sfx2000

Members
  • Content Count

    410
  • Joined

  • Last visited


Reputation Activity

  1. Like
    sfx2000 reacted to ayaromenok in nvidia jetson nano   
    it's a good idea to use tegrastats utility(for logging) or top-looks jtop from Raffaello Bonghi - it's show CPU\GPU\total power consumption, frequency and load graph for GPU too(on second tab - press 2)
  2. Like
    sfx2000 got a reaction from NicoD in nvidia jetson nano   
    Blender on Jetson Nano as requested - with Cuda, there's no other boards other than nVidia
     
    Time 01:11:37.77 | Mem:140.48M, Peak 142.73M
     
    I would expect Shield TV Box to do better, along with more recent tegra's that support Pascal and later GPU's.
     
  3. Like
    sfx2000 reacted to TonyMac32 in nvidia jetson nano   
    Assuming you have good Dupont connectors, yes.
     
    I use a 5V 14 A chassis supply adjusted to 5.25 for my mess, I think I've got the right barrel jack on it.  Been reworking some floating ground pins on a TRRS jack... 
  4. Like
    sfx2000 got a reaction from NicoD in nvidia jetson nano   
    Hmmm... giving it a try...
     
    With the current L4T - looks like it's about an hour and a half - CPU load is pretty low so it's all GPU...
     
    Jetson - she be getting pretty warm there - for folks with MicroUSB, make sure you're in the low power model (nvpmodel=1)
     
    Anyways - under full power... work in progress... CPU load is actually pretty low, so my thoughts are that this is all running on the GPU itself, because CUDA...
     

  5. Like
    sfx2000 got a reaction from NicoD in nvidia jetson nano   
    Proper power - 5A/5VDC over the barrel... @tkaiser - opened an issue to report.
     
    https://github.com/ThomasKaiser/sbc-bench
     
    http://ix.io/1I4j
     
    Again, general compute is not where nVidia is focused on, but there, as a big ARM, it does ok...
  6. Like
    sfx2000 got a reaction from qblueRed42 in Announcement : Odroid N2   
    Normally with big.LITTLE, tasks migrate from the LITTLE core to the big core - much depends on the kernel scheduler in use, and whether that scheduler is aware of ARM MP in a big.LITTLE context - in any event, one does need to consider performance per core, and how to schedule tasks. Done right, it works pretty well, done wrong and one can get into contention conflicts between cache, main memory, and arbitration on who has control of the CCI at any given time.
     
    nVidia also has the 4 Cortex-A53's along with the 4 Cortex-A57's, but they don't expose the A53's directly to the kernel - management is done below the kernel in the hardware, which might make scheduling easier in the linux kernel.
     
    Worth reading - it's a few years old - https://www.arm.com/files/pdf/big_LITTLE_Technology_the_Futue_of_Mobile.pdf
     
    More recently, there's been a lot of work around energy aware scheduling, and ARM is one of the big players there - https://developer.arm.com/tools-and-software/open-source-software/linux-kernel/energy-aware-scheduling
     
    It's in Kernel 5.0, but perhaps some folks might consider back-porting it if they have the spare time to do so.
     
    Anyways - more cores doesn't mean faster - there's a limit there - and this is more of a systems engineering issue, along with SW... and this goes back to 1967 with Amdahl's presentation...
     
    https://en.wikipedia.org/wiki/Amdahl's_law
     
    Much like Shannon with Information Theory and Wireless Communication - one can only put so many bits into a medium before it becomes error limited.
     
    Anyways - it's an interesting time where we're getting the big ARM cores in boards that are $100USD or lower - but it also starts to point out that we need to have better solutions to power these larger cores...
  7. Like
    sfx2000 got a reaction from NicoD in Announcement : Odroid N2   
    Anyways - would be fun to see a cage match between Nano Pi M4 vs. Odroid N2 vs nVidia Jetson Nano...
     
    Traditional benchmarks as well as some VPU/GPU work - as RK3399 has Mali T864 vs. Amlogic with G52 vs. nVidia with Maxwell
     
    Since all are supported in some way with Ubuntu 18.04LTS with Vendor supplied BSP's...
  8. Like
    sfx2000 got a reaction from NicoD in Announcement : Odroid N2   
    Would be interesting to see how it competes with Jetson Nano with a good power supply...
     
    Jetson has a fairly large heatsink, but it's generally cool to the touch right now with 2.5A over MicroUSB - 5 amp Power Supply is scheduled to arrive tomorrow, and then we can turn up the Jetson clocks and see what it can really do...
     
    (Jetson Nano under full load can easily consume 20w when doing GPU compute tasks - not too much different than Google Coral)
  9. Like
    sfx2000 got a reaction from qblueRed42 in Announcement : Odroid N2   
    Would be interesting to see how it competes with Jetson Nano with a good power supply...
     
    Jetson has a fairly large heatsink, but it's generally cool to the touch right now with 2.5A over MicroUSB - 5 amp Power Supply is scheduled to arrive tomorrow, and then we can turn up the Jetson clocks and see what it can really do...
     
    (Jetson Nano under full load can easily consume 20w when doing GPU compute tasks - not too much different than Google Coral)
  10. Like
    sfx2000 got a reaction from gounthar in nvidia jetson nano   
    It got here yesterday (4.26) - first glances...
     
    (Note - this is a big/LITTLE - the A53's are 'hidden' but they're present - to Linux, only the big A57's are shown - there's some shenanigans going on inside the chip for power management that is not exposed to the OS)
     
    1) When they say 5V 4A - they mean it - and it's best to power this device thru the barrel jack
    2) Can power via microUSB - and there, have a very, very, very good power supply - even then
    2a) powering thru the MicroUSB means that the Tegra is going to into a lower scale power domain - it adjusts the clocks according to not exceed 2A
    3) Display - Display Port to a monitor works, HDMI to a FullHD (1018P) works, HDMI to DVI adapters - strongly not recommended - I had an old one that was supplied with a Mac Mini 2010, and it kinda works
     
    Some tips - use a powered USB hub for Keyboard/Mouse if running from MicroUSB as depending on peripherals, including a WiFi adapter, takes away from that 2A power budget, and can cause stability issues.
     
    (on big chips like this where the board has a barrel jack - don't use MicroUSB)
     
    SD Card - use a quality UHS-1 or better card - I'm using a Samsung Pro Plus - which is a Class 10 UHS3 card, and it works fine.
     
    Info below - running from MicroUSB - I have a 5VDC 4A adapter on order...
     
     
     
     
     
    dmesg - bootup
     
     
  11. Like
    sfx2000 got a reaction from qblueRed42 in nvidia jetson nano   
    It got here yesterday (4.26) - first glances...
     
    (Note - this is a big/LITTLE - the A53's are 'hidden' but they're present - to Linux, only the big A57's are shown - there's some shenanigans going on inside the chip for power management that is not exposed to the OS)
     
    1) When they say 5V 4A - they mean it - and it's best to power this device thru the barrel jack
    2) Can power via microUSB - and there, have a very, very, very good power supply - even then
    2a) powering thru the MicroUSB means that the Tegra is going to into a lower scale power domain - it adjusts the clocks according to not exceed 2A
    3) Display - Display Port to a monitor works, HDMI to a FullHD (1018P) works, HDMI to DVI adapters - strongly not recommended - I had an old one that was supplied with a Mac Mini 2010, and it kinda works
     
    Some tips - use a powered USB hub for Keyboard/Mouse if running from MicroUSB as depending on peripherals, including a WiFi adapter, takes away from that 2A power budget, and can cause stability issues.
     
    (on big chips like this where the board has a barrel jack - don't use MicroUSB)
     
    SD Card - use a quality UHS-1 or better card - I'm using a Samsung Pro Plus - which is a Class 10 UHS3 card, and it works fine.
     
    Info below - running from MicroUSB - I have a 5VDC 4A adapter on order...
     
     
     
     
     
    dmesg - bootup
     
     
  12. Like
    sfx2000 got a reaction from gounthar in Which boards to get a VPN server and a VPN client?   
    But it works... wg is cool as CPU load is minimal, compared to the CPU intense OpenVPN -
     
    Enter MIPS24Kc, which would be comparable to ARM11 - Pre-ARM Cortex-A7... the other day, two clients attached, and CPU load was 3 percent, and that's with Stubby and WG active over a 4G mobile hotspot (USB-150 was WiFi to the phone as WAN, and routing from there to two WiFi clients)
     
    I suppose this is the difference between an Application focused SoC vs a Network focused platform.
     
    There's another thread where I chat a bit about MIPS - interesting arch there for networking stuff...
     
    Anyways - the openwrt repo's also support Tor, which may be of some interest to some...
  13. Like
    sfx2000 got a reaction from gounthar in Which boards to get a VPN server and a VPN client?   
    Something to consider...
     
    https://www.gl-inet.com/products/gl-usb150/
     
    MIPS based - but full OpenWRT/Linux under the hood - wireguard client/server and OpenVPN client/server - plug into the PC/Mac/Linux and it is an ethernet device, but not needed, just give it power and it'll run...
     
    If one is bent toward WiFi hacking, the SoC WiFi is ATH9K based, which is one of the better WiFi chips for documentation purposes...
  14. Like
    sfx2000 reacted to PDP11 in Geezer goes GaGa over Terminus console font!   
    It's *amazing* what you can find reading the Armbian documentation! 
     
    I dig the shell, even old-school in a virtual terminal or minimal server type setup.  Normally that means going through some squinty-eyed hoops to get the Terminus console font installed with larger sizes in the first place on large monitors.
     
    Wait - what's this?
     
    sudo dpkg-reconfigure console-setup
     
    I did this in a virtual terminal, and allowed setup to pick the right character set for me.  Then I blasted the framebuffer with the largest Terminus font size allowed.
     
    Perfection!  Not a big deal to most, but having this be part of the standard distribution image made my day.  Soooo easy.  thanks!
     
  15. Like
    sfx2000 got a reaction from qblueRed42 in raspberry pi 4 portotype review   
    Erm - no....
     
    Someone is playing games here - the chip count alone on the BOM suggests not.... they need to hit $35 APS, which considering BOM and GOGS, this board is not an RPI4...
     

     
    RPI4 - likely VC5 with 2 or 4 Broadcom B53 cores - USB3 and 4K might be nice, and USB3 gives Gigabit Ethernet. VC5 does extend the addressable memory space, so maybe 2G (would be nice)
     
    Just IMHO...
  16. Like
    sfx2000 got a reaction from Jack953 in Orange Pi 4G-IOT   
    Not ready to say, as it fills a gap that is currently missing...
  17. Like
    sfx2000 got a reaction from TonyMac32 in What does your workbench look like?   
    Hehe - my first calculator back when I was 12 or so...
     
    http://www.vintagecalculators.com/html/novus_650_-mathbox-.html
     
    So stepping into HP RPN calcs was not a problem
  18. Like
    sfx2000 reacted to TonyMac32 in What does your workbench look like?   
    I see dust and overpriced coffee... 
     
    I have a 50G as part of having been an EE major.  The professors wouldn't support the TI hardware.  And, since I'd done assembly coding of an 8087 math co-processor, RPN made perfect sense to me right away...    I also got the 50 and not the 49 because the 49 was complete garbage, quality wise.
     
     
  19. Like
    sfx2000 got a reaction from TonyMac32 in Armbian in 3D   
    @NicoD - Nice concept...
     
    Put some muscles on the penguin - big ARM's for Armbian...
  20. Like
    sfx2000 got a reaction from lanefu in Reliable SMTP for forum notifications?   
    Goes back to what I was saying about DKIM/SPF/DMARC as this establishes the trust and credibility for being an SMTP sender... Requires a bit of DNS work to update the records, but once done, generally works.... then just keep on top of certificates, which isn't much overhead there once they're set up.
     
    One can use postfix, qmail, or sendmail even (don't consider sendmail - it's powerful, but it's a curve all it's own).
     
    postfix works well as a smarthost for outbound, and works equally well for inbound as an full blown MTA
     
    BTW - I used to be the MMS admin for the 7th largest Wireless Telecom with over 7m users in the US - and MMS has the SMTP gateway...
     
     
  21. Like
    sfx2000 got a reaction from rooted in Random Automotive off-topic   
    that's not the same Ranger...
     
    Going back to Slant Six - when I was a kid - my dad had a older Dodge Van (the one where you sat over the front wheels, and no protection if you had a front end hit...)
     
    It had I suppose 150,000 miles, and it threw a rod...  dad welded up the side of the block in the driveway, ground down the broken rod end, and drove it for 3 more years as a slant five...
     
    beast of a motor, even wounded...
     
  22. Like
    sfx2000 reacted to TonyMac32 in Random Automotive off-topic   
    Indeed.  Had failure of idiotic plastic thermostat housing (replaced with Aluminum), then the timing chains flew out the side and bent a bunch of valves at 117k (miles).  The shortest lived engine I've ever had in an American car (I had a Stratus, but I may have crashed it)
     
    To be fair to Cologne, the pushrod version of the engine was far better, but the weird fetish with overhead cams was it's undoing.  I'm swapping in a 5.0L V8 from an Explorer, since they were the same vehicle from the drivers seat forward it will be extremely easy, the donor vehicle with the engine in it costs about as much as a new head for that V6...  I agree the LS is solid, but the GM people are too insufferable, I can't do that to the Ford.  If a Dodge 318 fit in there, I'd gladly do that, the community would give me a high five for something different...
  23. Like
    sfx2000 got a reaction from Werner in Build Armbian as User   
    I have a dedicated KVM image, as I don't want to introduce changes from the distribution that might impact the build process - it's an old habit of mine.
     
    Anyways - root is a major deity on a *nix system, and an errant script could do a fair amount of damage - it doesn't have to be a "bad actor" but even a typo...
     
    Started looking at fakeroot as an option, as this does provide the opportunity to do file manipulation as needed to build an image.
     
  24. Like
    sfx2000 reacted to markbirss in Lichee Pi zero   
    Here is the Lichee Pi Zero with 4.3inch 800x480, keyboard and Li-Po battery. (also using a ESP8266 ESP-01 for Serial console over wifi)
     
     
     

  25. Like
    sfx2000 got a reaction from ftp-bin2fex in Post your Speed Tests   
    Couple of things - I used to work for a Wireless Carrier that did 3G and LTE - 7th largest in the US until it was consumed by AT&T...
     
    10.0.0.0/8 is private addressing - we used this plus the other ranges internally inside the engineering plant, from core to edge.
     
    100.64.0.0/10 - this is carrier grade NAT, we used this extensively on the customer facing network, as it scales nicely
     
    We didn't allow public IP's ever - just didn't happen, we didn't want to be hosting private servers on our network.