Performance of CA-10 versus new "Core 5" controller

[C4 User]

22 minutes ago, therockhr said:

clicking the watch button and then choosing the source on the remote seems quicker to me.

It is. And I use it sometimes. All I am saying is that I would be very disappointed if the OSD was eliminated and would like to see it upgraded to 4k.

[cnicholson]

15 minutes ago, C4 User said:

All I am saying is that I would be very disappointed if the OSD was eliminated and would like to see it upgraded to 4k.

The spec sheet strongly *suggests* that GUI will be 4K, although I guess it isn't explicit:  "...CORE 1 delivers an upgraded 4K HDMI output that provides a beautiful, intuitive, and responsive on-screen user interface..."

https://www.control4.com/docs/product/core-1/data-sheet/english/latest//core-1-data-sheet-rev-a.pdf

[Neo1738]

My guests use OSD in guest bedrooms almost exclusively. They use 7" TS mostly for music.

[DanITman]

19 hours ago, C4 User said:

No. I use OSD to select my video source, and even channel or app selection where available. If I’m at the tv and I want to see what is happening in one of the cameras, depending what Im watching, I may use the OSD or I may use iPad or iPhone. Or if I get up for some reason, I may use on of the T4’s. There are times that I select the source directly from the remote, but rarely. I almost always start the tv with the big red button and go from there. Same is true for my wife. And its especially true if we have a house sitter or guests using one of the TV’s.

Since I moved over to the neeo, the mini apps on the neeo make this way easier. Just pick the remote up and hit the netflix logo. I can see the camera thing being nice but not sure its worth it on every tv. To each their own

[therockhr]

Has there been one of those charts released yet that compares all the different controllers including the CORE series?

[msgreenf]

5 minutes ago, therockhr said:

Has there been one of those charts released yet that compares all the different controllers including the CORE series?

Yea

[DanITman]

9 hours ago, therockhr said:

Has there been one of those charts released yet that compares all the different controllers including the CORE series?

[therockhr]

Thanks. What I was looking for was a chart like this that listed the specs for each controller and compared it to the previous generation. This chart was nice in that it showed there is no difference in computing power between an EA-1 and EA-3.

ea-1-performance-comparison-rev-a.pdf

[cnicholson]

I'm surprised no one has done a Lua benchmark suite.   Unless the OS does per-thread throttling or something that prevents you from saturating CPU cores, I would think that would be helpful data versus trying to evaluate "snappiness."   I would not be surprised if the Core 5 has more compute power / system performance than CA-10.  

[DanITman]

1 hour ago, cnicholson said:

I'm surprised no one has done a Lua benchmark suite.   Unless the OS does per-thread throttling or something that prevents you from saturating CPU cores, I would think that would be helpful data versus trying to evaluate "snappiness."   I would not be surprised if the Core 5 has more compute power / system performance than CA-10.  

Only beta testers have access to the new CORE controllers to do a test.  I've owned both the EA5v2 and the CA-10 and the CA-10 is faster in all operations.  I think the new controllers will be running ARM CPUs so it will be interesting to see how they stack up. 

[South Africa C4 user]

13 hours ago, DanITman said:

Interesting to see the Core5 limited to 200 devices.  I am pretty sure that the EA5 was advertised (at some stage) as being capable of handling 700 or 800 drivers and devices and I would have thought that the Core5 could do the same.  I know that my EA5 was handling around 800 when I moved to a CA10.  Admittedly, it was creaking but it was also working…

[RAV]

2 hours ago, therockhr said:

Thanks. What I was looking for was a chart like this that listed the specs for each controller and compared it to the previous generation. This chart was nice in that it showed there is no difference in computing power between an EA-1 and EA-3.

ea-1-performance-comparison-rev-a.pdf 445.64 kB · 8 downloads

There are spec sheets for each posted.
The only difference noted is the Core5 has twice the ram.

My understanding is the new controllers went to a design where they can change chip at manufacture without redesign, so they aren't constrained to a specific chip production.
Whether all 3 use the same chip, or different chips in the family is difficult to say from the information provided.

[cnicholson]

Just now, South Africa C4 user said:

Interesting to see the Core5 limited to 200 devices

I assume not *really* a hard cap, right?   I think spinning it as 200 for Core 5 and 200+ for CA10 is just marketing.  I still think Core 5 might be faster.   

[therockhr]

56 minutes ago, RAV said:

There are spec sheets for each posted.
The only difference noted is the Core5 has twice the ram.

My understanding is the new controllers went to a design where they can change chip at manufacture without redesign, so they aren't constrained to a specific chip production.
Whether all 3 use the same chip, or different chips in the family is difficult to say from the information provided.

I just found the spec sheets and looked over them. It does seem that all 3 use the same CPU and that the only computing difference is that the CORE 5 has twice the RAM as 1 and 3. I think this is a smart move for Control4. No reason to dumb down the CORE 1 or 3. I would love to see them update the CA-1 with the same CPU and RAM as CORE 1 and 3. Perfect for my needs as I dont use Control4 audio or OSD and use an IO Extender.

[DanITman]

3 hours ago, cnicholson said:

I assume not *really* a hard cap, right?   I think spinning it as 200 for Core 5 and 200+ for CA10 is just marketing.  I still think Core 5 might be faster.   

No, just a recommendation.  Very low in my opinion. 

[Cyknight]

Certainly a soft cap. But I don't put a lot of value in that number either way - soooo much depends on what those devices are...and that could go both ways.

And to a a note here - that 200 devices (or other numbers) do NOT indicate the number of zigbee devices, nor could you use it to 'add up' - in other words, two core5 would not all of a sudden count as being able to handle 400 devices, that number solely counts (in as far as it does) for a system's main controller.

[cnicholson]

3 minutes ago, Cyknight said:

two core5 would not all of a sudden count as being able to handle 400 devices

Makes sense.   I assumed your main controller software runs on only one designated box and, therefore, that box needs the compute power for all your processing (your boxes don't team up and run as a compute cluster).  I guess maybe media processing might be offloaded such that one Core 5 could be wired for media and one could be left unburdened to just run the logic.  IDK.  Also don't know if there are other chips inside that offload media processing from CPU in any case (as would be the case for a general purpose computer).    

My main concern is: high-availability issues aside, am I dumb for deploying a CA-10 (in addition to the Core 5 that I am getting in any case) in search of max performance if the Core 5 would actually perform *better* (either by itself while also doing media or by using *two* Core 5's together).

[zaphod]

Over time there seems to be a big reduction in the number of relays and contacts - compare the HC-200 and HC-800 to the newer controllers.  What's the reason for this?  People not using them anymore?

[South Africa C4 user]

10 minutes ago, zaphod said:

Over time there seems to be a big reduction in the number of relays and contacts - compare the HC-200 and HC-800 to the newer controllers.  What's the reason for this?  People not using them anymore?

I hadn’t noticed that.  I must also say that I use lots of both but none on my main controller (even when it was an HC800).  The reason for me is location and thus I have used IO extenders and Z2IO a lot!

[therockhr]

25 minutes ago, South Africa C4 user said:

I hadn’t noticed that.  I must also say that I use lots of both but none on my main controller (even when it was an HC800).  The reason for me is location and thus I have used IO extenders and Z2IO a lot!

i think this is why. if its a new build where you can homerun wires back to a rack then you probably have enough things ran where you want to use an IO extender. If its retrofit, then you probably have to use Z2IO or something similar. 

Like I mentioned above, I hope they update the CA-1 with the same specs as the CORE 1 and 3 but without the IO/AV. 

[Cyknight]

yup, relatively cheap i/o extenders

[Andrew luecke]

8 hours ago, zaphod said:

Over time there seems to be a big reduction in the number of relays and contacts - compare the HC-200 and HC-800 to the newer controllers.  What's the reason for this?  People not using them anymore?

Probably because a lot of applications where you'd use it normally, people often use alarm relays these days, or use direct via IP. For something like Door access, you'd generally use a dedicated door access controller.

 

Most of the time I used the relays was for TV/projector lifts when I was an installer. Other options though have become available as well though for these situations , including RF via Bond or via IR for some of these applications.  In fact, in a lot of cases, it would be better to use a trigger from the device (and bypass C4 entirely) if possibly (particularly for projector lifts, which are best wired directly to projector

On 5/26/2022 at 8:37 AM, cnicholson said:

Makes sense.   I assumed your main controller software runs on only one designated box and, therefore, that box needs the compute power for all your processing (your boxes don't team up and run as a compute cluster).  I guess maybe media processing might be offloaded such that one Core 5 could be wired for media and one could be left unburdened to just run the logic.  IDK.  Also don't know if there are other chips inside that offload media processing from CPU in any case (as would be the case for a general purpose computer).    

My main concern is: high-availability issues aside, am I dumb for deploying a CA-10 (in addition to the Core 5 that I am getting in any case) in search of max performance if the Core 5 would actually perform *better* (either by itself while also doing media or by using *two* Core 5's together).

It makes a lot of sense to use CA-10 if you have a large number of devices, or even if you want to make life easier with troubleshooting.

The difference between a EA5 and CA10 isn't only performance, but the fact that if an EA5 fails, it anything fails, unless its the power supply, it is difficult to diagnose. With a CA10, installers can program it to  send an email on different faults increasing the efficiency of repairs (because an installer often might know what component is failing instantly, and Control4 may be able to still log into the system). Obviously, the system (including lighting) will generally also still be fully workable. 

When troubleshooting user systems, I come across far too many underpowered systems and the issues range from composer crashing during driver updates, to bad performance too. So if you have money to spare, it's always worth the extra (when it comes down to it, Home Automation is a tradeoff between cost, performance and features) 

Control4 development has also REALLY sped up recently imho (it seems like they're smashing out new features). Its also hard to say where the CPU will come in handy in 3 years time 

  

[therockhr]

On 5/27/2022 at 5:14 AM, Andrew luecke said:

When troubleshooting user systems, I come across far too many underpowered systems and the issues range from composer crashing during driver updates, to bad performance too. So if you have money to spare, it's always worth the extra (when it comes down to it, Home Automation is a tradeoff between cost, performance and features)

What are the devices/drivers/programming/whatever that puts heavy loads on a controller? My whole house runs on just a single EA-1 and looking at the resources it doesnt seem to be very loaded at all. I have 4 TV zones (4 SR remotes, Roku and Dish satellite), 9 audio zones (2 denon AVRs and Russound MCA-C5 sharing a Sonos Connect and a single Sonos One), honeywell alarm, 15 lights with Lutron Caseta, 2 Honeywell thermostats, 3 garage doors, kwikset door lock, 1 keypad, no touchscreens. Not a huge system but I wouldnt think that small either. I dont have much programming though. just a few notifications if the door is unlocked, garage door or windows opened after midnight or if the alarm is tripped. 

Just curious what things cause a system to slow down.

[Cyknight]

On 5/28/2022 at 8:34 AM, therockhr said:

What are the devices/drivers/programming/whatever that puts heavy loads on a controller?

Two-way drivers vs one way, polling drivers are worse still (ie anything using cloud or a bridge, including and usually even more involved is 3rd party music service devices, ie Sonos, BlueSound) - while programming is bad, a LOT of timers running consecutively can add up. Lot of programming in general on device changes.

 

On 5/28/2022 at 8:34 AM, therockhr said:

Not a huge system but I wouldnt think that small either

I would consider that a small system - with the understanding that single room solutions are it's own 'genre'

[Andrew luecke]

On 5/29/2022 at 12:34 AM, therockhr said:

What are the devices/drivers/programming/whatever that puts heavy loads on a controller? My whole house runs on just a single EA-1 and looking at the resources it doesnt seem to be very loaded at all. I have 4 TV zones (4 SR remotes, Roku and Dish satellite), 9 audio zones (2 denon AVRs and Russound MCA-C5 sharing a Sonos Connect and a single Sonos One), honeywell alarm, 15 lights with Lutron Caseta, 2 Honeywell thermostats, 3 garage doors, kwikset door lock, 1 keypad, no touchscreens. Not a huge system but I wouldnt think that small either. I dont have much programming though. just a few notifications if the door is unlocked, garage door or windows opened after midnight or if the alarm is tripped. 

Just curious what things cause a system to slow down.

The big hit (as mentioned) comes from devices that use polling or monitor a lot of the system state or lots of programming. Sometimes, it could also simply be caused by the way the driver is designed too. It's also difficult to know whether C4 is quoting these numbers based on the expectation of new features in their firmware in the future. Sometimes it could simply be a driver that updates a lot of things rapidly. 

Some of the things that slow the driver down too, aren't always really considered bugs (but stuff we can optimise for) 

In the case of our Shelly driver though, Beta testers will be getting a new release today, and it actually has a huge impact on eliminating CPU spikes some people were seeing every few minutes(which may have affected performance in slower systems or extremely large systems intermittently, but the users may not have even noticed it, because the mobile app still would have seemed normal, but some actions would have been slightly slower intermittently).

 

In most cases, you can use more devices though. But, I've seen a lot of struggling CA-1's out there (in particular)