Spyke

Posts

linuxphones·Linux PhonesbyTheMightyCat

Fairphone 6 + PostmarketOS review

This is my second linux phone and coming into it i had high hopes and low expectations, the result was a mixed bag but generally positive.

My first was linux phone was the PinePhonePro and i really liked it at the time as a concept, but it was hard to use it as a daily driver, i think i lasted around a month before buying an android.

The PPP had several problems, it was very slow, would get very hot, and my primary concern as i said here its battery was terrible.

And the battery has since fully died, so sadly i can't run side by side tests.

So that brings us to the FP6.

The installation was pretty smooth, follow the steps from the wiki and for me they worked perfectly.

The usage itself feels alot more snappy compared to the PPP, general UI work is comparable to a normal android i would say, and no longer each action has a delay.

So i wanted to tackle the battery point, i ran a few tests, please take these with a grain of salt as they were not done professionally, but they should be indicative:

Idle (not full suspend as the SSH connection was still active):
Power	0.43 W (98 mA @ 4.44 V)
Drain	2.2 %/hr (45.4h)
CPU	0.9%
GPU	264 MHz

Idle with screen on:
Power	0.95 W (214 mA @ 4.43 V)
Drain	4.9 %/hr (20.4 h)
CPU	3.0%
GPU	277 MHz

Watching 1080p H.264 video with BT headphones:
Power	1.57 W (359 mA @ 4.36 V) 
Drain	8.1 %/hr (12.3h)
CPU	20.7%
GPU     278 MHz

Peak CPU:
Power draw : 6.15 W (1437 mA @ 4.28 V)
Drain rate : 32.6 %/hr (3.0 h)
CPU usage  : 100.0 %
GPU        : 277 MHz

Temprature stayed fine across all tests only during the CPU test did it became warm but not too much, which is in stark contrast to the PPP which would get very uncomfortable to hold, and be severly battery drained.

I think these values are interesting when compared to the same test on e/OS, well atleast that's what i would say if i could gather it.

Sadly i was not able to gather detailed power data on android.

Idle with screen on:
CPU usage : 18.7 %
GPU usage  : 11.7 %

18.7% seemed very sus, but it checks out?

58% 3760/foundation.e.blisslauncher: 49% user + 9.3% kernel / faults: 13209 minor
21% 1476/surfaceflinger: 14% user + 6.9% kernel / faults: 18 minor
12% 1383/vendor.qti.hardware.display.composer-service: 6.7% user + 5.5% kernel / faults: 137 minor
4.9% 972/crtc_commit:137: 0% user + 4.9% kernel
4.2% 986/android.system.suspend-service: 0% user + 4.2% kernel
2.3% 3353/com.android.systemui: 1.6% user + 0.6% kernel / faults: 8031 minor
2.3% 2364/system_server: 1.1% user + 1.1% kernel / faults: 5054 minor
2.1% 26001/kworker/u16:8-qc_ufs_qos_swq: 0% user + 2.1% kernel
2.1% 25893/kworker/u16:3-kgsl_devfreq_wq: 0% user + 2.1% kernel
2% 539/irq/153-24091000.qcom,bwmon-ddr: 0% user + 2% kernel
2% 15/rcuog/0: 0% user + 2% kernel
1.7% 4284/foundation.e.advancedprivacy: 1% user + 0.6% kernel / faults: 693 minor
1.1% 14/rcu_preempt: 0% user + 1.1% kernel
0.9% 25150/kworker/u16:13-events_unbound: 0% user + 0.9% kernel
0.7% 973/crtc_event:137: 0% user + 0.7% kernel
0.7% 946/kgsl_hwsched: 0% user + 0.7% kernel
0.7% 25148/kworker/u16:11-wmi_rx_diag_event_work_: 0% user + 0.7% kernel
0.5% 23213/kworker/u16:0-wmi_rx_event_work_queue: 0% user + 0.5% kernel
0.6% 1393/vendor.qti.hardware.perf2-hal-service: 0.1% user + 0.4% kernel / faults: 38 minor
0.4% 26004/kworker/u16:12-events_unbound: 0% user + 0.4% kernel
0.3% 1338/android.hardware.audio.service: 0.1% user + 0.2% kernel / faults: 34 minor
0.4% 1362/android.hardware.wifi-service: 0.1% user + 0.2% kernel
0.3% 23453/kworker/0:2-events: 0% user + 0.3% kernel
Watching 1080p video:
CPU usage 19.6%
GPU usage 7.3%

it seems the launcher from e/OS really is slurping down CPU.

So where does this leave us?

Would i recommend this as a daily driver?

No. If i gave this phone to someone who is not a linux fan they would probably get very frustrated.

Is it usable as a daily driver?

I think this is the more interesting question, if you really wanted you could definitely use this as a working daily driver, if you don't mind not having a camera or audio that is. BT works fine, so when combined with voip its a usable phone. and does all the phone things you would want it to do. Also if you are photosensitive it might not be a fit as there are currently some graphical flicker glitches. The audio driver is in the works if i look at postmarketos so if that gets fixed it would become much more daily driver worthy.

For the people interested i have also uploaded the vulkaninfo

View original on ani.social

Very large amounts of gaming gpus vs AI gpus

cross-posted from: https://ani.social/post/16779655

GPUVRAMPrice (€)Bandwidth (TB/s)TFLOP16€/GB€/TB/s€/TFLOP16
NVIDIA H200 NVL141GB362844.891671257742321
NVIDIA RTX PRO 6000 Blackwell96GB84501.79126.088472067
NVIDIA RTX 509032GB22991.79104.871128422
AMD RADEON 9070XT16GB6650.644697.324110317
AMD RADEON 907016GB6190.644672.25389608.5
AMD RADEON 9060XT16GB3820.322351.282311867.45

This post is part "hear me out" and part asking for advice.

Looking at the table above AI gpus are a pure scam, and it would make much more sense to (atleast looking at this) to use gaming gpus instead, either trough a frankenstein of pcie switches or high bandwith network.

so my question is if somebody has build a similar setup and what their experience has been. And what the expected overhead performance hit is and if it can be made up for by having just way more raw peformance for the same price.

View original on ani.social

Very large amounts of gaming gpus vs AI gpus

GPUVRAMPrice (€)Bandwidth (TB/s)TFLOP16€/GB€/TB/s€/TFLOP16
NVIDIA H200 NVL141GB362844.891671257742321
NVIDIA RTX PRO 6000 Blackwell96GB84501.79126.088472067
NVIDIA RTX 509032GB22991.79104.871128422
AMD RADEON 9070XT16GB6650.644697.324110317
AMD RADEON 907016GB6190.644672.25389608.5
AMD RADEON 9060XT16GB3820.322351.282311867.45

This post is part "hear me out" and part asking for advice.

Looking at the table above AI gpus are a pure scam, and it would make much more sense to (atleast looking at this) to use gaming gpus instead, either trough a frankenstein of pcie switches or high bandwith network.

so my question is if somebody has build a similar setup and what their experience has been. And what the expected overhead performance hit is and if it can be made up for by having just way more raw peformance for the same price.

View original on ani.social

You reached the end