All Haswell users on Mavericks - Ping issues

[bp1000]

Do you guys have a ping issue and possibly a throughput issue?

I have a Haswell iMac and Macbook pro and both machines exhibit really high pings over wifi.

Open a terminal and try... e.g. :

press: cmd + spacebar
type: terminal
press: enter key
type: ping http://www.google.com

my typical output from my adsl connection

PING www.google.com (173.194.34.84): 56 data bytes
64 bytes from 173.194.34.84: icmp_seq=0 ttl=58 time=30.052 ms
64 bytes from 173.194.34.84: icmp_seq=1 ttl=58 time=133.302 ms
64 bytes from 173.194.34.84: icmp_seq=2 ttl=58 time=43.235 ms
64 bytes from 173.194.34.84: icmp_seq=3 ttl=58 time=65.324 ms
64 bytes from 173.194.34.84: icmp_seq=4 ttl=58 time=89.182 ms
64 bytes from 173.194.34.84: icmp_seq=5 ttl=58 time=112.213 ms
64 bytes from 173.194.34.84: icmp_seq=6 ttl=58 time=32.163 ms
64 bytes from 173.194.34.84: icmp_seq=7 ttl=58 time=54.748 ms
64 bytes from 173.194.34.84: icmp_seq=8 ttl=58 time=77.396 ms
64 bytes from 173.194.34.84: icmp_seq=9 ttl=58 time=101.273 ms
64 bytes from 173.194.34.84: icmp_seq=10 ttl=58 time=123.731 ms
c64 bytes from 173.194.34.84: icmp_seq=11 ttl=58 time=44.756 ms
64 bytes from 173.194.34.84: icmp_seq=12 ttl=58 time=66.446 ms
64 bytes from 173.194.34.84: icmp_seq=13 ttl=58 time=90.136 ms
64 bytes from 173.194.34.84: icmp_seq=14 ttl=58 time=112.819 ms

This is disappointing. I suspected wifi interference because i ping my routers ip it is the same high pings ruling out my ISP. So if it's local maybe my wifi is the problem. Tried all channels.

Switching to ethernet solved the issue, but i also have windows 8 on this iMac and i tested wifi ping and it is a solid 29ms - 31ms regardless of how long i run it for. So now it points back to a mavericks issue. If it is a mavericks issue, more people should see the problem.

My Macbook Pro 13" is exactly the same. So i don't think it is a hardware problem (2 different machines have the problem, but ok in windows).

Someone suggested this might be a power saving feature but my throughput is behaving erratically too.

thanks

 

[fyzle]

I wrote an app to fix this.

I had the same issue, went to an Apple retail store to get the wifi card replaced, and that didn't fix it!

So, I wrote an app to fix it using solutions / workarounds posted in various forums.

https://gum.co/PingFixer

Hope you like it!

Kevin

 

[georgegates5]

Do you guys have a ping issue and possibly a throughput issue?

I have a Haswell iMac and Macbook pro and both machines exhibit really high pings over wifi.

Open a terminal and try... e.g. :

press: cmd + spacebar
type: terminal
press: enter key
type: ping http://www.google.com

my typical output from my adsl connection

PING www.google.com (173.194.34.84): 56 data bytes
64 bytes from 173.194.34.84: icmp_seq=0 ttl=58 time=30.052 ms
64 bytes from 173.194.34.84: icmp_seq=1 ttl=58 time=133.302 ms
64 bytes from 173.194.34.84: icmp_seq=2 ttl=58 time=43.235 ms
64 bytes from 173.194.34.84: icmp_seq=3 ttl=58 time=65.324 ms
64 bytes from 173.194.34.84: icmp_seq=4 ttl=58 time=89.182 ms
64 bytes from 173.194.34.84: icmp_seq=5 ttl=58 time=112.213 ms
64 bytes from 173.194.34.84: icmp_seq=6 ttl=58 time=32.163 ms
64 bytes from 173.194.34.84: icmp_seq=7 ttl=58 time=54.748 ms
64 bytes from 173.194.34.84: icmp_seq=8 ttl=58 time=77.396 ms
64 bytes from 173.194.34.84: icmp_seq=9 ttl=58 time=101.273 ms
64 bytes from 173.194.34.84: icmp_seq=10 ttl=58 time=123.731 ms
c64 bytes from 173.194.34.84: icmp_seq=11 ttl=58 time=44.756 ms
64 bytes from 173.194.34.84: icmp_seq=12 ttl=58 time=66.446 ms
64 bytes from 173.194.34.84: icmp_seq=13 ttl=58 time=90.136 ms
64 bytes from 173.194.34.84: icmp_seq=14 ttl=58 time=112.819 ms

This is disappointing. I suspected wifi interference because i ping my routers ip it is the same high pings ruling out my ISP. So if it's local maybe my wifi is the problem. Tried all channels.

Switching to ethernet solved the issue, but i also have windows 8 on this iMac and i tested wifi ping and it is a solid 29ms - 31ms regardless of how long i run it for. So now it points back to a mavericks issue. If it is a mavericks issue, more people should see the problem.

My Macbook Pro 13" is exactly the same. So i don't think it is a hardware problem (2 different machines have the problem, but ok in windows).

Someone suggested this might be a power saving feature but my throughput is behaving erratically too.

thanks

I once did a university project and I realising pinging from Mac Os X (Was doing my project on Mountain Lion than) was way slower compare to pinging on Windows.

Windows will receiving a response at a lesser time(faster response) compare to Mac OS X. Not so sure why but according to my lecturer it is due to the way the OS was written.

I guess there nothing we can do about it but raise this issue to apple engineers.

 

[CarreraGuy]

If you are testing your wifi you should ping the first "hop" in your network. Typically home routers have a private LAN ip of 192.168.0.1, but you can figure this out by looking at the first entry of your traceroute results to yahoo, google, etc.

Then ping the first address you see in the traceroute results. This will give you a better idea of your wifi connection.

When I ping my local router with a wired connection I see results like:
PING 192.168.0.1 (192.168.0.1): 56 data bytes

64 bytes from 192.168.0.1: icmp_seq=0 ttl=64 time=0.752 ms
64 bytes from 192.168.0.1: icmp_seq=1 ttl=64 time=0.665 ms
64 bytes from 192.168.0.1: icmp_seq=2 ttl=64 time=0.683 ms

On my laptop via wireless it varies and I've seen timeouts and some round trip times > 100.00 ms.

 

[f54da]

I'm glad you posted this, I thought I was going crazy trying to debug this. I can debug finer by using "ping -i 0.1" to set ping frequency, and it basically shows that the latency spikes occur predictably every second or so (fancier people can take the output and do a timeplot and then fourier transform to get precise frequency, but I'm lazy). Wifi adapter going to sleep makes sense as the cause.

This also makes ssh'ing into the thing unbearable.

(Note this is on mavericks, on haswell-era hardware. Issues persist in safe mode, so rules out OS issues. Haven't tried running linux distro on this to see if it's a networking stack issue or not. Alternatively trying a usb wifi dongle would also work to isolate cause.)

 

[chrfr]

I'm glad you posted this, I thought I was going crazy trying to debug this. I can debug finer by using "ping -i 0.1" to set ping frequency, and it basically shows that the latency spikes occur predictably every second or so (fancier people can take the output and do a timeplot and then fourier transform to get precise frequency, but I'm lazy). Wifi adapter going to sleep makes sense as the cause.

This also makes ssh'ing into the thing unbearable.

(Note this is on mavericks, on haswell-era hardware. Issues persist in safe mode, so rules out OS issues. Haven't tried running linux distro on this to see if it's a networking stack issue or not. Alternatively trying a usb wifi dongle would also work to isolate cause.)

These computers are compatible with macOS all the way up to Big Sur. Do the newer OSes solve the problem, and is there a reason you’re still using Mavericks in 2022?

 

[f54da]

@chrfr because I like mavericks, and newer OSs have much worse issues. Also note that people here are talking about egress ping, while I only observed it on ping packets addressed TO the machine. So it really is less of an issue, since the only situation where this becomes an issue is in low-traffic connections where ingress latency is important like SSH. (I assume the NIC firmware or whatever power control mechanism was updated to prioritize egress packets, so the only case in which power-save affects now is on incoming packets).

I'd be curious to know if newer OSs or using linux solves it or not. Anyone with a haswell era machine want to try pinging their laptop and posting results? Ideally once in internet recovery (which I think should boot into 10.9 even if it was updated?) and once in whatever latest OS it's running.