PC Engines apu2 experiences
-
Have recently upgraded my apu2b4 to 2.4 RC.
Prior to the upgrade I used the flashrom utility to upgrade BIOS to v4.0.7.
For my configuration (LAN, WLAN + dual WAN) it was a smooth upgrade - kudos to pfSense team!
Some updated OpenSSL benchmarks, showing GCM performance improvement (ready for OpenVPN 2.4):
openssl speed -elapsed -evp aes-128-cbc type 16 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes aes-128-cbc 121916.58k 174865.28k 214507.52k 226282.50k 230230.70k openssl speed -elapsed -evp aes-128-gcm type 16 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes aes-128-gcm 43423.18k 113562.28k 276355.50k 358399.32k 431721.13k -
Hi. I'm coming late to the party and bought an APU2C4 board, running the latest 2.40RC based on FreeBSD 11.1. I'm using the full install on an mSATA board. BIOS v4.0.7
I read every post in the thread… How many of these tweaks (from the defaults) are still needed on the latest pfSense?
- Temperature probe changes
- PowerD
- Trim support
- Hardware TCP & large receive offload changes
- AES-NI changes
Also, do you recommend the 32-bit or 64-bit version for this platform?
Thanks!
-
Hi. I'm coming late to the party and bought an APU2C4 board, running the latest 2.40RC based on FreeBSD 11.1. I'm using the full install on an mSATA board. BIOS v4.0.7
I read every post in the thread… How many of these tweaks (from the defaults) are still needed on the latest pfSense?
- Temperature probe changes
- PowerD
- Trim support
- Hardware TCP & large receive offload changes
- AES-NI changes
Also, do you recommend the 32-bit or 64-bit version for this platform?
Thanks!
It;s a 64 bit processor, so what do you think? :)
-
Not sure. It only has 4GB RAM, and some 32-bit operating systems are faster.
How about those other things?
-
I'm using the full install on an mSATA board. BIOS v4.0.7
Would be also my personal choice and is also highly recommended from netgate in their Version 2.4.0 blog.
_A few noteworthy items about running or upgrading to 2.4:-
32-bit x86 and NanoBSD have been deprecated and are not supported on 2.4. Hardware capable of
running 64-bit images should be reinstalled with a 64-bit version. -
NanoBSD installs on 64-bit hardware should be reinstalled as a full installation.
-
32-bit x86 hardware can continue to run pfSense software version 2.3.x, which will
receive security updates for at least a year after 2.4.0-RELEASE._
I read every post in the thread… How many of these tweaks (from the defaults) are still needed on the latest pfSense?
You don´t need to setup tune or organize something, but you cold do so and that is more a additional option for the
admins and nothing you must be pressed to do.- Temperature probe changes
Do you need the temperature sensor?
- PowerD
Is the cpu scaling from the lowest bottom to the highest top as needed by default?
- Trim support
Is this a mSATA or SSD that supports TRIM?
- Hardware TCP & large receive offload changes
Is all running fine without any issues, like you was imagine it?
- AES-NI changes
Are you using the OpenVPN or IPSec VPN method?
Some need it, some want it and some are aware of doing something in that direction!
Also, do you recommend the 32-bit or 64-bit version for this platform?
Since version 2.4.0 there will be only the amd64Bit version and ARM images only!!!
- no NanoBSD
- no 32Bit OS image
- AES-NI support in the CPU is a must be since version 2.5.0
Not sure. It only has 4GB RAM, and some 32-bit operating systems are faster.
Today this information is really outdated, it was due to the changing between 32Bit and 64Bit hardware
where 64Bit hardware were brand new on the marked, but today you may better of using the 64Bit image
if you have real 64Bit hardware, due to many differences likes;- drivers will be not even ported back to 32Bit hardware (due to the lag of time and from the hardware side)
- nearly any new hardware will be 64Bit and is more capable as the older 32Bit technic
- since the ARM support other things are gone, or we all must waiting a longer time to get updates/upgrades
How about those other things?
pfSense 2.4.0 blog
_A few noteworthy items about running or upgrading to 2.4:-
32-bit x86 and NanoBSD have been deprecated and are not supported on 2.4.
Hardware capable of running 64-bit images should be reinstalled with a 64-bit version. -
NanoBSD installs on 64-bit hardware should be reinstalled as a full installation.
-
32-bit x86 hardware can continue to run pfSense software version 2.3.x, which will
receive security updates for at least a year after 2.4.0-RELEASE._
-
-
Not sure. It only has 4GB RAM, and some 32-bit operating systems are faster.
How about those other things?
It uses very little RAM. As my tag says APU2C4. I've never seen RAM usage over 30% and I use pfBlocker too, run my own web and mail servers too. 2Gb would probably be sufficient.
It works perfectly well with a default install. With the fact it supports AES you can enable that… or not, it's optional but enable it anyway.
As has been said already 32bit is deprecated anyway.
It's a really good device and takes only minutes to install and set up pfSense from scratch.
-
Do you need the temperature sensor?
Yep
Is the cpu scaling from the lowest bottom to the highest top as needed by default?
Not sure. It was talked about in this thread. Do I?
Is this a mSATA or SSD that supports TRIM?
Of course.
- Hardware TCP & large receive offload changes
Is all running fine without any issues, like you was imagine it?
Of course, but maybe it can run better, which is obviously why I asked.
- AES-NI changes
Are you using the OpenVPN or IPSec VPN method?
Of course, that's why I asked.
-
I noticed today that the APU2 makes a quick short beep every time I login via the GUI.
Have I just never noticed this or is it new in 2.4.0?
-
I noticed today that the APU2 makes a quick short beep every time I login via the GUI.
Have I just never noticed this or is it new in 2.4.0?
Mine always did this from SSH via the speaker on the computer running putty but it's new for me the speaker of the apu2 device now beeps kind-a cool really
-
I have to ask this but which is better between the two BIOS series releases for the APU2? 4.0.x or 4.5.x/4.6.x?
-
Never had issues with either of them.
-
Any ideas why I'm getting a board mismatch error:
Calibrating delay loop… delay loop is unreliable, trying to continue OK.
coreboot table found at 0xdffae000.
Found chipset "AMD FCH".
Enabling flash write... OK.
Found Winbond flash chip "W25Q64.V" (8192 kB, SPI) mapped at physical address 0x00000000ff800000.
This coreboot image (PC Engines:PC Engines apu2) does not appear to
be correct for the detected mainboard (PC Engines:PCEngines apu2).
Aborting. You can override this with -p internal:boardmismatch=force.As you can see in the logs, they're the same exact board so I'm not sure why there's a mismatch?
EDIT: Nevermind, I already figured it out. It seems to be caused by the space between PC and Engines so forcing it fixed the problem.
-
Any ideas why I'm getting a board mismatch error:
Calibrating delay loop… delay loop is unreliable, trying to continue OK.
coreboot table found at 0xdffae000.
Found chipset "AMD FCH".
Enabling flash write... OK.
Found Winbond flash chip "W25Q64.V" (8192 kB, SPI) mapped at physical address 0x00000000ff800000.
This coreboot image (PC Engines:PC Engines apu2) does not appear to
be correct for the detected mainboard (PC Engines:PCEngines apu2).
Aborting. You can override this with -p internal:boardmismatch=force.As you can see in the logs, they're the same exact board so I'm not sure why there's a mismatch?
whats the firmware filename and model apu you have?
-
@marjohn56:
Any ideas why I'm getting a board mismatch error:
Calibrating delay loop… delay loop is unreliable, trying to continue OK.
coreboot table found at 0xdffae000.
Found chipset "AMD FCH".
Enabling flash write... OK.
Found Winbond flash chip "W25Q64.V" (8192 kB, SPI) mapped at physical address 0x00000000ff800000.
This coreboot image (PC Engines:PC Engines apu2) does not appear to
be correct for the detected mainboard (PC Engines:PCEngines apu2).
Aborting. You can override this with -p internal:boardmismatch=force.As you can see in the logs, they're the same exact board so I'm not sure why there's a mismatch?
whats the firmware filename and model apu you have?
I have the APU2C4. See my edit notes above, I already fixed it.
-
Any ideas why I'm getting a board mismatch error:
Calibrating delay loop… delay loop is unreliable, trying to continue OK.
coreboot table found at 0xdffae000.
Found chipset "AMD FCH".
Enabling flash write... OK.
Found Winbond flash chip "W25Q64.V" (8192 kB, SPI) mapped at physical address 0x00000000ff800000.
This coreboot image (PC Engines:PC Engines apu2) does not appear to
be correct for the detected mainboard (PC Engines:PCEngines apu2).
Aborting. You can override this with -p internal:boardmismatch=force.As you can see in the logs, they're the same exact board so I'm not sure why there's a mismatch?
They're not exactly the same: somebody decided things would look nicer if they added a space, so now you're forced to override the safety check to install the firmware. Let the author know the missing space didn't really bother you.
-
yes, the force option, I did not need that, strange.
-
Any ideas why I'm getting a board mismatch error:
Calibrating delay loop… delay loop is unreliable, trying to continue OK.
coreboot table found at 0xdffae000.
Found chipset "AMD FCH".
Enabling flash write... OK.
Found Winbond flash chip "W25Q64.V" (8192 kB, SPI) mapped at physical address 0x00000000ff800000.
This coreboot image (PC Engines:PC Engines apu2) does not appear to
be correct for the detected mainboard (PC Engines:PCEngines apu2).
Aborting. You can override this with -p internal:boardmismatch=force.As you can see in the logs, they're the same exact board so I'm not sure why there's a mismatch?
They're not exactly the same: somebody decided things would look nicer if they added a space, so now you're forced to override the safety check to install the firmware. Let the author know the missing space didn't really bother you.
I must have gone from a different firmware version :)
-
Any ideas why I'm getting a board mismatch error:
Calibrating delay loop… delay loop is unreliable, trying to continue OK.
coreboot table found at 0xdffae000.
Found chipset "AMD FCH".
Enabling flash write... OK.
Found Winbond flash chip "W25Q64.V" (8192 kB, SPI) mapped at physical address 0x00000000ff800000.
This coreboot image (PC Engines:PC Engines apu2) does not appear to
be correct for the detected mainboard (PC Engines:PCEngines apu2).
Aborting. You can override this with -p internal:boardmismatch=force.As you can see in the logs, they're the same exact board so I'm not sure why there's a mismatch?
They're not exactly the same: somebody decided things would look nicer if they added a space, so now you're forced to override the safety check to install the firmware. Let the author know the missing space didn't really bother you.
Yes, I figured it out after a few seconds of Googling and stumbling upon your post in the PCEngines forum :)
@marjohn56:
yes, the force option, I did not need that, strange.
Probably because when you bought the device, you had a BIOS with version of 4.0.8 or above. I bet yours is pretty new.
-
Possibly, I have two APU's one that's live and one that's test. The live version runs 4.0.7, from the pcengines site.
The other has bounced around a bit, so maybe I did have to force it, or not…. can't remember. ::)
-
Has anybody had any luck getting the front LED's to work with pfSense?
