Thanks for that extra info! I checked this out and I’ve got some more ideas. It honestly looks like for some weird reason you’re getting packet floods. Can you create a WinMTR test which catches a freeze up. The fact that limiting client/server send rates seems to have changed the behavior to be worse makes me want to double down on connectivity issues being the most likely cause, especially since we’ve hit most of the hardware stuff already.
For that winMTR you want to restart the test every 10-15 minutes until you catch one of these problems. Once you’ve caught one red handed, run the test for about 5 more minutes. This will let us look for problems between your house and our servers which may cause this.
If you have problems with posting it due to a link error, go ahead and copy paste this into your next post, and replace “WinMTR goes here” with your test.
~~~~
WinMTR goes here
~~~~
By the way! You can turn the limits back off. If they didn’t help, they don’t need to be on still.
Edit2: I am really suspecting my PSU. My GPU (r9 390) is very power hungry. I’'m getting a new, more energy-efficient GPU the day after tomorrow. Hopefully that will fix it.
I normally suggest using the one from in-game on the netgraph instead of the support page. Should be in the top left corner. Should look like 8.8.8.8:12345 or similar. The part after the : colon is a port so don’t use that in the test, just the IP address.
That said, if limiting PPS did nothing it may be that the PPS spikes are actually a result of the freeze rather than the cause. In which case we’d want to look for system bottlenecks, software conflicts, or power issues.
I noticed you mentioned replacing the GPU. If there is still any problem after that please reach out to our support team directly so we may further investigate.
I used Asus DOCP to auto-OC my RAM to 3000mhz (which it’s rated at). I never saw this as a ‘real’ OC since I set the RAM to what they were advertised as. Sadly this can happen a lot from what I understand. I’ll have to set the autoOC loweror manually configure the timings and voltages.
I loaded the optimized defaults in my bios, which sets the memory at 2400mhz, and I didn’t have the freeze in 2 games. Will investigate further.
Edit: This could also mean it’s still a power issue, since OCing requires more power
It is the MSI R9 390. As you can see here it’s very power hungry:
www.tomshardware. com/reviews/gpu-hierarchy,4388.html
Today I’ll receive one of the best energy-efficient cards: the 1070. It requres like half the power in watt!
I’ll test my complete system with my RAM set at different clock speeds and keep you updated .
Edit: Well my shipment got delayedvso I’m not getting my GPU today… sucks
Ah I still have the freezes on bios default settings (so no OC on memory at all). They are just less frequent and a lot shorter. This strengthens the PSU-theory
It also means that physical RAM issues are possible. Not only would RAM without XMP be less power hungry, but it’d also be more stable. With all this in mind, have you run Memtest86+ ever? If you have a USB drive hanging around somewhere, you can run it overnight. You can also use the windows built in memory checker but it’s somewhat less accurate.
This is going a bit beyond Blizzard-supported, so we Blue posters probably have to duck out beyond this since it seems to be a hardware issue, but between some tests on your own and possibly a PC tech you should be able to nail down the common cause and replace whatever’s not working correctly.
I ran memtest86+ 2 times. I posted about i in this topic. Once on 3200 mhz, that gave me a lot of errors, and once on 3000, no errors there.
EDIT:
Ok so I got my new GPU:
( Gigabyte GeForce GTX 1070 G1 Gaming (rev 2.0)
and I still have freezes/hiccups. I know for sure now that power/PSU was probably not the issue. This GPU is very energy efficient.
The freezes/hiccups are a bit shorter and less annoying though. I get around 160/170 fps and the fps just drops to 110/120 during the freeze. I ran WinMTR correctly this time, using the IP adress from the menu, please tell me if you see anything weird::
Sorry about that, looks like I missed your earlier RAM tests in this big ol’ thread.
It looks like your packet loss is back. I’m somewhat curious if this is something with your router firewall or something like that. We can test that by installing a VPN and seeing if it improves. Can you find a VPN with a free trial, install it, and see if the issue is still there when running the VPN? You may find that your overall ping is higher, but it’ll let us know if it’s an ISP or Router Firewall issue if it works around the problem.
No need to apologize you all have been very helpful so far. <3
You mention that my packet loss is back, can you see that from the winMTR statistics? how? I’d like to be able to read the WinMTR statistics.
You also advice to try gaming with VPN enabled. No problem, I’ve got GhostVPN. I will try out a couple of games with VPN enables and post the WinMTR results
EDIT
Still had the freeze with VPN enabled. I was on a network that was still in the Netherlands:
These nodes are experiencing at least 1% data loss. Even though one node is showing 2%, that 2% doesn’t continue throughout the test and is likely set up to ignore some of the traffic that WinMTR uses to test the network.
The new test looks really laggy, with a worst ping over 300ms, but no data was lost.
This is the best guide I know of for learning how to read them
According to many ISPs (in the US at least), 2-5% is sometimes acceptable and doesn’t qualify for repairs. To an fps gamer, 1% is not good and can cause you to miss shots.
I’ve had some contact with blizzard support. Seems like they have no clue either, really sucks. It actually got really worse to a point where the freezes occured 2 seconds long.
I installed UI for ETW as per suggestion in this thread. No idea how to read anything in there though. I recorded a trace each time a freeze occured: (remove the space before .com)
A while back someone else was getting microfreezes across their whole OS, and they said ILSC (intelligent standby list cleaner) fixed it for them. It’s a tool made by the same person who creates Display Driver Uninstaller.
Unfortunately, this requires downloading a ZIP file, and I don’t download files from anyone outside my co-workers
I didn’t realize this was data from the app the other player mentioned. I am not familiar with that software, and won’t be able to answer any questions about its output data.
I took a peek at your ticket, and I think we should head back to hardware. You got improvements from downclocking your RAM. In each of your screenshots, you have one large spike of issues (that white vertical bar) which coincides with “Packets Per Second” spikes. Now, typically we’re going to look at a connection error for this, but in very rare and pesky situations, the “Packets per second” will actually stop being processed by your PC rather than received or sent by the connection. With the MSinfo on that ticket, we get a much clearer picture. I noticed that way back in the day you had some bluescreens that actually called out your RAM having hardware issues. Doom and Overwatch would both use more RAM than many of the other games I see you playing on that PC in the MSinfo.
0x1a_61941_PAGE_TABLE_RESERVED_BITS_SET_IMAGE_hardware_ram, type 0
Event Name: BlueScreen
Let’s go a little harder on this step and see if things improve further. Do you have better luck if you set your RAM back to fully STOCK speeds? (i.e. no XMP or overclocking on them at all.)
.
