Quirky Xerus 8.1.4 for Raspberry Pi2 and 3

[daturach]

@BarryK
Thanks for your quick answer.

I have a general question about these puppies based on DEB packages. I should be able to install any apps through the Package Manager, right?
I tried to install apt-get. Well, did not manage to make it work. Here is the message:

If that's something easy to fix, let me know! Thanks

A couple of thoughts about this...

If you chose to update the package database in the Puppy Package Manager, that might be the cause. There might be an update of that lib that is causing trouble.

Quirky has it's own package manager, Puppy Package Manager. The Debian package management database and infrastructure is not there, apt-get probably won't work properly.

You are right and I see the problem now...I managed to get rid of this error by installing a newer library (libapt-pkg5.0_1.2.15_armhf.deb) BUT I need now to create a new user (_apt). This is not going to lead me anywhere because adduser command doesn't exist (single user distro! )
End of game!

[don570]

So nothing is wrong. You wrote an 8GB image to a 16GB drive, leaving about 8GB empty space. QuickSetup offers to fix this at first bootup.

In the start menu there is a menu item to run the first run setup again,
even if you have already setup your system, so you can expand the useable space to be greater than 8 GB.

____________________________________________________

[slavvo67]

Need to get me a Pi for the holidays. I'm really missing out, here!

[TeX Dog]

lol, right its really coming together with BKs vision, was reviewing stuff in the new kernel and its got so much neat stuff support compiled in and just waiting for use, the software defined radio stuff is there, joysticks, media like cdroms support there. I now can run puppy and download images from overhead satellites, track airplanes nearby etc. all those cool project that once needed a bloated RaspianOS are in sight.

[Sage]

Have powered up RPi3 with Quirky. It works, but:
Using an experimental layout of HW it was necessary to switch off - hard reset - a couple of times. This upset the booting procedure much more than expected, including a complete reset. During this stage, error messages overwrote the screen in several places making the underlying text unreadable. Something got written to the SD card, too which eventually rectified itself (I think). In my opinion, which may not be shared, I would like that an hard reset never writes to the SD card; maybe not possible?
After booting, there were too many options to run in the Quick Setup. In particular, selection of the video output default caused problems and freezing before it could be changed. Suggest the default is recast to always be AUTO.
There were some minor issues running the browser at first (not 'first run' related) but these also seem to have been cleared.
It was good to find Hardinfo retained because this is vital for HW experimenting. Thank you.
I fixed a small HS temporarily on the cpu, but it showed 50C in free air (no fan - yet). This confirms my recommendations above, viz. this system needs an heatsink AND fan combo, the former being affixed with a silver-loaded adhesive paint.
Otherwise, an interesting project all round.
For the record, I used a PSU like the one illustrated by BK on his blog and attached it directly to pins 4 & 6. This is an high quality switched-mode item by Cisco so I'm trusting that bypassing the on-board resettable fuse probably isn't going to be an issue.
There was some 'hash' on some areas of screen, hdmi, which did not go away. My hdmi-vga converter, sold for use on RPi B+ didn't work - it does on B+ and Zero.

Now, I'm away to get my Christmas lights coupled to a motion detector and the door video-phone to activate using the GPIO pins.

Sad!

PS: Please can we have a completely updated image. Some of us aren't too hot fixing SW!

[TeX Dog]

@sage I had similar but not as bad experience, check your fstab file in /etc for a malformed mount of root, BK says the upgrade does not touch fstab ( which I did check, he is right OF COURSE, but its been malformed since 8.1.1 ) the new kernel has a issue with that where last versions somehow worked, ( I can't see how /dev//sdb2 never existed for me ) with new kernel all versions back to 8.1.1 freeze unless changed.
Second, it is possible to run without sdcard, with either two methods I posted for you a few pages back, Dons ideas and Patks. My method is easiest, and moves root to a usb spinning harddrive so boot sdcard is removable.
On the heatsink front, The fanless metal case has worked well enough for me and I have run some CPU swamping software and played high bitrate full screen max res videos at same time, even case upside down (wifi issue work-around) it never got hotter than mid 50C

[Sage]

Many thanks, TXD. Your first two paras are outside my ambit - I do HW!
As for mid-50C, in my opinion it's too hot bearing in mind the physical size both of package and features inside package. The instructions are clear. Putting it in a case probably violates your warranty! No case to answer, so to speak!

[TeX Dog]

@Sage yeah right no one going to believe you do not know your way around a keyboard, text editor and that soft-mushy stuff called software, look at BK blog he is off on a multi-day massive compiling bender so quit pretending you need BK to roll a 8.1.3 clean image.
We ain't buying ( OH I AM JUST A HARDWARE GUY, DERP DERP!)

[Sage]

Yup, and I don't believe folks 'round here don't know which end of the soldering iron to hold and things like fitting heatsinks...

[TeX Dog]

Yup, and I don't believe folks 'round here don't know which end of the soldering iron to hold and things like fitting heatsinks...

Ah, yeah your right.. I have more soldering irons than radios But I am horrible about burning and cutting my fingers on sharp metal so not much desire. Heck last night I cut my finger on a plastic fake pineneedle.

[pakt]

Barry, I saw your blog comment "Running warm".

I think that you're running with the CPU scaling frequency governor set to the default powersave. That governor limits the CPU frequency to 600 MHz.

If you change it to performance, the CPU frequency will double to 1.2 GHz and stay there. You'll most likely notice the heatsink get hotter (SoC temp > 58 C) but also a bit better performance when compiling

Here's how to make the changes. First check which governor is being used:

Code: Select all

# cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor
powersave
#
# vcgencmd measure_clock arm
frequency(45)=600000000

Now change to the performance governor:

Code: Select all

echo performance > /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor
cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor
performance
#
# vcgencmd measure_clock arm
frequency(45)=1200000000

The SoC core temperature can be measured like this:

Code: Select all

# vcgencmd measure_temp
temp=36.5'C

[TeX Dog]

So this is not right? does't seem to change?

Code: Select all

#  cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor 
powersave
# vcgencmd measure_clock arm 
frequency(45)=600064000
# 
# echo performance > /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor 
# vcgencmd measure_clock arm 
frequency(45)=600000000
# cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor 
performance
# vcgencmd measure_clock arm 
frequency(45)=600000000
# 

[Sage]

You'll most likely notice the heatsink get hotter (> 58 C)

If that is the external HS contact temperature, I wonder what it is inside the cpu?!
Sensors on my 2x and 4x AMD x86 show 33C under load and that is supposed to be reading from the internal diode, not a contact thermistor. The whole idea of ARM architecture would seem to be lost viz, reduced instruction set, reduced feature size, low power, small footprint?
It's not just the heatsink size, but how quickly the heat can be removed. Al and copper have complementary heat conduction and heat capacity, but only a fan can physically remove the heat into the environment. Enclosures are not a good idea. I saved some large full-size PC cases to run my main systems. The micro boards inside crouch over in one corner. The whole thing is a large air chest from which the 5" PSU fan, located in the opposite corner, can easily remove heat. Some nodding acquaintance with fluid mechanics/aeronautics helps - probably someone reading this has a lot more knowledge and experience than me.

[pakt]

TeX, have you changed any 'turbo' settings in config.txt ?

I'm running with the default settings. What do you get when you run

Code: Select all

vcgencmd get_config int

Here's the output I get when running the 'performance' governor

Code: Select all

# vcgencmd get_config int
arm_freq=1200
audio_pwm_mode=1
config_hdmi_boost=5
core_freq=400
desired_osc_freq=0x36ee80
disable_commandline_tags=2
disable_l2cache=1
force_eeprom_read=1
force_pwm_open=1
framebuffer_depth=32
framebuffer_ignore_alpha=1
framebuffer_swap=1
gpu_freq=300
hdmi_force_cec_address=65535
hdmi_group=1
hdmi_mode=16
init_uart_clock=0x2dc6c00
lcd_framerate=60
over_voltage_avs=0x186a0
pause_burst_frames=1
program_serial_random=1
sdram_freq=450
temp_limit=85

[pakt]

You'll most likely notice the heatsink get hotter (> 58 C)

If that is the external HS contact temperature, I wonder what it is inside the cpu?!

That's me being unclear. I actually meant that the SoC would get hotter, ie, greater than 58 C, making the heatsink hotter in turn.

[TeX Dog]

Code: Select all

# vcgencmd get_config int
arm_freq=1200
audio_pwm_mode=1
config_hdmi_boost=5
core_freq=400
desired_osc_freq=0x36ee80
disable_commandline_tags=2
disable_l2cache=1
force_eeprom_read=1
force_pwm_open=1
framebuffer_depth=32
framebuffer_ignore_alpha=1
framebuffer_swap=1
gpu_freq=300
hdmi_force_cec_address=65535
init_uart_clock=0x2dc6c00
lcd_framerate=60
over_voltage_avs=0x124f8
pause_burst_frames=1
program_serial_random=1
sdram_freq=450
temp_limit=85
# 

[pakt]

Code: Select all

# vcgencmd get_config int
arm_freq=1200

It says right there that your CPU is running at 1.2GHz.

Hmm, nope, that's not right. 'arm_freq=1200' doesn't change when I change the governor, so I'm reading that wrong.

It's probably the maximum frequency the GPU allows the CPU to run at. This SoC is complicated because it's actually the GPU that controls the CPU here. GPU settings made at boot time control, in turn, the settings of the ARM chip

If I understand it right, when the Raspberry Pi was designed, the Broadcom engineers added a 1-core ARM CPU to the GPU they had in production. The Pi boots the GPU first - that's where you see the colors flash on the screen. Then the ARM CPU is booted by the GPU. So the 4 ARM cores of Pi3 are just an appendage of the GPU
Simple, right

[TeX Dog]

Me either going to google online maybe I got a defective unit?? You are using a pi3 like me right? I have added HMDI setting to configs to match yours and warm rebooted. Same results for me.

[TeX Dog]

Not a peep online.. did find this cool oneliner cut/paste to terminal

Code: Select all

echo -e "CPU => $(echo "scale=1; $(cat /sys/class/thermal/thermal_zone0/temp)/1000" | bc)'C\nGPU => $(/opt/vc/bin/vcgencmd measure_temp | sed "s/^.....//g")"

results:

Code: Select all

# echo -e "CPU => $(echo "scale=1; $(cat /sys/class/thermal/thermal_zone0/temp)/1000" | bc)'C\nGPU => $(/opt/vc/bin/vcgencmd measure_temp | sed "s/^.....//g")"
CPU => 39.1'C
GPU => 39.7'C
# 

[TeX Dog]

Other oddball stuff happen, tried to play video and a frozen image of a prior <days ago> was blended in alpha channel. Did a cold reboot, redid commands and ALL IS NORMAL acts like yours. This was running for days and even during a bad T-Storm when I unplugged everything else from wall, forgot it was still on battery backup phone charger so stayed running.