Create the file /boot/loader.conf.local (if it doesn't exist) then add to it the line: amdsmn_load="YES" Steve

Run top -aSH at the command line and see what's using the CPU per core. It might be stuck at a power saving CPU frequency or something similar. Steve

@stephenw10 Thanks for the information. I tried that and it ended up making the performance worse, down into the low 700s on download. I tried a couple hacky things including trying to add someone elses BSD patch for muti-threaded PPPoE and that blew the whole thing up. I'm going to table this for the time being and I was able to improve the performance to ~890/930 by using OpenWRT with some tweaking. Thanks for all your help!

Yeah, cables have been changed actually, I know nothing lasts forever and sometimes is open box bad. However this is a first for me. AND exactly as you described it.

Ah, that's probably OK then. The defaults we yuse are set hiadaptive for all three. It depends what the BIOS is reporting which is not reliable. Powerd and apcupsd are not connected. Steve

@setarcos you are right, I dont know if there is a way to check this in runtime. I check the release notes etc. e.g. freebsd 12.2 release notes (https://www.freebsd.org/releases/12.2R/relnotes/) say: „The ixl(4) driver has been updated to version 1.11.29. [r363876] (Sponsored by Intel Corporation)“. There should be a directory where you can see all packages and their versions in a specific distribution, not sure how it is called, I am also new to freebsd. It is also possible to update the driver to not use the one in the kernel, but use a newer version as an external module, but I am not doing that considering it is a less tested scenario. Also pay attention not all NVM update paths are tested same or supported, so you might need to do more than one updates (x to y, y to z, instead of x to z) to be on the safe side.

@mr-rosh try some options on ali express https://www.aliexpress.com/popular/1u-network-firewall.html

Well, glad you were able to get installed eventually.

Ah, yes not much we can do then. At least that explains it. Steve

@nickehallgren i've been using solarflare 10g network cards with pfsense. they work great. you can find them for cheap on ebay. Don't about EU though.

@nogbadthebad said in Hardware for 10-25 Gbit/s WAN/NAT: A home user will never fully saturate a 25Gbps circuit. Sounds like a home user that needs to try harder. All you need is 25 computers running speed test at the same time... A YouTube video idea was just born.

@attilay2k I have tried a few Linux based firewall distros on Raspberry Pi over the years just out of curiosity. I have rejected all of them because the log files are useless when a large number of entries appear stamped 1 Jan 1970. The RPi does not have a battery backed Real Time Clock so it relies on syncing with a NTP source. However, quite often log entries are made before time sync has occurred. Yes you can add a RTCmodule to a Pi but this puts the cost up further. My preference is to use repurposed thin client terminals. There are a number of low power consumption 64-bit intel based thin client terminals that have been produced by various brands. My favourites use 1.3Ghz intel Atom 3825 dual core CPU with AES-NI. Used prices and models vary, but I often can buy one used and upgrade it with 32GB mSATA SSD, 4GB-8GB or RAM for less cost than a comparable RPi4. The RTC, AES-NI and SSD using ZFS make it far better than a RPi. The only downside is the number of Ethernet interfaces, usually just one. Up until 2.4.5p1 I have successfully used them in 24/7 use cases with USB Ethernet. However, with 2.5.2 and 2.6.0 now with ure driver, reliability on USB Ethernet has gone. Fortunately, for all but a few I have reconfigured them to use VLANs with a VLAN capable switch. You may find this link useful.

Most of those Sophos boxes are from Portwell/Caswell I believe. I'm not aware of any that had bypass NICs but I haven't looked that hard. However when you do have bypass NICs they generally still attach OK but just no traffic. You can see failures for NICs using odd PHYs, like attached to internal switches for example. I don't think that applies to the UTM 320. @Juanesptux You should probably start your own thread unless you are using the same UTM 320. Steve

@stephenw10 Thank you for the response and the explanation. While researching the I219-V problem I found some other forums where users where trying to resolve this issue by compiling their own drivers. I will try and find those posts and mention the 2.6 snapshot as a potential solution instead of compiling drivers and mucking with the kernel config.

@setarcos It's the 4 port QNAP card you first mentioned - the QXG-2G4T-I22 (I purchased it thru B&H). It's sitting in my Dell R220 running v2.5.2. Northbound it's connected to my Moto MB8611 modem and southbound to a Netgear M5300-52G for the time being (I haven't decided on a southbound switch yet). Everything seems to be working fine. I know I am not getting my full speed yet, but that will come in time. Once I disabled HW cheksum offloading only then could I pull a ipv6 address from comcast. I could ping NB ipv4 (e.g. 8.8.8.8), but not much else since DNS still wanted to resolve addresses as ipv6 and I didn't want to start making a lot of changes to the config to see if v4 worked on it's own.

Try these. I can't test them against hardware but they load fine in 2.5.2: [2.5.2-RELEASE][admin@252dev.stevew.lan]/boot/modules: ls bwi_v3_ucode.ko if_wg.ko linker.hints t4_tom.ko toecore.ko [2.5.2-RELEASE][admin@252dev.stevew.lan]/boot/modules: kldload t4_tom.ko [2.5.2-RELEASE][admin@252dev.stevew.lan]/boot/modules: kldstat Id Refs Address Size Name 1 15 0xffffffff80200000 3aea720 kernel 2 1 0xffffffff83f19000 1000 cpuctl.ko 3 1 0xffffffff83f1a000 2698 intpm.ko 4 1 0xffffffff83f1d000 b40 smbus.ko 5 1 0xffffffff83f1e000 344d8 if_wg.ko 6 1 0xffffffff83f53000 137b0 t4_tom.ko 7 1 0xffffffff83f67000 c7e toecore.ko t4_tom.ko.txt toecore.ko.txt Remove the .txt extension. Steve

@hijikta said in Pfsense & Intel card Need Help please: Thank you i will try it . You wrote it works under Win, did you encounter the wrong FW checksum in that environment?

Nothing specific at this time. All the parts need to be in place and tested before it can happen.

@utha Any luck?