• Content count

  • Joined

  • Last visited

About Shimon

  • Rank
    Advanced Member

Recent Profile Visitors

63 profile views
  1. Do any of you know of an Android 5 ROM for this box? That way, we could run a few benchmarks under termux, at least.
  2. Second that, yandex causes all sorts of problems. Only recently have I overcome the problem with downloads not starting (Load denied by X-Frame-Options) by setting network.http.enforce-framing.soft to false in firefox. No idea how to make it work in chromium, though.
  3. Yeah, haven't tried that yet. Could be trivial or a huge maintenance burden. Not too helpful a name, judging from @zamrih's post it's meant to make a difference for 100M LAN and/or Realtek WiFi?
  4. I've seen this suggested before but I have doubts if it's still applicable to the Debian of today. I'll probably have to experiment with Jessie based images.
  5. @balbes150 Searching the forum for Devuan yields just two results, so there's probably not much chance of seeing a Devuan based image?
  6. Yes, that's exactly it. An A/V composite + stereo audio 3.5mm socket. The other one is an IR socket indeed.
  7. @James Kingdon Judging from the specs, Ugoos UM3 Ubuntu image might be your best bet:!CotRiQII!4l2xX9sKHwsF7_QOdo_XWk7xWA4XRYDJkwnipfndG10 If it doesn't work, here's the mount command for a quick peek: sudo mount -o offset=$[81920*512] sdcard_Xubuntu_0.3.1_um3.img /mnt Nope, it couldn't possibly work, /boot is empty.
  8. Wait a sec, it seems UGOOS UT3S is even marketed as an Ubuntu 14.04 box, it would be great if someone owning/familiar with that box could describe the state of Linux as it existed before the arrival of these new RK3288 SBC's.
  9. That's a typo, 8 cores were definitely used. Other than that, it seems the unholy alliance of big.little scheduling and DVFS settings are causing this issue. You could either experiment with some cpufreq tunables or try rebuilding the kernel with: `CONFIG_ARM_DT_BL_CPUFREQ=y`. Not sure if it's possible to disable `CONFIG_ARM_BIG_LITTLE_CPUFREQ` altogether.
  10. Speaking of GPU's and binary blobs, I wonder if OpenCL and ARM's ComputeLibrary are going to be usable on these boxes.
  11. As well as these options, needed for running iotop: - I/O accounting support (CONFIG_TASKSTATS, CONFIG_TASK_DELAY_ACCT, CONFIG_TASK_IO_ACCOUNTING) Yeah, that 2nd benchmark utilises just 2 cores, so it's clear the number of active cores doesn't matter as long as just one little core remains online. Amlogic should be asked to clarify what's going on here. I sincerely hope we're not witnessing another S905 2GHz situation. @buvaluy Could you try using the hotplug governor? It should be able to work around this limitation, giving us either 1-4 fast cores or 5-8 slower cores automatically.
  12. Actually, I remember noticing a few pages back in c-ray, 8-core performance was not scaling as expected, so the issue was already there as well. We were trying to test HMP at the time, that's why nobody thought about doing single-threaded tests. Turning off little cores makes the frequency scaling behave like any 4-core S905 system, including single-core performance. That's the problem (bug?) here, that is, as long as all cores are online, the frequency gets lowered, even if just 1-2 cores are being used. @buvaluy Could you try the most obvious test, activating just one little core to see what happens? Provided you already have just 4 cores active, add one more: echo 1 | sudo tee /sys/devices/system/cpu/cpu4/online
  13. Thanks, I completely forgot the s905_autoscript was hardcoded to query just usb 0:1 @sukanime In the interest of readability, please use the spoiler tag when pasting huge walls of log output. Look here for a neat example:
  14. @balbes150 I think I've just discovered a bug in the booting sequence. Trying to boot from a USB disk while another USB drive is connected, ends up booting Android from eMMC. Booting from an SD card works as expected, and the USB drive can be left connected.
  15. Cool, thx! @buvaluy @balbes150 Do you think it's a bug or a feature of DVFS?