What's new

BCM4916 process node?

  • SNBForums Code of Conduct

    SNBForums is a community for everyone, no matter what their level of experience.

    Please be tolerant and patient of others, especially newcomers. We are all here to share and learn!

    The rules are simple: Be patient, be nice, be helpful or be gone!

jixiangyuan

Regular Contributor
Hi, I read that the BCM4908 is made using the 28nm process, and the B4912 is made using TSMC’s 16nm process, but I can’t find any information about the BCM4916 regarding process node online. Does anyone know which process it use?
 
Hi, I read that the BCM4908 is made using the 28nm process, and the B4912 is made using TSMC’s 16nm process, but I can’t find any information about the BCM4916 regarding process node online. Does anyone know which process it use?
Why would anyone care what size process it uses? :confused:
 
Maybe because it has something to do with the temperature and power usage of the cpu? If we are starting to ask questions, why would anyone reply to a thread about which they know nothing
It just seemed like a strange question. I buy a router based on its features, availability and price. The size of the process used to create one of the chips inside it is not a consideration for me.
 
It just seemed like a strange question. I buy a router based on its features, availability and price. The size of the process used to create one of the chips inside it is not a consideration for me.
I understand that. It just one of those things I want to know after purchasing any tech products. Also there were numerous people posting their concerns about the temperature of their CPU inside of their router, particularly with BCM4908, so I was wondering if this will happen to BCM4916 as well in the summer, if it use the same process, with higher clock speed (compare to BCM4912) and with fan-less design (ASUS BE98)
 
Broadcom has always been very secretive with their specs, often only providing them on a need-to-know basis to actual customers. The only information publicly available is their specs sheet:


Based on the significant clock increase between the BCM4912 and BCM4916 while still using the same B53 cores, it would make sense to assume they also went through a node shrink. However any improvement in their power budget was most likely reallocated to the increased clock and increase in the available 10 Gbps interfaces, so I would expect thermals to be fairly similar to the BCM4912.
 
Broadcom has always been very secretive with their specs, often only providing them on a need-to-know basis to actual customers. The only information publicly available is their specs sheet:


Based on the significant clock increase between the BCM4912 and BCM4916 while still using the same B53 cores, it would make sense to assume they also went through a node shrink. However any improvement in their power budget was most likely reallocated to the increased clock and increase in the available 10 Gbps interfaces, so I would expect thermals to be fairly similar to the BCM4912.
This could be true, but after I saw ASUS advertise that the BE98 has 18% better heat dissipation, I started to think that Broadcom might still be using 16nm.

It should be pretty easy to find out if we can monitor the temperature of the CPU for both models. The BCM4916 is a Quad B53 at 2.6GHz, while the BCM4912 is a Quad B53 at 2GHz; that’s a 30% increase.

Let’s assume that the power consumption/temperature has a linear relationship with frequency (I know it’s not exactly that). Combined with 18% better heat dissipation, if Broadcom still on the 16nm node, the CPU temperature would definitely be higher than that of the BCM4912. If Broadcom moved onto the 12nm node, then the CPU temperature could potentially be lower or at the same level as the BCM4912.

Oh, I also emailed Broadcom and asked them about this, but they didn’t want to disclose any information other than what’s on their website.
 

Attachments

  • Screenshot_20240419_103234_Chrome.png
    Screenshot_20240419_103234_Chrome.png
    246.5 KB · Views: 16
Broadcom has always been very secretive with their specs, often only providing them on a need-to-know basis to actual customers. The only information publicly available is their specs sheet:


Based on the significant clock increase between the BCM4912 and BCM4916 while still using the same B53 cores, it would make sense to assume they also went through a node shrink. However any improvement in their power budget was most likely reallocated to the increased clock and increase in the available 10 Gbps interfaces, so I would expect thermals to be fairly similar to the BCM4912.
The CPU Temperature I got for my BE98 (BCM4916) is 61.021 at idle, using the command :
cat /sys/class/thermal/thermal_zone0/temp | awk '{print $1 / 1000}'

Wondering if you or anyone with AXE16000 (BCM4912) would like to share their CPU temperature. Thanks.

My Room Temperature is at 25 degrees Celsius with Humidity at 50% (and pressure at 1023hPa).
 
Wondering if you or anyone with AXE16000 (BCM4912) would like to share their CPU temperature. Thanks.
My GT-AXE16000 sits at 75C, however that doesn't necessarily say much since there's often been a lot of variance in temperatures between routers of a same model.
 
My GT-AXE16000 sits at 75C, however that doesn't necessarily say much since there's often been a lot of variance in temperatures between routers of a same model.
Thanks. Yes, I know there are quite a lot of variables, but it's still good to see that the temperature of my BE98 is lower, rather than higher, thanks to the new thermal design and possible node shrink.
 
my AX1100 pro runs at 64c today room temp 24c chip is listed as bcm4912
 
my AX1100 pro runs at 64c today room temp 24c chip is listed as bcm4912
Thanks for sharing. I assume you are on stock firmware? Merlin is using his custom firmware, so I won't be surprised that his CPU temperature is higher than normal.
 
Thanks for sharing. I assume you are on stock firmware? Merlin is using his custom firmware, so I won't be surprised that his CPU temperature is higher than normal.
no I always run Merlin FW
 
no I always run Merlin FW
Hmmm, interesting. Maybe it has something to do with the workload of his router. But I guess if you are getting 64c, it’s still quite possible that Broadcom has moved to a newer node to manufacture the BCM4916.
 
mmm, interesting. Maybe it has something to do with the workload of his router.
It's just there's a high variance between devices, as the QA isn't the best it could be. Quite a few people discovered that in their specific router the thermal pads weren't properly installed, leading to their router having a 10-15C temperature delta with other people with the same model.
 
It's just there's a high variance between devices, as the QA isn't the best it could be. Quite a few people discovered that in their specific router the thermal pads weren't properly installed, leading to their router having a 10-15C temperature delta with other people with the same model.
yes I have seen huge differences in thermal pads or whatever in routers GPU and computers I have taken apart . QC is not what it should be
 
I have RT-AX88U Pro (BCM4912) with the stock firmware.
It seems to be running pretty cool.
`/sys/devices/virtual/thermal/thermal_zone0/temp` says 55C (ambient is 20C).

The stock firmware has the `stress` utility build-in. I tried to run for 5 minutes with 8 threads to load the cpu (`stress cpu -e 8 -t 600`) and I only got to 62C.

It would be nice if there was a way to see at which frequency is running under stress. Specifically if's holding 2 GHz on all cores.
I tried to look in /sys and do some googling but I can't find any cpufreq reporting.
 
Last edited:
I have RT-AX88U Pro (BCM4912) with the stock firmware.
It seems to be running pretty cool.
`/sys/devices/virtual/thermal/thermal_zone0/temp` says 55C (ambient is 20C).

The stock firmware has the `stress` utility build-in. I tried to run for 5 minutes with 8 threads to load the cpu (`stress cpu -e 8 -t 600`) and I only got to 62C.

It would be nice if there was a way to see at which frequency is running under stress. Specifically if's holding 2 GHz on all cores.
I tried to look in /sys and do some googling but I can't find any cpufreq reporting.
Thank you! That’s roughly 60°C if the ambient temperature is at 25°C. It’s surprising to see your AX88U Pro is cooler than the AXE16000 posted by the others earlier, as the latter is bigger and in theory should have better thermal performance.
 
I am not too surprised. I do think that the jump to a newer process node makes a difference and should both reduce power consumption and heat generation. I think the AX88U Pro also comes with a smaller power supply as a result vs the AX88U (non-pro).
The 200Mhz difference between the BCM4908 and BCM4912 is not a major step. On the contrary, the presence of the 10G network interface and the even higher clocks of BCM4916 could offset the gains of using a newer node.

What surprised me is the relative minor change in temperature under stress test, only 7c in my case. I wonder if `/sys/devices/virtual/thermal/thermal_zone0/temp` is a cpu sensor or just a generic sensor within the device.

Also, as @RMerlin pointed out, the quality of the heatsink mount makes a lot of the difference.
 

Sign Up For SNBForums Daily Digest

Get an update of what's new every day delivered to your mailbox. Sign up here!
Top