OT: Laptop Render Benchmark Results

ebergerlyebergerly Posts: 3,255
edited November 2017 in The Commons

In another thread there was a discussion regarding render performance of laptops. I said I would test my piece of junk laptop to see if it could render continuously without problems. I think the main issue we’re trying to resolve is this:

  • Can even a very low end laptop render continuously without throttling (ie, slowing down its rated performance) or even shutting down completely due to thermal issues?

Now clearly we can’t resolve that for every laptop out there, but I think if we test some lower end devices at least we can see if they seem to maintain a safe temperature over time. Also, and more importantly, if the render time per frame stays constant, it’s a good indication that performance throttling hasn’t occurred. Presumably, if low end devices can perform, then higher end devices should also perform (other than some specific BIOS issues, etc.). 

So I’m testing the following laptop:

  • Intel® Core™ i3 3217U Processor (2 cores, 4 threads, 1.8GHz, 17 watts TDP)
  • 2 GB DDR3 1600MHz SDRAM
  • Integrated Intel HD Graphics
  • Linux Mint

And I’m rendering using the following software:

  • Blender 2.78
  • BMW 27 benchmark scene

My results so far (I've only rendered ONE frame):

  • Render time per frame = 1 hour 20 minutes
  • CPU Temperature: rose to 62C shortly after render started and stayed at 62C through the 80 minute render;
  • Fan speed: rose to 3200 rpm shortly after render started and stayed at 3200rpm through the render.

Attached is a snapshot of the render, along with a Linux thermal sensor app showing CPU temps and fan speed. Note that in the yellow highlighted areas, the “High” rating for that CPU is shown as 87C, and the “Critical” rating is 105C. Presumably, High is when performance throttling occurs, and critical is when the system is shut down by the BIOS. So far, the 80 minute peak temperature is 25C below even the “high” rating.

My plan is to render maybe 20 frames of the same scene (as an animation), and verify if the temperatures stay at or below 62C throughout, and whether the total render will complete in about 1600 minutes (20 frames x 80 minutes per frame). And since 1600 minutes is over 26 hours, so that should be a reasonable indication of continuous performance.

BTW, I'm a bit surprised that this laptop's peak CPU temperatures are almost identical to my Ryzen 7 1700 CPU temps during long renders. 

Comments and suggestions are welcome. 

LaptopRender.png
1310 x 689 - 440K
Post edited by ebergerly on
«13

Comments

  • nonesuch00nonesuch00 Posts: 18,032
    edited November 2017

    I recently in September bought a HP 8470P with i5-3230qm (2 core - 4 threads) as a medical waste product from eBay for $50 and the CPU, since saved and put away - is wisper quiet even when rendering and stays cool although not 62C cool like yours but very close. It would render a BOSS Lighting scene like the boxer in my gallery at 2000 iterations in about 60 hours.

    I then turned around and bought an i7-3630qm CPU with 4 cores and 8 threads for that laptop again as medical waste from eBay for another $50. That thing is loud as soon as I render or even open a program. The loudness during renders though stays loud for the duration of the render. It immediately shots up in temperature too when I render and instantly cools off when rendering starts. So quickly in both hot and cold that it is physically impossible from it to actually heat up that hot so quickly or cool off so cool that quickly. It also runs at 3.3GHz which it is only supposed to do rarely when it needs that. There is something odd about that CPU. I think it's been tampered with to speed it up. Anyway that same DAZ Studio boxer scene in my gallery takes only aboy 27 hours with the i7-3630qm CPU so less than half the time of the i5-3230qm (or maybe it's i5-3220qm).

    The good thing about both of those though is they do directX11.1, openGL 4.0, and openCL 1.2. My previous laptop that had an i5-2520m CPU did only directX10.1, openGL 3.1, and no openCL at all so it was very limiting with DAZ Studio, Blender, and other SW.

    Post edited by nonesuch00 on
  • ebergerlyebergerly Posts: 3,255
    edited November 2017

    After 3 hours of rendering the temperatures and fan speed have been surprisingly consistent. Temps between 61C and 63C (far below the 85C high rating), and fan speeds from 3100-3300 RPM. And consistent render times per frame of about 1:20.

    Anyway, I'm expecting the 20 frames to finish in a little less than 24 hours unless there is some throttling, so we'll see if we're still getting 1:20 and 62C tomorrow. 

    Post edited by ebergerly on
  • JamesJABJamesJAB Posts: 1,760

    I would not expect any of your machines to overheat.  (even if they get hotter than you as the end user want it too)  All modern Intel CPUs are rated for opperation at junction temperatures (AKA Core temp) up to 105c.  Theyh also have advanced power management systems integrated into the CPU that will reduce power delivery and keep the CPU under the 105c cap.

    On a side note, your i5 is most likey the Core i5 3230m.  the QM designator is the notation for quad core mobile chips, and I've only ever seen it used on the Core i7 line.
    https://ark.intel.com/products/72056/Intel-Core-i5-3230M-Processor-3M-Cache-up-to-3_20-GHz-BGA

    The machine getting loud makes sense as your old i5 is a 35w CPU and the new i7 is a 45w CPU.  Looking at the cooling setup this machine uses a single heatpipe and fan setup to cool both your CPU and discrete GPU.  (This is one of my things that is on my "No Go" list when laptop shopping)

    On the bright side, once you pull that cover off the bottom of the computer, you only need to remove 2 screws and you have full access to cleaning the fan side of the heatsink (This is where all the dust bunnies live).
     

  • ebergerlyebergerly Posts: 3,255

    Okay, after 6 hours of continuous rendering the temperatures and fan speed haven't changed from 62F and 3200RPM. And the render time per frame is still about 1:20 (the last frame was about 2% longer, or 1:22). 

  • ebergerlyebergerly Posts: 3,255

    After 14 hours of continous rendering the temperatures and fan speed havent changed from 62F and 3200 RPM. Though temps now are a bit lower at 61F because the room temperature has dropped overnight from 69F to 60F. Render time per frame is still around 1 hour 20 minutes (the last was 1 hour 24 minutes). 

  • ebergerlyebergerly Posts: 3,255

    Well we're coming up on 22 hours of continuous rendering by my piece of junk laptop, and CPU temps and fan speeds and render times per frame have been rock solid the entire time. CPU temps are 61-63C, which is FAR below the "high" temps (presumably where performance throttling occurs) of 87C, and super far below the "ciritical" (ie, shutdown) temp of 105C. Fan speeds have been solid at 3200-3300 rpm the entire time, and most importantly, render times per frame have been solid at about 1 hour and 20 minutes, though occasionally I'd get a 1:22 frame. Perhaps that's when the OS was doing some internet updating or whatever. Though the difference is minor, and a fairly clear indication that no performance throttling has occurred. 

    And if you look at a resulting render below, it did pretty well, IMO. Looks like a professional render to me. 

    Now, to certify it is actually a piece of junk (as I recall it's about 5 years old), I checked Newegg to see if you can even buy an i3, 1.8Ghz, laptop with only 2GB of ram. And all they have is one used one for sale. Though there's a bunch of refurbished ones with 4GB of ram, but nothing new. So unless you have a mobile 286 or whatever from the 90's, I don't think you can get more of a piece of junk.

    Now of course, with any laptop or workstation desktop, if you're using an i3 with internal graphics, you're not going to get the best render times. However, being a laptop doesn't automatically mean it can't handle continuous rendering. I'm more convinced that designers design their systems to provide continuous performance for the components they include in the design. Unless of course there are some strange BIOS issues where fans aren't correctly controlled based on the temps.  

    Anyway, I'll let it continue for a few more hours until it finishes, but I think the answer is clear. It would be interesting if others could do something similar... 

     

    BMW.PNG
    1210 x 682 - 990K
  • nonesuch00nonesuch00 Posts: 18,032
    edited November 2017

    Thanks for the info JamesJab & ebergerly.

    My now i7-3630qm HP 8470p Elitebook does have the type heatsink & fan as pictured. It was very easy to change the CPU. I saved the old CPU because my renders immiediately cause the temperature to shot up and fluctuate between 90C - 104C. The temperatures fluctuate as the rendering parent process doles out work for child threads to the different threads on the cores.

    I am using Core Temp 1.11 to monitor the temps. I never bothered to monitor my old HP 8460P as I never considered that it was overheating but as I found it must have but, and this is the tricky part, not on the CPU but on the GPU and the system board area associated with the GPU. Even if your CPU never gets near 105C to ruin, constant rendering will cause your system board to break circuits in the area of the GPU, at least on these older model laptops. The problem is often hidden until you turn off the laptop and then turn it back on after it has cooled completely down. I reckon the newest have even lowered temperature but improved performance to even solve that problem.

    My last laptop, the i5-2520m HP 8460P Elitebook actually broke when I turned it off and let it completely cool down over a week. I know it was because the system board cooled off after expanding from the sustained heat of running 24/7 and often being used to do DAZ Studio renders. As the chassis & rest of the laptop is so nice & functional I will replace the system board, but since I don't use it I'm not in a hurry.

    Post edited by nonesuch00 on
  • nonesuch00nonesuch00 Posts: 18,032
    edited November 2017
    ebergerly said:

    Well we're coming up on 22 hours of continuous rendering by my piece of junk laptop, and CPU temps and fan speeds and render times per frame have been rock solid the entire time. CPU temps are 61-63C, which is FAR below the "high" temps (presumably where performance throttling occurs) of 87C, and super far below the "ciritical" (ie, shutdown) temp of 105C. Fan speeds have been solid at 3200-3300 rpm the entire time, and most importantly, render times per frame have been solid at about 1 hour and 20 minutes, though occasionally I'd get a 1:22 frame. Perhaps that's when the OS was doing some internet updating or whatever. Though the difference is minor, and a fairly clear indication that no performance throttling has occurred. 

    And if you look at a resulting render below, it did pretty well, IMO. Looks like a professional render to me. 

    Now, to certify it is actually a piece of junk (as I recall it's about 5 years old), I checked Newegg to see if you can even buy an i3, 1.8Ghz, laptop with only 2GB of ram. And all they have is one used one for sale. Though there's a bunch of refurbished ones with 4GB of ram, but nothing new. So unless you have a mobile 286 or whatever from the 90's, I don't think you can get more of a piece of junk.

    Now of course, with any laptop or workstation desktop, if you're using an i3 with internal graphics, you're not going to get the best render times. However, being a laptop doesn't automatically mean it can't handle continuous rendering. I'm more convinced that designers design their systems to provide continuous performance for the components they include in the design. Unless of course there are some strange BIOS issues where fans aren't correctly controlled based on the temps.  

    Anyway, I'll let it continue for a few more hours until it finishes, but I think the answer is clear. It would be interesting if others could do something similar... 

     

    That looks good. An old laptop will be able to produce images that are just as professional as a workstation with the most expensive nVidia video cards but will take much longer to do so.

    For me that isn't a problem except it severely limits the speed of which I can learn or experiment with artistic looks via DAZ Studio and other software via their rendering and other settings like light, surfaces, and so on.

    This is the image produced by 60 hours of render time in DAZ Studio 4.10.x on an HP 8470P i5-3230m with 16GM RAM. The CPU never exceeded the mid-60Cs in temperature either. It was 2000 iRay iterations. I also used the dForce and the catalyst product from the LAMH team.

    I am running the same scene now again but with my new CPU, the i7-3630qm. I will post results tomorrow when it probably should be done. Only 2000 iterations again.

    Post edited by nonesuch00 on
  • ebergerlyebergerly Posts: 3,255
    edited November 2017

    nonesuch, thanks. Regarding your issue with component failure on your old laptop:

    I think it's reasonable that if the cooling system can maintain the component temperatures at a safe level, like I showed with my piece of junk tests, then presumably as long as that cooling/fan system is functional then the temperatures would stay in the safe range (eg, 65C), even if you kept rendering for a month, right? 

    So then the question is, is the 65C safe temperature really safe over the long term? Or maybe it's safe for 24 hours but not 96 hours? Even though it's way below the "high" and "critical" temperature ratings? 

    Maybe designers set the cooling for a safe, continuous temperature that isnt' really safe and continuous, but I really doubt it. But when it comes to component damage I think it's a lot more complex than that, and a LOT of other things can cause damage, and those things can REALLY cause temps to spike and possibly cause damage (even though the BIOS should sense that and shut things down first). Stuff like failure of the BIOS to control the fans correctly, blocked fans and vents, component failure unrelated to temps (old age, etc.), mechanical damage, and so on.

    Yeah, if temps regularly get up past the "high" range then I can understand component issues. But maybe if they do that there's something else going on?        

    Post edited by ebergerly on
  • nonesuch00nonesuch00 Posts: 18,032
    edited November 2017
    ebergerly said:

    nonesuch, thanks. Regarding your issue with component failure on your old laptop:

    I think it's reasonable that if the cooling system can maintain the component temperatures at a safe level, like I showed with my piece of junk tests, then presumably as long as that cooling/fan system is functional then the temperatures would stay in the safe range (eg, 65C), even if you kept rendering for a month, right? 

    So then the question is, is the 65C safe temperature really safe over the long term? Or maybe it's safe for 24 hours but not 96 hours? Even though it's way below the "high" and "critical" temperature ratings? 

    Maybe designers set the cooling for a safe, continuous temperature that isnt' really safe and continuous, but I really doubt it. But when it comes to component damage I think it's a lot more complex than that, and a LOT of other things can cause damage, and those things can REALLY cause temps to spike and possibly cause damage (even though the BIOS should sense that and shut things down first). Stuff like failure of the BIOS to control the fans correctly, blocked fans and vents, component failure unrelated to temps (old age, etc.), mechanical damage, and so on.

    Yeah, if temps regularly get up past the "high" range then I can understand component issues. But maybe if they do that there's something else going on?        

    Well your laptop is safe at 62C I think, even the GPU portion of it that doesn't reside on the CPU die (or so that is true if what I read  was correct - they as an HP HW support tech should know more than I do about that). And if I were to continue to use my i5-3230m CPU I think that would last. With my i7-3630qm CPU though that I am using now, I think however were I to do sustained rendering frequently that over the course of a year or two my system board would be ruined at the area on it where the Intel HD Graphics 4000 chips and i/o controllers and whatever reside.

    I plan to model and animate more in Blender and in DAZ do pose sets and animations but in both to use shader styles such as the pwToon instead of iRay PBR materials though so I likely won't burn out this laptop like the last one.

    Post edited by nonesuch00 on
  • JamesJABJamesJAB Posts: 1,760
    ebergerly said:

    nonesuch, thanks. Regarding your issue with component failure on your old laptop:

    I think it's reasonable that if the cooling system can maintain the component temperatures at a safe level, like I showed with my piece of junk tests, then presumably as long as that cooling/fan system is functional then the temperatures would stay in the safe range (eg, 65C), even if you kept rendering for a month, right? 

    So then the question is, is the 65C safe temperature really safe over the long term? Or maybe it's safe for 24 hours but not 96 hours? Even though it's way below the "high" and "critical" temperature ratings? 

    Maybe designers set the cooling for a safe, continuous temperature that isnt' really safe and continuous, but I really doubt it. But when it comes to component damage I think it's a lot more complex than that, and a LOT of other things can cause damage, and those things can REALLY cause temps to spike and possibly cause damage (even though the BIOS should sense that and shut things down first). Stuff like failure of the BIOS to control the fans correctly, blocked fans and vents, component failure unrelated to temps (old age, etc.), mechanical damage, and so on.

    Yeah, if temps regularly get up past the "high" range then I can understand component issues. But maybe if they do that there's something else going on?        

    Well your laptop is safe at 62C I think, even the GPU portion of it that doesn't reside on the CPU die (or so that is true if what I read  was correct - they as an HP HW support tech should know more than I do about that). And if I were to continue to use my i5-3230m CPU I think that would last. With my i7-3630qm CPU though that I am using now, I think however were I to do sustained rendering frequently that over the course of a year or two my system board would be ruined at the area on it where the Intel HD Graphics 4000 chips and i/o controllers and whatever reside.

    I plan to model and animate more in Blender and in DAZ do pose sets and animations but in both to use shader styles such as the pwToon instead of iRay PBR materials though so I likely won't burn out this laptop like the last one.

    For your HP, I would probably look into setting a more agressive fan speed profile. 
    My Desktop GTX 1080 ti on the stock profile lets the GPU get up to 85c before setting the fans to 100% and at idle it shuts the fans off completely with gpu idle temps hitting around 60c...
    I set mine so the fans never completely shut off and spin up to 100% at 70c.  Now the card idles at 29c and renders at 73c.

    Your HP is probably turning off the cooling fan when it desn't think that it's needed.  When you hit render the chip heats up quick and the computer says "Oh crap, where did this come from!" and kicks the fan up to 100%  I would see if you can force the fan to 100% using system monitoring software before hitting render.

  • Kevin SandersonKevin Sanderson Posts: 1,643
    edited November 2017

    Since I don't have a laptop, ebergerly, could you render a scene for a few frames in Iray? By anecdotal accounts, not by your preferred hard data as none exists, Iray really stresses out a system full tilt - Blender Cycles doesn't according to anecdotes. My PC hits over 70c when rendering in Iray, maybe case and fans issues, though they roar when rendering in Iray, though it's wonderful for most tasks. It's at 61C now as I type this and fans are quiet. CPU heats up and fans speed up when rendering in 3Delight since it runs full bore, and heats my apartment a little, but is still stable. Long term Iray use is not stable, it really heats up and DAZ Studio crashes and will not start until my PC cools down. It might be a revealing test for a laptop.

    Post edited by Kevin Sanderson on
  • TheKDTheKD Posts: 2,677

    Hmmmm, you might want to look into getting a large case with top vents and plenty of fan housings. My PC hs two intake fans in the fron, one in the back for exhaust, and a top vent, GPU and CPU never even hits 70. My 1070 gets some heat from the 960 under it, but still never hits 70, it runs a few degrees hotter than the 960.

  • Kevin SandersonKevin Sanderson Posts: 1,643
    edited November 2017

    That's what I was thinking, but it only does that when running Iray, so I thought it would be a more suitable stress test for ebergerly to try on a laptop since that's what the thread is about. My case is already pretty big and has appropriate fans and vents for most tasks. I've been looking at a case with more cooling just because of Iray renders causing issues.

     

    Post edited by Kevin Sanderson on
  • wsgentrywsgentry Posts: 572

    I used to use an 8 year-old Dell XPS 17.  It was perfectly fine.  I used to let it render for up to 24 hours.  Now I have an Alienware R417, naturally, it's much faster with rendering.  I have the graphics amplifier with a GTX 1080ti in it.  Renders that used to take 24 hours are now taking about an hour or so. 

  • ebergerlyebergerly Posts: 3,255

    Since I don't have a laptop, ebergerly, could you render a scene for a few frames in Iray? By anecdotal accounts, not by your preferred hard data as none exists, Iray really stresses out a system full tilt - Blender Cycles doesn't according to anecdotes. 

    I guess it depends on what you mean by "full tilt" smiley

    I think that's one of the issues in the tech world, people get these general ideas without really knowing exactly what it is in the real world. For example, when I rendered that 24 hour Blender scene, my CPU was at 100% utilization the entire time. Which means it was processing data continously, on both cores. Now, would Iray somehow make even more use of those cores than Blender cycles? Would the CPU utilization show 120% or something? I kinda doubt it. smiley

    In any case, I'm not sure how I would do an iray render on this since it's a Linux machine and I don't think I set up Wine on it. And honestly, I think if folks still think that laptops aren't "as good" as desktops when it comes to continuous use, I think it's up to them to provide their own data to prove their point. There's only a certain amount of time I'm willing to spend on stuff like this. But knowing me, I'll probably take it as a Sunday challenge and see if I can come up with something. Unless I get distracted by an "OOO SHINY !!"  smiley

  • ebergerlyebergerly Posts: 3,255
    edited November 2017

    Also, keep in mind that my results showed the cooling system maintaining about 62C. And the "high" limit for this CPU is 87C. That's the point that performance throttling occurs. Not damage, just throttling. And the danger point, when the system shuts down, is 107C. So even if Iray somehow made it run 25C hotter, you're still at just the throttling point, where the CPU would run slower. You'd still render the frame, but it would take longer. Only if Iray cause it to run 45C hotter would the system shut down. Which means not only did Iray do something special to cause the CPU to run "more", but the cooling system was incapable of keeping up. 

    Now, like I said, there are a lot of things that can cause temperatures to increase. One time you're rendering with the laptop in your lap and you're blocking a vent with your pants, or there's a BIOS issue, or there's dust in the filter, or the thermal paste needs replacing, and so on. 

    I don't know, maybe I'm missing something, but from what I've seen after monitoring CPU and GPU temps on my desktops and this laptop, it all seems pretty consistent. 60's or 70's C for continuous rendering is normal, and the kind of temps I'm always seeing, whether it's Iray or Cycles or whatever. Seems like for CPU's and GPU"s, that's what designers design their cooling systems to maintain, because that's what semiconductors and stuff likes. Yeah, if it got into the mid-high 80's or 90's I'd start to worry, but I just haven't seen it. If others have, and are sure there's not some other issue causing a heating problem, it would be nice if they could provide some test data. It's really not that difficult to set up some renders and let it go while you watch temps and render times.  

    Post edited by ebergerly on
  • FishtalesFishtales Posts: 6,104
    edited November 2017

    My laptop idles around 65ºC and renders at 89ºC and drops to 82ºC for a little before rising again.

    The longest I have had a render run was an animation which took four days during which I had another Studio running and I did renders on it at the same time.

    ScreenShot_20171119114813.png
    430 x 390 - 16K
    ScreenShot_20171119114953.png
    430 x 390 - 16K
    Post edited by Fishtales on
  • ebergerlyebergerly Posts: 3,255
    edited November 2017
    Fishtales said:

    My laptop idles around 65ºC and renders at 89ºC and drops to 82ºC for a little before rising again.

    Now that's something I'd start to be concerned about. smiley

    Sounds like the throttling is probably kicking in at the high 80's, so probably not a problem. It's still staying below the damage point (over 100C or whatever). But I'd worry there's some other issue going on that's causing it to idle so high. 

    IMO, if it's IDLING that high there might be a system problem going on. 

     

    Post edited by ebergerly on
  • ebergerlyebergerly Posts: 3,255

    Fishtales, have you opened up the laptop and looked inside to see if there's some blockage or something? If all looks clear I'd tend to think about replacing thermal paste. Also check if your fans are spinning at the correct speed. Maybe you need a BIOS update. Sounds like if it's a W7, i5 it's a few years old?

  • ebergerlyebergerly Posts: 3,255
    edited November 2017

    Fishtales, I took a look at Toms Hardware, and found this regarding the Intel CPU's:

    "If your hottest Core is near it's specified Tj Max Throttle temperature, then your CPU is already too hot. Experts agree it's prudent to observe a reasonable thermal margin below Tj Max. So regardless of environmental conditions, hardware configurations, workloads or any other variables, Core temperatures above 85°C aren't recommended."

    Now, apparently for the i5 the TjMax is the important measurement, and to measure it you can use a free (?) app called "Core Temp". It gets kind of complicated, based on exactly what part of the CPU is being measured (core temps, case temps, TjMax, etc.), but it sounds like this app will give you the important TjMax measurement. You might want to try it and see if something's going on with your laptop. 

    Post edited by ebergerly on
  • FishtalesFishtales Posts: 6,104

    When I say idle there is always something running in the background and Opera is always open with at least ten tabs open at any one time :) It has been running like that for at least five years, if not longer, and is only shut down for updates, when I move from one location to the other and then it is only in sleep mode so it starts where I left off, and when I go on holiday abroad for four weeks or so a year. As I said in a previous thread I had to replace the fan as it wore out and was making one heck of a racket every time the RPM increased :) My last laptop the plastic base plate warped with the heat and it still worked although the hard drive finally failed which is why I got this one. The old one was a lower spec but it still works and I will replace the hard drive at some point, I've been saying that since it died. I did buy another hard drive but ended up using it as a drive for my daz library on this laptop instead :) I also have a 3.5 inch external drive connected as a backup drive and I also use it for video work. I don't believe in mollycoddling stuff and generally use things as I want rather than how they are supposed to be used. If they can't keep up they are replaced. I have had my desktop for thirty years or so and it is on its third motherboard and is due for another one as there is something overheating in it and it keeps shutting down. It is about ten years old so no shame for it :)

  • ebergerlyebergerly Posts: 3,255
    edited November 2017

    Also, they mention this about desired idle temperatures, and that matches what I've seen in practice with my PC's:

    "2nd through 8th Generation processors should idle less than 10°C above Ambient. Certain 1st Generation variants and most Core 2 processors may idle several degrees higher."

    So if your room is at, say, 73F/23C, then your CPU idle temps should be around 30-35C.

    Post edited by ebergerly on
  • ebergerlyebergerly Posts: 3,255
    edited November 2017

    BTW, I just checked my desktop (Ryzen 7 1700 plus GTX 1080ti and GTX 1070) and with a room ambient temperature of 70F/21C, my idle temps are as follows:

    • CPU: 32C
    • 1080ti: 33C
    • 1070: 47C (hmmm...not sure why this is higher at 0% load...maybe the Founders Edition runs warm?)

    EDIT: Oh, apparently because the 1070 is running the 3 monitors it has to do stuff that warm it up in idle...

    Post edited by ebergerly on
  • ebergerlyebergerly Posts: 3,255
    edited November 2017

    Wow, here's something I didn't know...apparently the "throttling" point, when the CPU speed is automatically lowered to bring temperatures down, is the 100C point, not the 87C (or whatever) point. Maybe the "high" point is just when the fans get boosted:

    " Also called "Tj Max" (Temperature Junction Maximum), this is the Thermal Specification that defines the Core temperature limit at which the processor will Throttle (reduce Core speed) to protect against thermal damage. Throttle temperatures are shown below for several popular processors, including Thermal Design Power (TDP) and idle Power, which are expressed in Watts (W).

    Core

    8th Generation 14 nanometer i7 8700K / i5 8600K (TDP 95W / Idle 2W),
    7th Generation 14 nanometer i7 7700K / i5 7600K (TDP 91W / Idle 2W),
    6th Generation 14 nanometer i7 6700K / i5 6600K (TDP 91W / Idle 2W):
    Tj Max (Throttle temperature) = 100°C

    5th Generation 14 nanometer i7 5775C / i5 5675C (TDP 65W / Idle 2W):
    Tj Max (Throttle temperature) = 96°C

    4th Generation 22 nanometer i7 4790K / i5 4690K (TDP 88W / Idle 2W),
    4th Generation 22 nanometer i7 4770K / i5 4670K (TDP 84W / Idle 2W):
    Tj Max (Throttle temperature) = 100°C "
     

    Post edited by ebergerly on
  • nonesuch00nonesuch00 Posts: 18,032
    edited November 2017

    My i7-3630qm render is at 76%, 1531 iterations, after 20 hours 25 minutes 42 seconds that I will be posting later in comparision to the 60 hour 2000 iterations render I posted earlier. Had I done the render on my old i5-2520m CPU laptop it likely would have exceeded 72 hours and maybe on some renders 96 hours. The reason though I'm making this post is to give the idle temps of the two CPUs I've used in this HP 8470P Elitebook laptop.

    i5-3230m - 32C was the coolest always when I was away from the laptop I saw but typical non-rendering temperatures are in the 40C - 55C range. The TDP is 35W.

    i7-3630qm - 34C was the coolest also always when I was away from the laptop. Typical non-rendering temperatures are in the 55C - 65C range. The TDP is 45W.

    Post edited by nonesuch00 on
  • nonesuch00nonesuch00 Posts: 18,032
    edited November 2017

    OK, my test render with the roo boxing the man is done now using the i7-3630qm. It took 26 hours 34 minutes total to do 2000 iterations. That is less than 1/2 the time it took the i5-3230m to render essentially the same image although the 1st image was slightly simpler because it didn't use SimonWM's Wet iRay Materials Bundle for G3M & G2M and JGreenless's Shimmer iRay Shaders.

    Also here is a screen shot of my Core Temp readings...two of my cores actually topped at at 105C and 100% utilization while core 0# typically was at about 55% utilization and 85C - 95C because I always uncheck core 0" thread #0 thread in affinity setting on the Detail Tab in Task Manager when starting DAZ Studio so my PC doesn't become i/o bound.

    Hmmm, I guess I shouldn't have my CPUID on that image but I think they are no longer unique per CPU.

    Core Temp Readings after about 23 hours render time.png
    683 x 371 - 36K
    Post edited by nonesuch00 on
  • ebergerlyebergerly Posts: 3,255

    nonesuch thanks for the data. Though I'm really confused. Sounds like you bought a used, "medical waste" laptop (not sure what that is, and whether it may have been damaged, etc.), then bought a used medical waste CPU, and then you installed the used CPU in the used laptop? And you also think the CPU may have been tampered with? 

    And the resulting max rendering temps are at the maximum TjMax levels? 

    Honestly I'm not sure that it's telling us much. Seems like the hardware you're testing isn't really representative of a standard, off-the-shelf laptop. Also, I didn't see render time per frame data, which would indicate if performance throttling occurred. Seems to me there might be some other issues with that system that are causing high temperatures?    

  • ebergerlyebergerly Posts: 3,255

    Also, IMO the really telling thing is your idle temperatures. Based on the info I posted from Tom's Hardware, they should be about 10C above ambient, but looks like your max idle temps are 55-65C, which implies your ambient is 45-55C, or 120-130F smiley  I hope not.

    What that tells me is that even with little or no load on your CPU/GPU the sytem is running too hot. Again, it makes me think there's a problem with your cooling system or other components of the laptop are damaged. 

  • If this thread turns into a continuation of the war in a previous thread it will be locked. Please keep the tone civil and address comments to topic not poster.

Sign In or Register to comment.