37 replies to this topic

[BTGbullseye]

Horseman, on 14 January 2018 - 10:54 PM, said:

59 under load? That's fairly cool, anything under 80 is a non-concern.

83°c before it starts throttling, 93°c before it shuts itself off. Neither temp will actually damage the card, as the components are rated to over 100°c.

Horseman, on 14 January 2018 - 10:54 PM, said:

59 idle? You have a problem.

Not necessarily... Idle temps are a useless metric, unless you're looking to try and run it as quiet as possible.

[Not A Real RAbbi]

It was 59C under a load. As high as 67 today after a few straight hours of gaming. Sounds like it's not a problem.

The CPU thermal paste looked good. Fully covered without spilling over the sides. Of course, since I pulled the heat sink
anyhow I went ahead and replaced the paste. Still doing fine.

Started getting the stutters today after a few hours of Destiny 2; took 15 minutes off, turned settings all down to medium or high. Been pretty solid since then. Video card memory usage was at like 79% with the previous settings (of 3GB), and none of the CPU cores was getting pegged.

***EDIT***

So, finally did it again just a bit ago (I'm on a different machine now). FPS cut to 27-37, and felt worse than that. Nearly unplayable. If you're familiar with Destiny 2, it was just at the end of a public event (the one with the Fallen servitor) in EDZ, and there was a lot going on. 3-4 other players there, one of 'em had dropped Nightstalker super and grenade. Servitor blowing up. Etc. Got the single-digit framerate stuff for a few seconds, followed by being stuck at 30-ish.

So that Open Hardware Monitor was a bust. Installed a different one (mentioned here in a previous comment; thanks for the recommendation!). But by then, temps were pretty normal. CPU temp was falling from 35.0 by the time I got it installed and running. It DID record a total CPU usage of 92.0% and 100.0% on one of the cores, but not really sure what that was about.

SO, gonna let it sit and cool off overnight. If it's still being weird in the AM, after 7-10 hours of being shut down at room temp, then I'm gonna be like 99% sure I've got a software/system/driver problem. If it's back to normal, then I'm gonna run it with the new monitor running in the background and see what it does.

It had been running for 9-10 hours straight, with breaks from gaming lasting 7-30 minutes, when this started. When I had the case open, I was kinda shocked at how little space there is between the GPU heat sink and the bottom of the power supply. The graphics card itself seems to mostly block air from coming from anywhere other than the back of the card. This just bums me out though...

Edited by Not A Real RAbbi, 15 January 2018 - 07:54 PM.

[Horseman]

if your CPU is maxing out like that, that's the most likely culprit right there. Check the task manager - are any other programs hogging CPU time, or is the game responsible for most of it?

[BTGbullseye]

Sounds mostly like a background program eating CPU, though it is possible it's a bad RAM module, BIOS glitch, (if the BIOS is setting the timings wrong on your RAM) or Windows is using a power saving feature. (always change it to the "high performance" mode for gaming)

If you're on Windows 10, it is entirely possible that the telemetry reporting system that you can't ever turn off is the culprit. (I've seen it numerous times, it eats CPU, RAM, drive access IO, and internet bandwidth, not to mention that it runs with the highest process priority)

Edited by BTGbullseye, 16 January 2018 - 04:01 AM.

[Not A Real RAbbi]

After a nice long night of being shut off, it's still there. Framerates are lower than usual. Also, LOAD TIMES seem a lot longer. When running Destiny 2 (for instance; it's the only game I've run since the last wipe), it's using about 2.5 GB of RAM (sounds about right), and the next thing down the list (Blizzard app) is using like 150-ish MB. Nothing else even comes close. GPUID HW monitor says nothing of concern regarding temperatures or usage. Everything on the list on Task Manager looks right. I've uninstalled every Windows 10 P.O.S. app it'll let me, the HP junk that came on the computer (I HATE their stuff), etc.

So, about that RAM/BIOS glitch... (Faction Rally starts today... )

Quote

Not necessarily... Idle temps are a useless metric, unless you're looking to try and run it as quiet as possible.

Well, 59.0 C was picked up as MAX after shutting down the program, but it was settling down by about 1 degree C per 1.5-2.0 seconds. Cooled off pretty quickly.

Edited by Not A Real RAbbi, 16 January 2018 - 06:33 AM.

[BTGbullseye]

RAM usage is totally unimportant... (it doesn't affect anything unless you're exceeding your total RAM amount, which will then force Windows to star using virtual RAM, which is an exceedingly rare occurence on machines less than 20 years old) You're looking for CPU priority, and CPU usage.

I highly recommend grabbing Process Hacker. It's a free open source Task Manager replacement that does way more than the one built into Windows. If you're comfortable with it, grab a screenshot of your running processes, and we can go from there.

There might also be a few services running that are interfering, they would take a lot longer to troubleshoot though. (mainly any of the remote access ones will bog your system for no apparent reason)
______________________________________________________________________________________

If it's failing RAM, it'll affect everything on your system, and you'll need to have a replacement stick ready to use. The only reliable way to test it is to have a known good stick there to swap out with a suspected bad stick, and compare.

If it's a BIOS error, the only one that can fix it is the motherboard manufacturer. Check for firmware updates on the manufacturer site, and download them only from there. If it isn't fixed by that, it might need manual timings, which requires a lot more specific knowledge of the hardware than you've provided. (RAM manufacturer, RAM speed, factory rating for the timing of the RAM, motherboard manufacturer, motherboard product designation, BIOS version, etc.)

There are other hardware-based possibilities, but I suggest we focus on the software side first, as it's a lot less expensive to fix.

Edited by BTGbullseye, 17 January 2018 - 12:04 AM.

[Tarl Cabot]

Quote

prebuilt bought on sale on Amazon. HP Pavilion.
Intel Core i7-7700
Nvidia GTX 1060 (3GB)
16 GB DDR4/2400

HP Pavilion model 570-p047c or 570-p030?

Lets back up some. What model and was it new or used/reconditioned? When the side case is off, what is the PSU? And did it come with the Nvidia GTX 1060 or was that added later?

Something like this ?

https://www.amazon.c...0/dp/B06XDSMXR1

Also, in the Nvidia UI, change the Power Management mode from adaptive to Max Performance. This ensures the Nvidia card is being used and not the Intel GPU on the CPU.

Edit. The models I listed came with a 180w PSU... I do hope yours has a better replacement PSU.

Edited by Tarl Cabot, 18 January 2018 - 09:32 PM.

[Tarl Cabot]

Just to edit the above post some. If your system is that pre-manufacturer with the 1060 added on with that underpowered PSU, the system may be dropping the GPU settings to a low setting, running it so it is not much better than the built-in Intel GPU on the CPU.

https://www.amazon.c..._sp_detail_page

I think I found what he is using. It may be driver related or heat related.

Edited by Tarl Cabot, 21 January 2018 - 01:57 PM.

[Not A Real RAbbi]

THIS one. This is where I bought it. This is the one I bought.

HERE is a graph from a process monitor. Older is to the left, newest data at the right edge. You see to the left, near the edge, where the CPU usage went from reasonable to HOLY-MOTHER-OF-TRICERATOPS. That corresponded to the sudden change (during Destiny 2 session) from a smooth 60 FPS to 22-27 FPS and sluggish control response. The big drop, of course, was me shutting down the game. Spike toward the end is firing up Firefox. Now, if I go back to that computer (I'm using a different one ATM) to fire up ANY video game, I'll get the same crap frame rates and load times, and it doesn't matter what time of day it is or how long it's been shut down. Only thing that fixes it is a system wipe, as far as I can tell.

This time around I did NOT install Steam or any other game. Also didn't uninstall anything. Only stuff on there that's not as-delivered is Razer whateverit'scalled stuff (wasn't on there the first two times this happened), and Blizzard.net app with Destiny 2. All installed on the SSD, with plenty of room to spare. OH, and Firefox of course. On previous occasions, It's had Steam with Borderlands 1/2/TPS, DooM, WildStar, Paladins, Gigantic, MWO, and probably something else. Once installed the old manager software for that ancient Nostromo N52, but haven't put it on there since the first time I had to wipe.

I'm about to set fire to HP. The whole company. Employees, their families and friends, pets, homeless folks sleeping behind their facilities, etc. LOTS of fire. Help me not be a criminal, please?

[Horseman]

Disable Nvidia's overlay along with its' recording functions. That might be part of the problem.

Generally try to get rid of any extra software that runs on startup that you're not certain you'll need.

Edited by Horseman, 28 January 2018 - 12:58 AM.

[Not A Real RAbbi]

Horseman, on 28 January 2018 - 12:56 AM, said:

Disable Nvidia's overlay along with its' recording functions. That might be part of the problem.

Generally try to get rid of any extra software that runs on startup that you're not certain you'll need.

Tried that one run. No effect. Still did what it did. And doesn't explain the nature of the problem: It starts at some point 2-8 days after fresh installation of Windows, and persists thereafter. The Nvidia stuff being present or absent seemingly has no effect. Same for virtually anything else installed on the system.

Currently installing all the HP-recommended updates, including some BIOS update that was hiding in there. We'll see what that does for us. A promising note: Razer Synapse loaded up super quick on restart after the BIOS update, as did the OneDrive deal (which I intend to uninstall later, as I have no use for it really). Once the rest of these updates are done, we'll restart and see where we stand.

Also posted about this issue on my Facebook page (got some friends who are pretty into this stuff, including a game dev working on WoT) and on a computer-issues subreddit (just now; waiting to see what shakes out there). Windows updates, and Nvidia driver updates, have so far failed to have any measurable impact.

[Tarl Cabot]

What program are you utilizing to monitor CPU and GPU heat? - The Task manager is showing the drop in CPU usage but it does not provide any additional system information such as CPU/GPU heat, CPU/GPU speed (throttling down due to heat), GPU usage, etc.

After reading some of the reviews, I believe that system has like one case fan. Reviewer in the thread below - first suggestion is to remove the side panel. Test this and then even have a fan blowing through/over it. Another option would be to cut/install one or two case fans.

Stress tests. Notes of the GPU throttling down due to heat. You noted you did have the side case off but did it stay off? And just FYI, it does not take hours for CPU/GPU to cool down, tis not an engine..

https://slickdeals.n...9#post109075599

System power Options, set to High Performance. Nvidia UI, 3D, Power Management mode, set to Max Performance.

HWMonitor - https://www.cpuid.co.../hwmonitor.html
Open Hardware Monitor - http://openhardwaremonitor.org/
HWInfo64 - https://www.hwinfo.com/download.php

Edited by Tarl Cabot, 28 January 2018 - 01:10 PM.

[Not A Real RAbbi]

Tarl Cabot, on 28 January 2018 - 12:46 PM, said:

What program are you utilizing to monitor CPU and GPU heat? - The Task manager is showing the drop in CPU usage but it does not provide any additional system information such as CPU/GPU heat, CPU/GPU speed (throttling down due to heat), GPU usage, etc.

After reading some of the reviews, I believe that system has like one case fan. Reviewer in the thread below - first suggestion is to remove the side panel. Test this and then even have a fan blowing through/over it. Another option would be to cut/install one or two case fans.

Stress tests. Notes of the GPU throttling down due to heat. You noted you did have the side case off but did it stay off? And just FYI, it does not take hours for CPU/GPU to cool down, tis not an engine..

https://slickdeals.n...9#post109075599

System power Options, set to High Performance. Nvidia UI, 3D, Power Management mode, set to Max Performance.

HWMonitor - https://www.cpuid.co.../hwmonitor.html
Open Hardware Monitor - http://openhardwaremonitor.org/
HWInfo64 - https://www.hwinfo.com/download.php

Used open HW monitor first time around. NOT a heat issue. Don't remember which one I used after that, but openHW didn't show GPU temp. The latter one did. Neither ever approaches a concerning temp.

Got a HP guy remoted in right now scratching his head on this one. Slippery little b****rd tried to skeeze out on me after he merely changed a setting concerning page files. Kept him on while I fired up D2. NOPE. Still suckin' wind.

There's an error in the system or bios somewhere.

[Tarl Cabot]

Understand, you are providing only bits and pieces, even though several other items have been discussed.

There is heat for the components that can be read, provided the heat is being read correctly, then there are the other components. Have you kept the side case/panel off? Again, simply to remove most of that from the equation...

But heat was not the only thing I was asking about. When it goes south, are either/both CPU/GPU running at their respective speeds or do the monitoring software indicating reduced speeds?

Also, when it does go south, is the harddrive light spasming?

Again, mentioning this - Computer Power Options, set to High Performance (prevents throttling unless heat related). Nvidia UI, 3D, Power Management mode, set to Max Performance (ensures using Nvidia GPU and not the CPU's Intel GPU).


The last item is the system, ie motherboard may be defective, ie there may be a defective component on it since if it was bios there would be tons more issues being reported online, especially tons of sales on Black Friday and it would not wait for several hours to appear.

Edited by Tarl Cabot, 28 January 2018 - 03:32 PM.

[Not A Real RAbbi]

Potential fix applied.

Used Park Control and disabled CPU core parking. Test runs with Borderlands 2 and Doom2016 seem to verify the fix. Will test with MWO and Destiny 2 in the next 18 hours. More to follow, but it appears there's a not-uncommon glitch with Power Saver and newer Intel Core ™ processors that winds up constantly trying to park CPU cores, even when something's trying to
use them. Something like that.

If this isn't a stunning success, I'll follow every last letter of Tarl Cabot's advice and see where that lands us.

[Tarl Cabot]

You are using Power Saver as your power settings? wtf, seriously? Of course it will aggressively core park and throttle CPU, that is its purpose.. that is what it is meant to do. But your computer is not a laptop constantly running off of batteries where you are trying to make that battery charge last between charges.. It is a gaming system, you do not put chains on it to keep it down...

When the Ryzen chips came out, that is one thing AMD said needed to be done, to set Power settings to High Performance due to Windows stuff. It was a few months before AMD added their own power settings that would allow a player to run Balanced power without adversely affecting performance.

And have you really calculated the money being saved vs the frustration you are putting yourself through, the time loss, etc, etc, etc. I mean.. a few websites noting how to saving money? Ditch the desktop and get a laptop and save approx $11/year on electricity.....

If anything, at least change to to Balance, and if it is not already setup then put a timer to turn off your monitor if you are really concerned about saving little bit of energy...

But if you are really, really concerned, a shortcut to your desktop for power options then when playing games select High Performance. Keeping it on Power Saver while messing with other Parking/etc will have problems crop you that will take you awhile to figure out due to all the other settings that are changed when moving from Power Saver, Balanced and High Performance.

Oh well, look like my job is done here...

Edited by Tarl Cabot, 30 January 2018 - 08:22 PM.

[Not A Real RAbbi]

Tarl Cabot, on 30 January 2018 - 08:00 PM, said:

You are using Power Saver as your power settings? wtf, seriously? Of course it will aggressively core park and throttle CPU, that is its purpose.. that is what it is meant to do. But your computer is not a laptop constantly running off of batteries where you are trying to make that battery charge last between charges.. It is a gaming system, you do not put chains on it to keep it down...

When the Ryzen chips came out, that is one thing AMD said needed to be done, to set Power settings to High Performance due to Windows stuff. It was a few months before AMD added their own power settings that would allow a player to run Balanced power without adversely affecting performance.

And have you really calculated the money being saved vs the frustration you are putting yourself through, the time loss, etc, etc, etc. I mean.. a few websites noting how to saving money? Ditch the desktop and get a laptop and save approx $11/year on electricity.....

If anything, at least change to to Balance, and if it is not already setup then put a timer to turn off your monitor if you are really concerned about saving little bit of energy...

But if you are really, really concerned, a shortcut to your desktop for power options then when playing games select High Performance. Keeping it on Power Saver while messing with other Parking/etc will have problems crop you that will take you awhile to figure out due to all the other settings that are changed when moving from Power Saver, Balanced and High Performance.

Oh well, look like my job is done here...

Didn't intentionally set Power Saver. It was a default setting. Interestingly, I didn't have this problem with my last PC, which ran a FX8350 CPU and Windows 10 64-bit. Anyhoo, problem appears to be solved for real. D2 is back to its original beautiful self. Time to re-re-reinstall MWO and get to work!

Thanks everyone for the advice here! I really do appreciate it!

[BTGbullseye]

Tarl Cabot, on 30 January 2018 - 08:00 PM, said:

You are using Power Saver as your power settings? wtf, seriously? Of course it will aggressively core park and throttle CPU, that is its purpose.. that is what it is meant to do. But your computer is not a laptop constantly running off of batteries where you are trying to make that battery charge last between charges.. It is a gaming system, you do not put chains on it to keep it down...

I did actually mention this fairly early on...

BTGbullseye, on 16 January 2018 - 04:01 AM, said:

or Windows is using a power saving feature. (always change it to the "high performance" mode for gaming)

Not A Real RAbbi, on 30 January 2018 - 08:35 PM, said:

Didn't intentionally set Power Saver. It was a default setting.

That's always the default for a fresh install of Windows ever since Vista. Prior to that, the default was "balanced". Any other setting that you see by "default" was actually changed by something aftermarket.

Here's a message everyone should get:
ALWAYS SET YOUR POWER SAVER SETTING TO "HIGH PERFORMANCE" FOR ANY DESKTOP, AND MAKE SURE IT'S SET FOR THAT ANY TIME YOU GAME ON ANY SYSTEM!!!

Edited by BTGbullseye, 31 January 2018 - 12:41 AM.