Temperature Reading on AMD
-
Any performance figures for this interesting cpu setup? Routing throughput & openvpn?
I have a Netgear Nighthawk router plugged into the Fitlet pfSense box using bridge mode. I use that as my wifi access point.
Between two wired boxes connected locally there, using iPerf3, I get close to theoretical max throughput:
$ iperf3 -c fitlet -f m -n 1024M iPerf Client Connecting to host fitlet, port 5201 [ 5] local 192.168.x.x port 54552 connected to 192.168.x.x port 5201 [ ID] Interval Transfer Bandwidth [ 5] 0.00-1.00 sec 113 MBytes 947 Mbits/sec [ 5] 1.00-2.00 sec 112 MBytes 940 Mbits/sec [ 5] 2.00-3.00 sec 112 MBytes 940 Mbits/sec [ 5] 3.00-4.00 sec 112 MBytes 940 Mbits/sec [ 5] 4.00-5.00 sec 112 MBytes 940 Mbits/sec [ 5] 5.00-6.00 sec 112 MBytes 940 Mbits/sec [ 5] 6.00-7.00 sec 112 MBytes 940 Mbits/sec [ 5] 7.00-8.00 sec 112 MBytes 940 Mbits/sec [ 5] 8.00-9.00 sec 112 MBytes 940 Mbits/sec [ 5] 9.00-9.13 sec 14.4 MBytes 940 Mbits/sec - - - - - - - - - - - - - - - - - - - - - - - - - [ ID] Interval Transfer Bandwidth [ 5] 0.00-9.13 sec 1.00 GBytes 941 Mbits/sec sender [ 5] 0.00-9.13 sec 1024 MBytes 941 Mbits/sec receiver
I don't have a stellar Internet connection (Comcast, 170Mbps down/12Mbps up) but I'm able to get those speeds through pfSense with xfinity speedtest:
I have OpenVPN setup…I'll try and get on wifi somewhere else and do an iPerf while VPN'd in.
-
That test tells us the switching speed of your night hawk :)
If possible you should setup an additional interface on the firewall on its own subnet, plug a pc into that and measure with another pc on the lan/different subnet.
You could then probably enable an openvpn server on this new interface, connect to that from a pc/openvpn-client then do a speed test to a client on the lan. -
Just sent you a PM.
-
Using AT&T 4G, I VPN'd back home. I have iPerf3 running on another box internally and ran a test. I got 3.61 Mbits/sec.
Connecting to a wifi network with 10mbps up and down, I ran the test again and got 8.47mbps. I'll see if I can find a faster wifi network to test on.
**Update, got 17.9mbps connecting back to VPN on a friend's wifi with 40down/20up.
-
Just upgraded to 2.3.1 and temp widget is broken again… :(
redid your procedure and it's back... :) -
Just upgraded to 2.3.1 and temp widget is broken again… :(
redid your procedure and it's back... :)Thanks for the heads up, mine just did the same thing on 2.3.1 (not to be confused with 2.3_1 :o)
-
haven't done 2.3.1_1 yet…
-
I just did, and had to reload amdtemp.ko. I ended up writing a shell script to copy the files and enable the service.
-
Wish they would include your mod. 2.3.2 time….
-
Wish they would include your mod. 2.3.2 time….
Same. I think FreeBSD has to include it in their release though. Someone else already submitted a bug (https://lists.freebsd.org/pipermail/freebsd-bugs/2016-February/066274.html); not sure how to get FreeBSD to include the fix in a release though…
-
Fahrenhe1t, I'm running pfSense 2.3.2p1 w/ A6-5400K CPU with set to AMD and dashboard temps read 0C to 1C, varies. No where close to accurate. So tried your recompiled 10.3 amdtemp.ko file. Didn't change anything. My two other boxes with Intel CPU and A6-6400K work fine without a modified amdtemp.ko file. Ideas, mistakes I may have made?
-
Fahrenhe1t, I'm running pfSense 2.3.2p1 w/ A6-5400K CPU with set to AMD and dashboard temps read 0C to 1C, varies. No where close to accurate. So tried your recompiled 10.3 amdtemp.ko file. Didn't change anything. My two other boxes with Intel CPU and A6-6400K work fine without a modified amdtemp.ko file. Ideas, mistakes I may have made?
Strange! From the command line, what does the OS report if you run this: sysctl -a | grep temperature
-
dev.cpu.1.temperature: 0.0C
dev.cpu.0.temperature: 0.0CSame as before the modified file was copied to the box.
-
BIOS reports 33C
-
Do you know if the temperature is able to be read from a different operating system?
Maybe you could boot Ubuntu from USB on your pfSense box (without installing it), install the lm-sensors package, then run: sudo sensors
If it returns accurate CPU temperature, then you know it's a problem with FreeBSD not detecting your CPU properly.It looks like the A6-5400K CPU was released in 2012. If I had to guess, I bet FreeBSD (and the and specifically amdtemp.ko) doesn't detect the CPU properly. The CPU ID would have to be determined, the new sensor line would have to be added to the amdtemp source code, then it would have to be recompiled with the new CPU information. I'm just guessing though.
-
Thats a good idea but then I still have a broken temperature status. Your right the cpu is a bit old. I think I'll cut my time losses and try an A6-7470K. If it works I'll keep the 5400 as a back up.
-
Actually, I think the CPU is fine, it's just that amdtemp.ko isn't always updated with "new" CPU's. If you don't have any luck with the new CPU, I might be able to try and compile a new amdtemp.ko, you'd just have to tell me what the CPU ID is. In the source code, the CPU ID's look like this:
#define DEVICEID_AMD_MISC14 0x1703 #define DEVICEID_AMD_MISC15 0x1603 #define DEVICEID_AMD_MISC16 0x1533 #define DEVICEID_AMD_MISC16_X3 0X1583
The last one is the ID for my A10 Micro-6700T. I have no idea how that can be found though.
-
The newer cpu reports the same. I'll look for the DEVICEID.
-
The only ID I find is 630F81. Not sure where those txt strings come from in the amdtemp.ko file.
-
Hey, I found a thread which has updated DEVICEID's (http://www.pcengines.info/forums/?page=post&id=795B2ACC-F4B0-4181-9B4A-54EC757D4001&fid=DF5ACB70-99C4-4C61-AFA6-4C0E0DB05B2A). The guy (Stephan) said he compiled the latest version and uploaded it here: http://s000.tinyupload.com/index.php?file_id=70328254396185987242
Maybe you can try that one?