Initial Raspberry Pi 4 Performance Benchmarks

[thesandbender]

There is no such thing as a dumb USB-c cable, at least not according to the specs. Sure, you can use a non-compliant USB-c cable which, if you try to charge your cellphone from accidentally, may or may not fry it.

To clarify, cables rated for 5A must have an e-marker chip to be compliant. Cables rated for 3A do not need to have this chip. Power adapters that are compliant and rated for 5A will refuse to provide 5A over a cable with no e-marker or to a device that doesn't report as compatible (see below). If you use a cheap, non-compliant adapter that doesn't support the e-marker but will still drive 5A, the Pi 4 will work. So AndyChow is correct in that there are no "dumb" cables... just non-compliant cables (with the added twist of 3A vs 5A e-marker compliance). It's the power adapters that are "smart" and "dumb".

The root cause of the problem with the Raspberry Pi 4 is that there are two pins on the Raspberry USB-C connector that are supposed to have their own 5.1K ohm resistor (as clearly stated in the USB-C spec). The Pi 4 has a single 5.1K ohm resistor on both pins. This causes 5A e-marked cables and power adapters to see it as an audio device and refuse to provide power. It's a "simple" fix but has to be done on the Pi 4.

[Brunnis]

Tests are weird, differences between rp4 2gb and rp4 4gb

I’m running the Phoronix Test Suite on my Pi 4 4GB right now. It’s cooled by a USB powered desk fan that blows down on the board. No throttling will occur with this setup. So far (the test is still running) my Pi 4 4GB appears to perform identically to the 2GB version in the article. So, it indeed seems to be a thermal issue.

[coder]

but how does it stack up vs pc hardware

I really want to see a benchmark vs. ODROID-H2, which is a SBC based on the Gemini Lake x86-64 SoC found in many low-end Chromebooks.

https://www.hardkernel.com/shop/odroid-h2/

However, it'd be nice if we got a comparison with like a Skylake i3-6300U dual-core 15W laptop, as an additional point of reference.

[coder]

from the article:
"If you want to see how your own x86/Arm hardware compares to the Raspberry Pi 4, simply install the Phoronix Test Suite and run phoronix-test-suite benchmark 1907127-HV-RASPBERRY81 for your very own fully-automated, side-by-side benchmark comparison."

No. I don't want to see how it compares with my PC hardware...

[coder]

The Raspberry Pi 4 is a huge improvement, no doubt about it. However, they definitely released it prematurely.

The irony is that it was originally scheduled for a launch next year. So, they had plenty of time in the schedule to do some extra testing and get it right.

See the FAQ section, here: https://www.raspberrypi.org/blog/ras...e-now-from-35/

[betam4x]

It really is too bad the Tinker Board S doesn't have better software support. I own one and it works quite well in Armbian, but the last I checked the kernel version was lagging, keeping my dual TV tuner from working. I will be keeping an eye on the Pie 4 to see how things play out. I can't wait until these things can use Cemu at 4k60.

[blackshard]

Tests are weird, differences between rp4 2gb and rp4 4gb, aswell as the stark difference between tinkerboard and firefly (both use rk3328 and should be reasonable close).

Running without heat sinks, or using different software stacks?

Differences between the Rpi4 2gb and 4gb are due to the thermal throttling and the 4gb being in closed case while the 2gb version is the bare board with a heatsink on it. Not the professional way to conduct a test, IMHO.

The tinkerboard has a rk3288 (Cortex-A17 ARMv7 32 bit Out-of-order CPU), while the Firefly has a rk3328 (Cortex-A53 ARMv8 64bit In-order CPU). The rk3288 is more similar to the Cortex-A72 in the RPi4

[deadite66]

Would be nice to see some comparisons to the ODroid boards.

[M@yeulC]

Negative press about the USB-C implementation on the Pi4 is starting to grow. Word on the street is its a foobar.

The USB-C cable guru at Google has declared it a problem already, which is not a good sign.0

Look for a new port design out of the foundation soon.

Also some people have already hacked the Pi4 and wired a PCIe 1x slot to it. Pretty wild.

The cable situation doesn't matter that much, as e-marked USB-C cables aren't that common. (Well, I do not own a single USB-C device yet).

I didn't find any issues regarding the benchmark methodology, the difference was clearly stated multiple times in the article, and helps shed light on that possible issue. Of course, additional benchmarks will be required to investigate this specific matter.

I am especially looking forward to power consumption benchmarks

[dungeon]

It seems it needs fan not to throttle even with board on open