What Pi Linux Benchmarks Would You Like To See?

[phoronix]

Phoronix: What Pi Linux Benchmarks Would You Like To See?

First up, I'm currently in the process of running benchmarks on the Raspberry Pi Zero and expect to have those initial results later today or tomorrow for this $5 USD ARM development board...

http://www.phoronix.com/scan.php?page=news_item&px=Pi-Benchmark-Requests

[Namenlos]

Storage would be interesting, usb and sd (and sata for those boards that have it). Also network performance.

[elliotpotts]

(accidental double post, ignore this)

[elliotpotts]

I'd like to see some gpu/video benchmarks using the open source driver. I'd like to see some power benchmarks as well - what power does it draw idle, under load, etc.

[miah]

Personally I'm more interested in the status of the development VC4 driver. Does it allow one to run Gnome or Gnome Wayland? What is working or still broken?

[alpha_one_x86]

All realated to server: network, cpu, storage, power consumption.

[Dick Palmer]

I'd be very interested to see a filesystem performance comparison on that wee CPU:

FAT32, exFAT, F2FS, XFS, EXT2, EXT4, XFS without journal, EXT4 without journal.
All run on both a partition on the Pi's SD card* and again on a USB mounted SSD. Particular attention to CPU burden.

Would also be interesting to see some of the various FS compression options added to those tests, to examine the bus-throughput vs CPU-overhead tradeoff on this comparatively bus and CPU bound system...

*Assuming it's possible to boot a Pi from a card with multiple partitions - don't own one (yet) so don't know.

[Pander]

http://www.midwesternmac.com/blogs/jeff-geerling/raspberry-pi-microsd-card

[willmore]

I would like to add my vote for uSD card testing--both with the onboard uSD interfaces and USB attached ones. I know it'll be a PITA as the differences between uSD cards would invalidate any direct comparisons--as the link Pander posts points out--so you'd pretty much have to reuse the same card--unless you were very confident that you got multiple cards that are truely identical. I've seen variation between cards with the same batch code, so this might not be very practical.

My standard CPU benchmark for these little bards is "openssl speed". Sadly, it's a test of wether the distro set the right build flags half of the time.

Thank you, Michael.

[eydee]

Graphics, even if it's some ancient game/benchmark or Xonotic or whatever. If it could be matched against some video cards that used to have similar power, even better.