Spitz AX (X3000) Download Speed Limited Unless Software Acceleration is Enabled

Routers
1

Hi everyone,

I've been generally happy with my Spitz AX (GL-X3000) router, but I've encountered a peculiar issue regarding download speeds that I wanted to share and see if anyone has similar experiences or insights.

I'm finding that I only achieve my full download speed of around 720 Mbps from the local Vodafone 5G cell (which I can confirm by testing with the same SIM card in my phone directly 100 meters from the cell tower) under a very specific router configuration.

Here's what I've observed:

  • If "Network Acceleration" (found under "Network" -> "Network Acceleration") is disabled, my download speed is capped at around 500 Mbps.

  • If "Network Acceleration" is enabled and the "Mode" is set to "Hardware Acceleration", my download speed is also capped at around 500 Mbps.

  • Only when "Network Acceleration" is enabled and the "Mode" is set to "Software Acceleration" do I get my full download speed of approximately 720 Mbps (the best the tower can deliver).

It seems that both having Network Acceleration completely off and having it on in "Hardware Acceleration" mode somehow limits the download throughput. I've been able to reproduce this consistently over about 40 atttemps during 14 days.

Is this expected behavior, or could this indicate a potential firmware issue? It feels counterintuitive that "Hardware Acceleration" would perform worse than "Software Acceleration" or even having the feature disabled, in terms of raw throughput.

For context, VPN is completely disabled during these tests of course and everything is connected via ethernet.

Here are my current reception values, in case they are relevant:

LTE FDD
Band: 1
UL Bandwidth: 20M
DL Bandwidth: 20M
RSRP: -74 (Excellent)
RSSI: -46 (Excellent)
RSRQ: -6 (Excellent)
SINR: 25 (Excellent)
Cell ID: 78E314

NR5G-NSA
Band: 78
DL Bandwidth: 80M
RSRP: -67 (Excellent)
RSRQ: -10 (Excellent)
SINR: 34 (Excellent)

Has anyone else noticed this? Any thoughts on why this might be the case?

Thanks!

3

Electromagnetic interference from the operator's signal tower, when the router has hardware acceleration enabled, causes the hardware module to be more sensitive to interference (for example, the anti - interference ability decreases during high - speed signal processing). This leads to packet loss and retransmission in data transmission, slowing down the speed.

4

Hi Alen,

Thanks for offering that explanation regarding potential electromagnetic interference (EMI) affecting the hardware acceleration module.

I understand that EMI can indeed cause issues like packet loss and retransmissions, which would certainly impact speed. However, I'm finding it a bit difficult to reconcile that explanation fully with my specific observations:

  1. Software Acceleration: Achieves full ~720 Mbps.

  2. Hardware Acceleration: Limited to ~500 Mbps.

  3. Network Acceleration OFF: Also limited to ~500 Mbps.

If EMI is significantly degrading the performance of the dedicated hardware acceleration module, leading to speeds of 500 Mbps, it's puzzling why disabling acceleration entirely (relying on the main CPU) results in the same 500 Mbps limitation. One might expect the CPU-bound processing (Acceleration OFF) to be less susceptible to the specific type of EMI that purportedly affects only the high-speed processing in the hardware module, or at least perform differently.

Furthermore, if the router's hardware components are generally susceptible to EMI from the cell tower, it's also curious why software acceleration (which still relies on the router's main CPU and other potentially exposed hardware) would be completely unaffected and deliver full throughput.

This leads me to wonder if the issue might be less about external EMI selectively impacting one processing mode and more about a potential firmware or driver-level matter. Perhaps there's an issue with:

  • The way hardware acceleration is implemented or optimized for certain types of 5G traffic or specific modem interactions.

  • A bug in the NPU (Network Processing Unit) driver or firmware layer that causes it to underperform in my specific scenario.

The fact that software acceleration works perfectly suggests the router's core processing capabilities and the modem itself are capable of handling the full speed, but the hardware offloading mechanism isn't delivering the expected benefit – in fact, it performs no better than having acceleration off.

Could this be something the firmware team could investigate further as a potential area for optimization or a bug fix related to the hardware acceleration pathway?

Thanks again for the input!