Welcome!

The kernel problem with recent updates has been solved. Find the solution here

Important information
-- Required MX 15/16 Repository Changes
-- Information on torrent hosting changes
-- Information on MX15/16 GPG Keys
-- Spectre and Meltdown vulnerabilities

News
-- Introducing our new Website
-- MX Linux on social media: here

Current releases
-- MX-18.3 Point Release release info here
-- Migration Information to MX-18 here
-- antiX-17.4.1 release info here

New users
-- Please read this first, and don't forget to add system and hardware information to posts!
-- Here are the Forum Rules

Conky CPU temp history graph [solved by using xfce4-hardware-monitor-plugin]

Help for Current Versions of MX
User avatar
Stevo
Forum Veteran
Forum Veteran
Posts: 19864
Joined: Fri Dec 15, 2006 8:07 pm

Re: Conky CPU temp history graph [solved by using xfce4-hardware-monitor-plugin]

#21

Post by Stevo » Tue Jun 18, 2019 6:03 pm

Just forget about temps. Means nothing to you.
Probably famous last words at Chernobyl*...


*OK, I know that what happened is well determined, but they did disable the emergency cooling system on purpose for the tests, so someone could have said that.

User avatar
srq2625
Forum Regular
Forum Regular
Posts: 128
Joined: Mon Jan 21, 2019 10:40 am

Re: Conky CPU temp history graph [solved by using xfce4-hardware-monitor-plugin]

#22

Post by srq2625 » Wed Jun 19, 2019 7:55 am

Davo wrote:
Tue Jun 18, 2019 4:44 pm
You have a high-end CPU, but it's throttled massively (and thus very energy efficient to boot) and a low-end GPU on a modern board coupled with a gigantic psu. No reason at all to ever look at temps here.
No real world scenario that would ever deliver too high temps across the board. Unless you massively overclock or ramp up cpu and ram speeds hugely.

Just forget about temps. Means nothing to you.
I suspect (hope) you intended your last statement above to be reassuring rather than condescending and in that hope (and with only the intent of good-will) I offer the following:

I can think of at least one "real world" use-case proving this to be a false statement/assumption - all without any over-clocking of either the CPU or the ram speeds.

Photo processing RAW image files to JPG with darktable, using a more-or-less common set of filters and/or transformations, absolutely pegs the CPU (all 16 thread - I have the same CPU and virtually the same MBO and have tested this). Batch processing a large number of photos will stress (but not over-stress) the system. But, I also have an non-OEM cooling solution so my temperatures are moderately controlled. We've already seen, in another thread, that the OEM cooling solution used by Eadwine Rose has a bit more trouble moderating the CPU temperatures.
AsusTek PRIME B450-Plus
AMD Ryzen 7 2700 (8-Core, 16 threads)
GeForce GTS 450/PCIe/SSE2 v: 4.6.0 NVIDIA 390.116
Realtek RTL8111/8168/8411 - driver: r8169
Samsung SSD 860 EVO 500GB
DDR4 3200MHz CL16 2x8GB

User avatar
Eadwine Rose
Forum Veteran
Forum Veteran
Posts: 8943
Joined: Wed Jul 12, 2006 2:10 am

Re: Conky CPU temp history graph [solved by using xfce4-hardware-monitor-plugin]

#23

Post by Eadwine Rose » Wed Jun 19, 2019 8:09 am

srq2625 wrote:
Wed Jun 19, 2019 7:55 am
We've already seen, in another thread, that the OEM cooling solution used by Eadwine Rose has a bit more trouble moderating the CPU temperatures.
Thanks for your post :)

It's doing its job so far, the fan speed is following a rise in temperatures very nicely and promptly and regulates it alright imho.

Of course it is not a solid temp at all times, except for when in the bios, but that is only logical as there is no load on the thing then.


I have been keeping an eye on things using that plugin and normal is anything between 38ish and 42-44ish. Above that means there is a heavier load (like when using DVDstyler), then it goes into the low to mid 50s, throwing up all the 8 cores' percentages. Nothing to get stressed out about in any case. And when the room temp is hotter, of course the system overall is running hotter.

Then again we won't know how reliable the values are that I got on the old system, with that temp being more constant on also the OEM cooling solution, but then that was a different model CPU with much less cores and all that jazz.



Basically: I am glad I can see the values, and over time I have seen what is normal for this CPU. I have gotten used to them and am not worried about it like I was in the beginning.



As far as not needing to see values.. that is like blindly trusting everything will always be ok. Sure.. but I also like to see my gas gauge in my car, even though I KNOW how much is in the tank because I know how much and far I have driven it. ;)
MX-18.3_x64 May 26 2019 * 4.19.0-5-amd64 ext4 Xfce 4.12.3 * 8-core AMD Ryzen 7 2700
Asus TUF B450-Plus Gaming UEFI * Asus GTX 1050 Ti Nvidia 390.116 * 2x16Gb DDR4 2666 Kingston HyperX Predator
Samsung 860EVO * Samsung S24D330 & P2250 * HP Envy5030

User avatar
Davo
Forum Regular
Forum Regular
Posts: 816
Joined: Sun Jul 30, 2006 12:38 pm

Re: Conky CPU temp history graph [solved by using xfce4-hardware-monitor-plugin]

#24

Post by Davo » Wed Jun 19, 2019 2:21 pm

srq2625 wrote:
Wed Jun 19, 2019 7:55 am
Davo wrote:
Tue Jun 18, 2019 4:44 pm
You have a high-end CPU, but it's throttled massively (and thus very energy efficient to boot) and a low-end GPU on a modern board coupled with a gigantic psu. No reason at all to ever look at temps here.
No real world scenario that would ever deliver too high temps across the board. Unless you massively overclock or ramp up cpu and ram speeds hugely.

Just forget about temps. Means nothing to you.
I suspect (hope) you intended your last statement above to be reassuring rather than condescending and in that hope (and with only the intent of good-will) I offer the following:

I can think of at least one "real world" use-case proving this to be a false statement/assumption - all without any over-clocking of either the CPU or the ram speeds.

Photo processing RAW image files to JPG with darktable, using a more-or-less common set of filters and/or transformations, absolutely pegs the CPU (all 16 thread - I have the same CPU and virtually the same MBO and have tested this). Batch processing a large number of photos will stress (but not over-stress) the system. But, I also have an non-OEM cooling solution so my temperatures are moderately controlled. We've already seen, in another thread, that the OEM cooling solution used by Eadwine Rose has a bit more trouble moderating the CPU temperatures.
This should be taken as a given, I do hope. No attempt being made or else to be implied as being made is/was meant (or ever intended) in any negative way. No boasting, self-promotion or any condescending angles ever intended.
Hence my claim that there is very little to worry about vs temps. There are close to 10m+ OEM PC offerings out there with almost identical specs and no dedicated psu and/or fan upgrade. Of course I do accept your point that anything can be made to fail under the right (or else the wrong) conditions.
But OEM's do not generally bring products to market that will very likely fail sometime soon and her specific components are just very widely used in OEM land.

I was however genuinely quite curious about the steps that she chose to arrive at her very recent self-build PC. Normally you self-build to escape (or else to completely avoid) the entrapment angle from buying an off-the-shelf PC.
However, several chosen items in her chosen build are quite "opvallend" here, to use a Dutch term, which translates to kinda really "standing out". Which then begs the question of how the decisions were arrived at, with hands-on extensive research actually being applied and all manners of due diligence too.

For example the Ryzen CPU chosen is the 2-2700, which is almost exclusively designed to being bulk bought by OEMS's like Lenovo, HP, Acer etc. And indeed about 98% of market is achieved in this way. Typical order magnitudes are in the region of 250K units bulk-bought and normal minimum production lines in a Chinese factory are at $50m for best discounts. It's very intentionally throttled-back by about 25% in order to not ever directly compete with AMD's own premium product i.e. the 2700X, but is only priced at about 25 euros lower. OEM's then use this CPU in 2 ways, namely 1) to leverage the version number in advertising
by subtly hailing it's (almost) premium status, without ever disclosing that it is kinda a crippled (or throttled back) version of the 2700X (thus targeting the lower 1000~1250 dollar or euro semi high-end consumer market) and 2) to exploit it's simply superb watt usage and it's stellar energy efficiency vis-a-vie other components that need to be supplied in any off-the-shelf PC. Think about psu (nothing above about 250watt-gold really required here ... by them) and with passive cooling too. OEM's are very price conscious atm. The 2700X draws about double the wattage btw, assuming no overclocking applied to either model Normally very few ever select the 2-2700, as it's more of a window-filler in self-builds. In self-builds, about 5 cpu's (both red and blue) get chosen 90% of the time and the other 100x offerings generally get passed over, mostly anyways.

The 1050ti geforce nVidia is also a fairly odd choice here. Again about 98% of it's sales are (or were once) gained from bulk orders from OEM's and very few from the direct self-build market. It IS available out there as a build option of course, but since late March of this year not so much anymore as it has even there been superseded and supplanted by the latest-greatest offering from nVidia namely the 1660ti (based on newest Turing stuff), which I do have myself atm and do like a lot as well. Or the more crippled 1660 (non-ti) and the even more crippled 1650. All sustained favorites of the OEM marketplace, as you (again) can always advertise a dedicated video card on-board and can thus claim superb gaming abilities (you just do not then mention that only 1080p applies though, or in the case of the 1660ti that only 1440p really applies as a hard limit). Plus passive cooling and very low wattage draws also allies very well with then shipping of minimum level psu's and/or no additional cooling stuff. 1050ti is not cheap by any means either atm (bitcoin inflation?), suffices for 1080p gaming atm anyways and is more than 99% of consumers now ever need, but it is very low-end qua gaming.

The 3rd choice that kinda stands out is the Kingston HyperX predator ddr4 Ram, clocked at 2666, albeit with a wonderful CL latency of just 13. However this is RAM exclusively designed for Intel CPU's primarily (consult the .pdf on their site). It is also dual rank ram, which can run a bit better/faster/more efficient than single rank ...but not quite so much with the Gen2 Ryzen Cpu's. To overclock this ram to normal AMD performance levels (XMP) and to then still retain the same great latency aspect is quite very highly problematic with Ryzen2. Intel i5 or i7 cpu's are no problem at all though. I normally use g.Skill with Ryzen, but only their ripjaw or trident stuff that specifically marries very well here (or else corsair vengeance ram which is also fully amd focused too). The g.skill sniper ddr4 stuff is very highly intel specific though in very many ways. Of course, you can 100% run HyperX under Ryzen too in a sort-of basic mode of operation. Kingston is from LA btw, has no factories and they just re-brand Korean SK.Hynix (or else Micron ram) ram as their own. Normally SK.Hynix also does supply the no-name ram that is mostly used by HP/Lenovo/Dell/Acer etc OEM's , to cut some costs.

Quite a fairly unique self-build ,in some ways - in that it almost equates to purchasing off-the-shelf in very many ways ... except with a racked-up psu wattage, fully dedicated coolers and a very high-end ssd added-in. Maybe better wiring too and a nicer case. Then temps should (ideally) be a non-issue here, as so many OEM's ship almost same-same stuff with about 75% less in the way of safeguards here. Which they would no do, if it did not prove to be adequate enough in testing the build scenarios.

Would be quite interesting to explore how the actual build choices are or were made or arrived at. As humans we kinda very fast arrive at a first option decision, then very often use confirmation bias to then justify this initial decision, then sometimes wildly depart from first to then the next option being considered.
Just to be clear, no criticism is ever intended (or is to be implied) ...and this build is just 99% good-to-go for the next 5 years probably/possibly (depending on tech advances meantime) ...but it is just a very curious alignment of build choices, either because of too much research or else of too little. Even under-specc'd the 2-2700 will deliver more than 99.9% of consumers/prosumers would ever use etc. So a moot point maybe, but normally you self-build to totally avoid the whole industry that pushes you in one direction

User avatar
Eadwine Rose
Forum Veteran
Forum Veteran
Posts: 8943
Joined: Wed Jul 12, 2006 2:10 am

Re: Conky CPU temp history graph [solved by using xfce4-hardware-monitor-plugin]

#25

Post by Eadwine Rose » Wed Jun 19, 2019 2:59 pm

Wow.. that is a LONG bit of text. If you look up and read through the build thread a lot of questions as to why I bought/chose what I did will be answered.

I have my reasons and the decision is mine to make, no matter what anyone else might think of them, nor is it going to make a difference discussing all of my choices now that the deed is done. Also.. very important: I don't decide on parts depending on what they were allegedly built for, I do not care about that one iota. It seems you are giving that quite some value. I however do not.


As for the comment you made on my wanting to see temperatures being intended to be reassuring, it sure did not come across that way, hence my response. Being told by someone what to think or feel is not something I will let anyone do. This long post isn't helping me get rid of that feeling.



You would do the build differently, sure, then please.. do so. Go ahead and build your own system, assemble it all yourself, and you can do it exactly the way you like. These were my choices and I still stand by them.
MX-18.3_x64 May 26 2019 * 4.19.0-5-amd64 ext4 Xfce 4.12.3 * 8-core AMD Ryzen 7 2700
Asus TUF B450-Plus Gaming UEFI * Asus GTX 1050 Ti Nvidia 390.116 * 2x16Gb DDR4 2666 Kingston HyperX Predator
Samsung 860EVO * Samsung S24D330 & P2250 * HP Envy5030

User avatar
richb
Administrator
Posts: 19268
Joined: Wed Jul 12, 2006 2:17 pm

Re: Conky CPU temp history graph [solved by using xfce4-hardware-monitor-plugin]

#26

Post by richb » Wed Jun 19, 2019 3:46 pm

When I do my build I will keep my choices to myself. :happy:
Forum Rules
Guide - How to Ask for Help

Rich
SSD Production: MX 18.1
AMD A8 7600 FM2+ CPU R7 Graphics, 16 GIG Mem. Three Samsung EVO SSD's 250 GB, 350 GB HD

User avatar
Eadwine Rose
Forum Veteran
Forum Veteran
Posts: 8943
Joined: Wed Jul 12, 2006 2:10 am

Re: Conky CPU temp history graph [solved by using xfce4-hardware-monitor-plugin]

#27

Post by Eadwine Rose » Wed Jun 19, 2019 3:53 pm

I think that is a smart thing to do Rich hahahaha


Do share it with me though wouldya? You know I'll think everything is ok as long as you're happy with it :)
MX-18.3_x64 May 26 2019 * 4.19.0-5-amd64 ext4 Xfce 4.12.3 * 8-core AMD Ryzen 7 2700
Asus TUF B450-Plus Gaming UEFI * Asus GTX 1050 Ti Nvidia 390.116 * 2x16Gb DDR4 2666 Kingston HyperX Predator
Samsung 860EVO * Samsung S24D330 & P2250 * HP Envy5030

User avatar
Davo
Forum Regular
Forum Regular
Posts: 816
Joined: Sun Jul 30, 2006 12:38 pm

Re: Conky CPU temp history graph [solved by using xfce4-hardware-monitor-plugin]

#28

Post by Davo » Thu Jun 20, 2019 2:22 pm

Eadwine Rose wrote:
Wed Jun 19, 2019 2:59 pm
Wow.. that is a LONG bit of text. If you look up and read through the build thread a lot of questions as to why I bought/chose what I did will be answered.

I have my reasons and the decision is mine to make, no matter what anyone else might think of them, nor is it going to make a difference discussing all of my choices now that the deed is done. Also.. very important: I don't decide on parts depending on what they were allegedly built for, I do not care about that one iota. It seems you are giving that quite some value. I however do not.

As for the comment you made on my wanting to see temperatures being intended to be reassuring, it sure did not come across that way, hence my response. Being told by someone what to think or feel is not something I will let anyone do. This long post isn't helping me get rid of that feeling.

You would do the build differently, sure, then please.. do so. Go ahead and build your own system, assemble it all yourself, and you can do it exactly the way you like. These were my choices and I still stand by them.
No criticism ever intended or else implied from me. More the abstraction angle intended.

if presenting as a public speaker, then primacy/regency angles have been investigated cq discussed at length in very many scientific papers post about 1962. Typically the first 5% and the last 5% of any public speaking are very well retained by the audience (in minutes about the first 3 and the last 3 of an hour lecture) at about 75% rates and everything in-between is retained at about rates of less-than 10%, which equates to a blur. Normally all public (or in-house) speakers were always taught to present to the public a light-hearted anecdotal-based introduction first-off then to present the main stuff mid-lecture and to conclude with a small recap. Not any more though. Now, you must offer key points instantly and then reiterate in summation. Same with job applications. Either lightning talks or extensive TED talks, or presenting to Mozilla or to Debcon or Suse confs. First and last tiny moments are well absorbed in human memory, everything in-between is simply lost.

Same idea with an ultra-log build log/blog/endlessly long forum discussion etc. Only first and last impressions are conventionally retained.

The very best idea here is to always try to deploy the Johair principle of cognitive research, available since circa late '60's in cognitive psychology. Map a single 4x box of known-knowns, unknown-knows, known-unknowns and unknown-unknowns fields. This method has been used by NASA, CIA, FBI, plus all military instances for the past 50 years etc. Also 100% applicable to self-builds. Every post and/or reply then auto gains 100% equal value and weight and is mapped to one of the 4 boxes - which is 1) all agree that this X is good, or subject states something that no one else mirrors or peers states something that subject does not know or else
Subject and peer interaction here is indifferent and what comes out of it. This is quite a very revealing method which is to assign equal value to all comments (and not only to just first and last).

BTW: I did self-build btw last 2 months. Cost maybe one hour max to assemble and with no issues whatsoever since. Not boasting btw. All Linux boots are just simple plug-and-play. since then. Assemble and then instantly forget. Good/bad who knows? Will be obsoleted in just 2.5 years anyways, very probably.

Again no criticism ever intended or implied.

Using Intel specific ram in a Ryzen set-up is 100% up-to-you stuff ...and it will work anyways and 99.99999% of normal users don't ever need to ever align the ram properly anyways in xmp stuff (which manually doing means setting about 30x parameters) and will have no issues with running ram at 40% less than it's actual capabilities in everyday life. If you are happy, then great. Good for you, but not quite best practices in general.

Again to over-emphasise no criticism is ever to be intended or else implied.

User avatar
Stevo
Forum Veteran
Forum Veteran
Posts: 19864
Joined: Fri Dec 15, 2006 8:07 pm

Re: Conky CPU temp history graph [solved by using xfce4-hardware-monitor-plugin]

#29

Post by Stevo » Sun Jul 07, 2019 5:09 pm

It turns out that the xfce4-hardware-monitor-plugin does build and display on the MX 19 alpha, so you can continue to use it! I'm sending the new packages up.

User avatar
richb
Administrator
Posts: 19268
Joined: Wed Jul 12, 2006 2:17 pm

Re: Conky CPU temp history graph [solved by using xfce4-hardware-monitor-plugin]

#30

Post by richb » Sun Jul 07, 2019 5:57 pm

Stevo wrote:
Sun Jul 07, 2019 5:09 pm
It turns out that the xfce4-hardware-monitor-plugin does build and display on the MX 19 alpha, so you can continue to use it! I'm sending the new packages up.
Good news. One of my favorite plugins.
Forum Rules
Guide - How to Ask for Help

Rich
SSD Production: MX 18.1
AMD A8 7600 FM2+ CPU R7 Graphics, 16 GIG Mem. Three Samsung EVO SSD's 250 GB, 350 GB HD

Post Reply

Return to “MX Help”