of improvement, but I still get occasional stuttering during conversations. But in general Updating the BIOS seems to have helped a bit.

[nextg3n]

i guess it wouldnt hurt to check ur gpu clock speeds and compare them to factory specs.. maybe previous owner oc-d it abit to much...  but i am at a loss since you should be able to run it easily... or is it really an bad ssd? doubt that...  my biggest worry would be the used gpu if it has been run as a mining card before it can be pretty banged up....

[nextg3n]

what PSU are you using?

[nextg3n]

sorry for the multiple posts also you could or probably should check the cards thermal compound and probably re apply it or take it to a proffesional who will do it if you dont know how... could mean that the card starts to thermal throttle automatically  if it hits its peak temps since the cooling is not as affective as it can be

[Nord]

Stuttering every 30sec is definitely not normal.
 

Aside from checking temps & clock speeds (CPU for each core, GPU -> Core & Mem), like suggest already (Afterburner can do all of this) it might be worthwhile to also make sure that your RAM channels are configured properly in the BIOS. Dual channel for 2x sticks, quad for 4x, single for 1 stick and same that the RAM are placed in the correct slots as stated in the motherboard manual.
 

As stuttering is usually caused by
a.) simply overheating or very unstable clock speeds
b.) communication problems between the hardware.
c.) unstable overclocks / too low voltage for said overclocks

 

My hunch would be towards b.), hence the suggestion with the RAM as it seems to be the most logical cause.
Assuming that you would have stated if you OC’d anything and also, while using afterburner, would have checked the temperatures at least a bit.

[Jean-Francois]

On 10/3/2019 at 3:03 AM, nextg3n said:

sorry for the multiple posts also you could or probably should check the cards thermal compound and probably re apply it or take it to a proffesional who will do it if you dont know how... could mean that the card starts to thermal throttle automatically  if it hits its peak temps since the cooling is not as affective as it can be

 

On 10/3/2019 at 11:59 AM, Nord said:

Stuttering every 30sec is definitely not normal.
 

Aside from checking temps & clock speeds (CPU for each core, GPU -> Core & Mem), like suggest already (Afterburner can do all of this) it might be worthwhile to also make sure that your RAM channels are configured properly in the BIOS. Dual channel for 2x sticks, quad for 4x, single for 1 stick and same that the RAM are placed in the correct slots as stated in the motherboard manual.
 

As stuttering is usually caused by
a.) simply overheating or very unstable clock speeds
b.) communication problems between the hardware.
c.) unstable overclocks / too low voltage for said overclocks

 

My hunch would be towards b.), hence the suggestion with the RAM as it seems to be the most logical cause.
Assuming that you would have stated if you OC’d anything and also, while using afterburner, would have checked the temperatures at least a bit.

So I changed my graphics card thermal paste (was pretty dry to be honest), I switched around my 2 ram sticks with one another, I changed my A-XMP profile in the BIOS.

The stuttering seems to be gone so far. I need to make a few more tests, but I had time to test it for maybe 30 minutes and I have not encountered stuttering so far.

 

One thing I noticed though, is my 2 ram sticks seems to have different clock speed. One is running at 2933 mhz while the other one is at 2800 mhz. I'm currently using the 2800 mhz profil but it seems odd to have them at different speed, unless I'm missing something.

 

Anyway. The issue appears to be fixed so far. So I thanks both of you for your help, it's super appreciated for real. That stuttering was starting to get depressing :^)

[Nord]

I don't know much about the new AMD platforms, but even then I doubt that you should have different RAM speeds, however im not entirely certain how this can even be possible.

 

2933mhz is not an available RAM speed, the nearest ones are either 2800 or 3000 and even if you did base clock overclocking, reaching 2933 would require you to key in 104,8 BCLK, which is an oddly specific value. Though in actuality it 104,8 x28 would still be 2934,4 mhz... but that's the only sense I can make of it.

I know that AMD has a base clock OC that is or can be locked down onto the CPU multiplier only, so maybe - and this is entirely theoretical from my side -  if you got the RAM wrongly placed or configured (channel setup and/or physically placed into the mobo) the BCLK value carries over to the one stick the system thinks is there (or does not think there is) even though there are two.
This would be utterly wired to be even possible but you never know.

It might also be a simple reporting error, depending on where you read the RAM speeds.
I mean AMD GPU’s have known problems to report clock speed incorrectly or accept custom fan profiles, so this might not be too far fetched.
Personally, I would trust CPU-Z and/or AIDA64 reportings for memory speeds - for cpu-z always take the reported values x2, because Double Data Rate.

[Jean-Francois]

30 minutes ago, Nord said:

I don't know much about the new AMD platforms, but even then I doubt that you should have different RAM speeds, however im not entirely certain how this can even be possible.

 

2933mhz is not an available RAM speed, the nearest ones are either 2800 or 3000 and even if you did base clock overclocking, reaching 2933 would require you to key in 104,8 BCLK, which is an oddly specific value. Though in actuality it 104,8 x28 would still be 2934,4 mhz... but that's the only sense I can make of it.

I know that AMD has a base clock OC that is or can be locked down onto the CPU multiplier only, so maybe - and this is entirely theoretical from my side -  if you got the RAM wrongly placed or configured (channel setup and/or physically placed into the mobo) the BCLK value carries over to the one stick the system thinks is there (or does not think there is) even though there are two.
This would be utterly wired to be even possible but you never know.

It might also be a simple reporting error, depending on where you read the RAM speeds.
I mean AMD GPU’s have known problems to report clock speed incorrectly or accept custom fan profiles, so this might not be too far fetched.
Personally, I would trust CPU-Z and/or AIDA64 reportings for memory speeds - for cpu-z always take the reported values x2, because Double Data Rate.

I already tested it with CPU-Z and it properly showed as 1399.2 Mhz so it's fine. Thats why I also find the 2933 Mhz weird. It's properly physically socketed and to my knowledge in the BIOS as well. 

 

This is is pretty much my total extend in computer knowledge and I don't want to overclock my ram because I don't personally think it's worth it. I'll just stay at 2800 mhz for the current time. 

[Nord]

There is an order in which you need to slot the RAM sticks into the motherboard, like usually you got 4 RAM slots in total where 2 are color A and the other are 2 are color B, to know which one you need to use for dual channel ram setup, you need to look it up in the motherboard manual. This does matter for performance.

 

But yea, if CPU-Z says the RAM MHZ is on 2800 than I would take it as such - you can btw also look up if you are using dual channel in cpu z too.

 

 

FYI: Activating XMP (which I guess you meant with overclocking RAM) is not overclocking it's just a different memory setting. RAM have to follow the JEDEC standard, which requires them to work basically no matter what on all “settings”.
XMP does not follow this codec and can be enabled by userchoice, so it might not work in some systems, in some cases but almost always it will work. It doesn't change the power draw, voltage or anything else on the RAM, it just makes them “faster”, with the only drawback being that you can't be 99% certain that it works on every PC you try it on. From what I recall the magic number for ryzen is actually 3000mhz on ram, so might give it a shot, you can't really do any harm enabling it.

[Jean-Francois]

1 hour ago, Nord said:

There is an order in which you need to slot the RAM sticks into the motherboard, like usually you got 4 RAM slots in total where 2 are color A and the other are 2 are color B, to know which one you need to use for dual channel ram setup, you need to look it up in the motherboard manual. This does matter for performance.

 

But yea, if CPU-Z says the RAM MHZ is on 2800 than I would take it as such - you can btw also look up if you are using dual channel in cpu z too.

 

 

FYI: Activating XMP (which I guess you meant with overclocking RAM) is not overclocking it's just a different memory setting. RAM have to follow the JEDEC standard, which requires them to work basically no matter what on all “settings”.
XMP does not follow this codec and can be enabled by userchoice, so it might not work in some systems, in some cases but almost always it will work. It doesn't change the power draw, voltage or anything else on the RAM, it just makes them “faster”, with the only drawback being that you can't be 99% certain that it works on every PC you try it on. From what I recall the magic number for ryzen is actually 3000mhz on ram, so might give it a shot, you can't really do any harm enabling it.

My mobo only has 2 ram slots, so I can assure you my 2 ram sticks are in the proper slots lol. 

 

I had 2 profil for A-XMP by default. The first one was 2800 mhz, and the second one was 2933 mhz. At first I didn't use any and my rams where running at very low clock speed so I turned it on. First I tried the 2933 mhz, the profile on what my games were stuttering. I'm on the 2800 mhz profil now and so far everythings good. I can't say for sure if the profils are what made my games stutter since I changed more than one thing, but I will make some more tests tonight.

 

I wonder if it has something do do with the fact that, like I said, one of my ram sticks is detected as 2800 mhz and the other one at 2933 mhz. Maybe running then both at 2933 is what's causing issues? I'll check it out, and try to set it at 3000 mhz if I can. 

[Nord]

Could have actually been the XMP profile than, yes.
Though if 2933 is the XMP (utterly weird number though, but ok), you most likely won't be able to do 3000mzh, unless the RAM is actually advertised for that speed on the packaging. 

Though at this point I kinda have to tap out anyway as I dont keep up enough with ryzen platforms to know more than what I already stated previously on the topic.

 

If it were an Intel platform, I would lean towards broken memory that refuses to downclock or simply an incorrect BIOS settings - though I have no idea what one would need to do to achieve this.
So the easiest way to get around it would be to reset bios to defaults and start a-new basically or boot the PC up with one stick at a time (on the same ram socket) and see how it behaves, both performance wise and what speeds it does report. However if your system is now working as one would expect it to be, it probably is not worth investigating further