Google Pixel 8 and Pixel 8 Pro confirmed to lack Wi-Fi 7 connectivity
Google Pixel 8 and Pixel 8 Pro confirmed to lack Wi-Fi 7 connectivity
The Pixel 8 and Pixel 8 Pro have now turned up on the FCC's database, just over a month before their official launches alongside the Pixel Watch 2. So far, the FCC has outlined five model number variants, all without rumoured Wi-Fi 7 connectivity.
So according to an FCC filing the Pixel 8 and Pixel 8 Pro will support Wi-Fi 6E but not the newer Wi-Fi 7.
For a comparison, you can get a max speed of 9.6 Gb/s with Wi-Fi 6E, while Wi-Fi 7 can reach a speed of 46 Gb/s.
You're viewing a single thread.
I'm just going to put this information here: the use case for 46Gb WiFi is going to be extremely niche. There is nearly no legitimate use case where you can achieve that speed on your phone.
The problem here is that:
- The majority of internet traffic is TCP
- TCP protocol processing is atomic (i.e. your speed is bottlenecked by a single CPU)
- The bottleneck is the receiver (i.e. downloader)
- TCP is too complex for efficient receiver-side hardware offloads (i.e. can't workaround this issue by adding more special hardware)
What does this mean?
Your connection speed on a wifi 7 device WILL be bottlenecked by your single-core CPU speed, even if you are doing absolutely nothing except transmitting data. This assumes you are only using a TCP single connection (e.g. downloading a file from a website). But that's the majority of use cases unless you are running a server (in this case on your phone).
I haven't checked what CPU the Pixel 8 uses. But my Pixel 7 has a Cortex A-78. I also don't have the raw data handy for the 3Ghz A-78, but I do have data from the 2Ghz A-53 connected to a 100Gbps Ethernet NIC which is around 8-9Gbps. The A78 generally outperforms the A53 by 1.5x (At least that's the characteristics on the Nvidia Bluefield DPUs). So we can assume 12-14Gbps max for a single connection with Wifi 7 running on a state-of-the-art ARM CPU.
That is still nowhere near 46Gbps. It's like mounting a Vulcan Minigun on a bicycle.
To use the full wifi bandwidth, you would need to have multiple connections running on different cores. That's also not including the switches/servers connected to the wifi AP. Unless you are running a Redis server on your phone, I see no reason why Wifi 7 would be needed unless the remaining hardware is upgraded significantly.
To use the full WiFi bandwidth you'd probably also need to connect your phones antennae to the access points antennae via coax.
Nobody's ever going to get 46Gb/s with 7 just like nobody actually gets 1gigabit with AC. Real speeds tend to be vastly lower than rated, and splitting the airwaves with other stations is reason enough to minimize transmit time in any case. A busy area is going to benefit from phones having Wi-Fi 7 even if those phones can't process a 46gbps tcp stream.
It really annoys me when people look at max rates of networking technologies as though they're minimums for that technology to be useful. You don't have to use all 10gigabits of 10gbps for the upgrade from gigabit to be worthwhile, same with all others.
There's no way the in order A53 from 2012 gets even close to the performance of the OoO A78 from 2020.
Didn't know one is in-order and the other is OoO. The A53 is still being used for new products by Nvidia in 2020 (Bluefield-2). So there must be some merit to it or Nvidia is cheaping out on stuff
The BlueField-3 uses the A78 and unfortunately I don't have one to test. I'm basing everything I know based on conference talks. I do know apparently the A78 does not have working performance counters for perf which makes it a pain to debug.
That being said, a 2023 Mid-end Xeon gets you up to 60Gbps TCP single flow (100Gbps ConnectX-6 NIC) So maybe that's a better comparison? Might need to account for all the other x86 optimizations
Also, I think the bottleneck for TCP processing is branching, not memory access. So I'm not sure if OoO execution would help much. Would the A78 have improved branch predictors?
A53 is used for low-power and low-cost applications... It's a "good enough" CPU that has really good performance/area.
Perfect performance counters for OoO is really hard.
OoO also makes BP more useful. An OoO processor without BP isn't very useful because there aren't that many instructions between branches... So, generally, modern OoO processors dedicate far more resources to BP.