ASUS Blue Cave AC2600 Dual-Band Wireless Router Reviewed

[RMerlin]

There was some discussion of the CPU specs in the Phicomm K3C thread. A site in Chinese said, "Network SoC: Intel AnyWAN GRX350 PXB4395EL1600 800MHz Dual Core Three-Thread"

At first I assumed that this referred to the offloading which Intel claims that each radio does but it really seems like the third core could be a result of Hyper-Threading.

Would you be able to confirm the clock speed of the CPU? Intel claims 1600 MHz but Phicomm support and the Phicomm UI show 800 MHz.

I saw 800 MHz here, but I didn't check if it was the max clock frequency, or just the current one.

 

[Easy Rhino]

Hmm, the main occupant of the third core seems to be wanduck. And when Tim was running the bufferbloat test without hardware acceleration the CPU screen was showing the third core maxed out. Wonder if that's related.

wish I had the link to the Chinese forum with the bootlog showing the cpu boot messages.

 

[RMerlin]

Hmm, the main occupant of the third core seems to be wanduck. And when Tim was running the bufferbloat test without hardware acceleration the CPU screen was showing the third core maxed out. Wonder if that's related.

wish I had the link to the Chinese forum with the bootlog showing the cpu boot messages.

Wanduck is responsible for checking that the WAN connection is up, handling failover/disconnections, etc... In itself it doesn't use much CPU.

The core used may change between reboots (or between software restarts), Asus hasn't applied any CPU affinity to wanduck, so don't read too much into that. Tim's high usage on a specific core is usually because that core was responsible for handling traffic/routing (that code isn't multithreaded unfortunately).

Asus applies core affinity only for a few specific processes, such as Samba, OpenVPN, switch handling, USB handling, etc...

 

[Easy Rhino]

hmm, this page:
https://www.techbang.com/posts/5696...lexa-and-ifttt-asus-blue-cave-wireless-router

has a printout of a /proc/cpuinfo from the blue cave. It lists 3 cpus, and one of them has "VPE=1" , which is MIPS lingo for a virtual processor, So maybe... just one of the CPU's is multithreaded?

And based on this thread:
https://www.mips.com/forums/topic/cpu-info-details/

I wonder if CONFIG_NR_CPUS could be changed to 4 to allow four logical processors?

 

[RMerlin]

has a printout of a /proc/cpuinfo from the blue cave. It lists 3 cpus, and one of them has "VPE=1" , which is MIPS lingo for a virtual processor, So maybe... just one of the CPU's is multithreaded?

It's possible. <shrug>

 

[CrystalLattice]

2600mhz npu for the anywan 550(not 350) intel says: https://www.intel.com/content/dam/w...uct-briefs/anywan-soc-grx350-grx550-brief.pdf

phicomm k3c is basically sitting near the top of router charts & router ranker of the only site which does scientific/engineering/computerscience/networking router testing. why ? the #cores or speed? just enjoy the temporary victory of an asian phone co. over the established competition!

how do we know the R7800 is really 1700mhz? or whatever? was it tested? or checked?

 

[thiggins]

No, I don't test the performance of any components. I test the system.

 

[sfx2000]

hmm, this page:
https://www.techbang.com/posts/5696...lexa-and-ifttt-asus-blue-cave-wireless-router

has a printout of a /proc/cpuinfo from the blue cave. It lists 3 cpus, and one of them has "VPE=1" , which is MIPS lingo for a virtual processor, So maybe... just one of the CPU's is multithreaded?

And based on this thread:
https://www.mips.com/forums/topic/cpu-info-details/

I wonder if CONFIG_NR_CPUS could be changed to 4 to allow four logical processors?

The SoC is a Dual Core SMT - one of the threads is not exposed to the userland via Linux.... and that's ok...

Many folks think ARM/x86 these days - with MIPS it's a bit different.

 

[sfx2000]

It's possible. <shrug>

it's MIPS and it's odd compared to other arch's for folks more used to ARM/x86...

Most of the recent MIPS progress has been in Russia (Baikal) and China (Loongson) for current chips based on MIPS - for general purposes, Imagination Tech that owned the rights to MIPS - sold it off to Canyon Bridge back in late 2017 - Canyon Bridge has close ties to the Beijing government there...

ARM is in an interesting place - Softbank now owns them - more directly, Masayoshi Son own's ARM - and he's 在日韓国人 with a $billion plus...

Yah, I know I'm getting a bit political at the moment...

 

[Easy Rhino]

The SoC is a Dual Core SMT - one of the threads is not exposed to the userland via Linux.... and that's ok...

Well now you've got me wondering what the secret VPE is used for!

 

[sfx2000]

Well now you've got me wondering what the secret VPE is used for!

Runs the main system and the hypervisor that runs the Linux userland?

That CPU would not be visible to Linux depending on how it's configured...

 

[CriticJay]

Yes i bought one, worked for about a week. Seemed to work great. Then the entire network completely stopped -- wired and wireless, had to reset, that was on a Friday night. Then 12 hrs later Saturday morning, wolk up was completely dead again. Returned to Amazon after that. Top was warm, bottom was quite hot. Suspect this router doesn't cool properly due to unique design. Went back to ASUS AC1900P, which is a boring choice, but least is a proven design.

I recently had a VERY similar Blue Cave experience to the person I quoted above.

After receiving my router, I upgraded firmware to the May 25 2018 one - the one with the myriad of security patches and MU-MIMO support, as posted on the Asus support website.

The router will appear to work fine for about 10 minutes ... then, pingtimes to all internet sites will increase dramatically. For example, pinging my ISP's dns server will average 600ms, and go as high as 2000ms. Typically, it's a 9-12ms ping. Similar results are found when pinging 8.8.8.8, 1.1.1.1, www.google.com, etc.

I can put my mediocre previous router back on (Archer C9) and internet ping times are all back to normal. Basically, I can get 9-12ms ping to www.google.com for 15 minutes straight! Then, if I put the Blue Cave back on, it'll appear to ping fine for the first 7-8 minutes (probably because it's "cool" since it was turned off for a bit), then start to degrade. At 10 minutes uptime, average ping to internet sites is about 200-300ms. At 30 minutes uptime (right now) I'm getting 750ms average ping!!

What's strange is that while all this is going on, local LAN-side pingtimes are unaffected, and still 2-3ms for wireless to wireless (i.e. laptop to laptop).

 

[DJ Trance]

New to this forum; read a lot of feedback and comments about this product but decided to give it a try...

Picked one up yesterday and updated it to the latest firmware Version 3.0.1.4.383.19330 (dated Jun.22.2018)...wasn't too excited setting it up; did a factory reset after the firmware update...

Did some WiFi speed test with my Dell laptop with Intel AC9560 2x2 WiFi; my Internet plan is max 1Gbps down and 940Mpbs up; my router is placed on the main floor and my laptop is at the second floor. Was able to pull some download speed of around 580 to 615Mbps download and upload of 270 to 290Mpbs; very impressed with the speed but slow upload speed.

I also have an Asus RT3100 using the latest stock firmware; could only pull 530Mpbs down and 400Mbps up if I am lucky...

 

[phusebox]

Looks like I picked the wrong routers from Asus. I have an AC86U here and Blue Cave the latter I have not yet tested but based on this topic is going straight back without even opening... how could this get so many rave reviews?

The AC86U isn't any good either. Bad range. Slow speeds. Missing many options.

 

[CriticJay]

Looks like I picked the wrong routers from Asus. I have an AC86U here and Blue Cave the latter I have not yet tested but based on this topic is going straight back without even opening... how could this get so many rave reviews?

The AC86U isn't any good either. Bad range. Slow speeds. Missing many options.

I completely disagree (with the bolded above); while I also had an atrocious experience with the gimmicky Blue Cave router, I have had an excellent experience with the AC86U - easily making it the best (consumer) router I've ever used in my home.

The best illustration of how good the AC86U is the fact that my office is located on the opposite side of my house from the router, on a higher floor, with 3 drywall walls in-between. And my desktop PC, in my office, has no problem doing wireless downstream (router to PC) of ~500mbps and wireless upstream (PC to router) of ~180-240mbps.

 

[phusebox]

I completely disagree (with the bolded above); while I also had an atrocious experience with the gimmicky Blue Cave router, I have had an excellent experience with the AC86U - easily making it the best (consumer) router I've ever used in my home.

The best illustration of how good the AC86U is the fact that my office is located on the opposite side of my house from the router, on a higher floor, with 3 drywall walls in-between. And my desktop PC, in my office, has no problem doing wireless downstream (router to PC) of ~500mbps and wireless upstream (PC to router) of ~180-240mbps.

Well I am comparing it to a five year old R7000 which beat the AC86U in every location in my home. I didn't get more than 100 Mbps when in the line of sight and at a distance of 3 to 5 meters .

 

[CriticJay]

Well I am comparing it to a five year old R7000 which beat the AC86U in every location in my home. I didn't get more than 100 Mbps when in the line of sight and at a distance of 3 to 5 meters .

That sounds like a defective product, no? Maybe try to get an exchange

 

[phusebox]

That sounds like a defective product, no? Maybe try to get an exchange

Well maybe. It is going back and I'm getting a (partial) refund.

For now I will have to put up with my old router, at least that gives me acceptable wifi speeds in the living room.

 

[av566788]

Correct.

I notice that the NAT Acceleration function in Blue Cave is called "PPA (Protocol Processing Accelerator)", not "CTF" I usually see in other ASUS router, is that any different?

I also noticed that even if I enable most functions (such as AiProtection, Adaptive Qos, Web & Apps Filters, Time Scheduling), the NAT Acceleration is still working.

However, there is one exception...when I want to limit the bandwidth of specific device in my network, the Bandwidth Limiter function will disable the NAT Acceleration and slower my network speed.

As I remember, the Bandwidth Limiter function does not affect the NAT Acceleration in my old AC68u router (C1 version). But overall Blue Cave still have better performance than AC68u (if the NAT Acceleration is on), it makes no sense to switch it back to AC68u...

updated 2018/9/18

It turns out that I still need the bandwidth limiter function, and the performance of Blue Cave is terrible without NAT Acceleration. So, I switch back to AC68u....

 

[Easy Rhino]

the name change seems believable with different chip manufacturers. CTF might be a broadcom specific term