System Total Freeze (NEED HELP)

[WesPlaysGuitar]

So I've just built my first PC ever. Pretty excited about it. It actually posts and runs smoothly at first glance. I start to download all of my games and what not onto it overnight and go to bed. So the next day I go to play League of Legends, a relatively low demand game and use OBS to record the gameplay just as a test. About ten to fifteen minutes in, the computer freezes completely, requiring a hard power off and reboot. This phenomenon continues during gaming, whether it be League, BF4, 7 Days, anything requiring moderate load. However, just browsing, it's never frozen. I've updated drivers, everything I can think of to no avail. Please help.

Specs:
AMD FX 8350 8 Core @ 4.0 GHZ
ASUS M5A99FX Pro R2.0 Mobo
MSI Radeon R9 290 Gaming Edition 4GB
Samsung 840 Evo 250 GB SSD
Corsair H60 Liquid Cooling
WD Caviar Blue 1TB HDD
Corsair CX600M PSU
Windows 8.1 Pro 64 Bit

 

[Darkbreeze]

Well, since the GPU really only draws power through the board until a demanding load is presented, it seems likely that it's either the PSU or the card itself. It could be a driver issue I suppose. Are you running the Omega Catalyst drivers that are the most currently available or the drivers that came with the card?

I'd download and run HWinfo, run sensors and post screenshots of the system voltages. All sensors wouldn't hurt actually.

http://www.tomshardware.com/answers/id-2173703/post-images-tomshardware-guide.html

 

[WesPlaysGuitar]

Well, since the GPU really only draws power through the board until a demanding load is presented, it seems likely that it's either the PSU or the card itself. It could be a driver issue I suppose. Are you running the Omega Catalyst drivers that are the most currently available or the drivers that came with the card?

I'd download and run HWinfo, run sensors and post screenshots of the system voltages. All sensors wouldn't hurt actually.

http://www.tomshardware.com/answers/id-2173703/post-images-tomshardware-guide.html

Is this what you're looking for?

 

[Darkbreeze]

No, not HWmonitor, HWinfo. Totally different utilities. HWmonitor is NOTORIOUSLY inaccurate reading sensor data. Honestly, I don't even know why they still offer it. CPU-Z works great, HWmonitor, not so much.

HWinfo: http://www.hwinfo.com/download.php


This is what the sensors window looks like. You'll need to scroll down and take a second screenshot to capture all the sensors.

 

[WesPlaysGuitar]

No, not HWmonitor, HWinfo. Totally different utilities. HWmonitor is NOTORIOUSLY inaccurate reading sensor data. Honestly, I don't even know why they still offer it. CPU-Z works great, HWmonitor, not so much.

HWinfo: http://www.hwinfo.com/download.php


This is what the sensors window looks like. You'll need to scroll down and take a second screenshot to capture all the sensors.

I can't quite give you the screenshot now as I've departed for college and left that behemoth of a computer at home. However, something very interesting occurred since your last reply. Upon starting to resort to the RMA process, I tried tinkering with Catalyst Control Center on the OC settings. I dropped the power limit and set the memory frequency (I think although I'm not sure) to around 700 MHz. After doing this, I ran the Furmark test for ten minutes with no problems whatsoever (I stopped it because I was cramped for time) and then proceeded to play a full length game of BF4 on High to Ultra settings. I'm not sure what part of whatever I did worked but as far as I can tell the problem has ceased or at least drastically subsided. As I said though, I ran the Furmark stress test for a good ten minutes, maxed out the load and temp at around 80C, and not even a hiccup out of it. It would seem I'm on to something although if you could tell me what could be up with this I'd greatly appreciate it.

 

[Darkbreeze]

Well, when you get back I'd be happy to look into this further with you. Not much point if you don't have the computer to apply potential solutions to.

 

[Rukeith]

I had this exact problem with the 290, i eventually ended up refunding it for something else because I couldn't fix it... Freezes that require reboot in Youtube, Games or whenever the card wanted to. There were a few "fixes" I found that are suppose to help the card. In Afterburner, go to Settings > General TAB > Enable both 'Force constant voltage' and 'Disable ULPS', then restart. You could also try under-clocking the card's Core Clock and see if that helps...

 

[WesPlaysGuitar]

Well, when you get back I'd be happy to look into this further with you. Not much point if you don't have the computer to apply potential solutions to.

I'll be sure to send you screenshots when I can.

 

[WesPlaysGuitar]

I had this exact problem with the 290, i eventually ended up refunding it for something else because I couldn't fix it... Freezes that require reboot in Youtube, Games or whenever the card wanted to. There were a few "fixes" I found that are suppose to help the card. In Afterburner, go to Settings > General TAB > Enable both 'Force constant voltage' and 'Disable ULPS', then restart. You could also try under-clocking the card's Core Clock and see if that helps...

I can't refund it sadly. I was going to RMA for a replacement but since then I've found that dialing back the power limit and toning down the memory frequency has fixed the problem. I'll be able to try this more thoroughly over the weekend and find the "stable point" but as it stands I'm still pretty happy with the solution as I didn't sacrifice a ton of performance by any means. I ran the Furmark stress test for over ten minutes averaging around 100 fps. Could still get better but I had to move in at school today so I haven't had much a chance to mess around with it. I'd like to raise the memory frequency if I can though. 700 MHz is on the lower end of the slider.

 

[Darkbreeze]

Here's my thought on the issue. Scaling back the settings having a result of the problem resolving itself leads me to believe that it might well be a power issue just as I first said. If the PSU isn't supplying stable and sufficient power with low levels of noise and ripple, or if the voltage regulation is all over the place like it generally is with low cost PSUs, the card will never run right at it's normal settings because the PSU isn't providing what the card needs. Downclocking the card smooths things out in these situations because now the card isn't placing the same demand on the PSU.

I wouldn't rule out the card either though, but I'd really want to see some GPU and PSU sensor readings before I even started pointing fingers at either of them. The CX600M isn't a terrible unit by any means, but it's a Tier 3 unit and really wasn't meant to be used with high end cards or gaming rigs that load the CPU up as well. That FX chip has a 125w TDP, which together with that cards requirements is fine normally, but when it's turbo core kicks in under load the system raises the voltage substantially to account for the increased multiplier and with the system controlled method it generally increases the voltage much more than you would if you overclocked it manually to ensure stability.

This increased draw along with the draw from a fairly high end card is likely to be more than that unit was meant to deal with. Again, seeing what's actually going on according to the sensors is the best place to begin making determinations. Having to turn DOWN the settings on the card to get it to work right is unacceptable. We want to be able to turn them UP, not down, so let's see what we see when we can.

Might even have just been unlucky and gotten a bad apple on that PSU, or on the GPU. Hard to try and make guesses.

 

[WesPlaysGuitar]

Here's my thought on the issue. Scaling back the settings having a result of the problem resolving itself leads me to believe that it might well be a power issue just as I first said. If the PSU isn't supplying stable and sufficient power with low levels of noise and ripple, or if the voltage regulation is all over the place like it generally is with low cost PSUs, the card will never run right at it's normal settings because the PSU isn't providing what the card needs. Downclocking the card smooths things out in these situations because now the card isn't placing the same demand on the PSU.

I wouldn't rule out the card either though, but I'd really want to see some GPU and PSU sensor readings before I even started pointing fingers at either of them. The CX600M isn't a terrible unit by any means, but it's a Tier 3 unit and really wasn't meant to be used with high end cards or gaming rigs that load the CPU up as well. That FX chip has a 125w TDP, which together with that cards requirements is fine normally, but when it's turbo core kicks in under load the system raises the voltage substantially to account for the increased multiplier and with the system controlled method it generally increases the voltage much more than you would if you overclocked it manually to ensure stability.

This increased draw along with the draw from a fairly high end card is likely to be more than that unit was meant to deal with. Again, seeing what's actually going on according to the sensors is the best place to begin making determinations. Having to turn DOWN the settings on the card to get it to work right is unacceptable. We want to be able to turn them UP, not down, so let's see what we see when we can.

Might even have just been unlucky and gotten a bad apple on that PSU, or on the GPU. Hard to try and make guesses.

Like I said, I'll be able to provide some better information on Friday as that's when I'll be home. Until then, I'm going to assume it's like you said, either the PSU not being able to meet the demands or a partially faulty video card. Either way, it's definitely fixable.

 

[Darkbreeze]

Sounds good bro. Until then.

 

[WesPlaysGuitar]

Sounds good bro. Until then.

Well I'm back at home and I've downloaded HWiNFO. What screenshot are you looking for?

 

[Darkbreeze]

Just like I explained up above, run the program, check only the box that says "sensors" and take screenshots of all the sensors. You'll need to take multiple shots to get all the sensor data.

 

[WesPlaysGuitar]

Just like I explained up above, run the program, check only the box that says "sensors" and take screenshots of all the sensors. You'll need to take multiple shots to get all the sensor data.

 

[Darkbreeze]

Perfect, now can you do the same thing with the system under load. Preferably something like running a demanding title, furmark or heaven with a fairly high demand, or just run something that normally causes the issues and get those same screenshots so we can see if anything has changed on the sensor readings.

 

[WesPlaysGuitar]

Perfect, now can you do the same thing with the system under load. Preferably something like running a demanding title, furmark or heaven with a fairly high demand, or just run something that normally causes the issues and get those same screenshots so we can see if anything has changed on the sensor readings.

 

[Darkbreeze]

Sensors look ok. Let's look around a bit. Why don't you run Memtest to see if there is perhaps a faulty module. Run it for seven passes on each module individually. Running on multiple modules often creates false test results. Also, run Seatools for windows, doesn't matter about brand, on all connected drives.

Memtest:http://www.memtest86.com/download.htm

Seatools: http://www.seagate.com/support/downloads/item/seatools-win-master/

 

[WesPlaysGuitar]

Sensors look ok. Let's look around a bit. Why don't you run Memtest to see if there is perhaps a faulty module. Run it for seven passes on each module individually. Running on multiple modules often creates false test results. Also, run Seatools for windows, doesn't matter about brand, on all connected drives.

Memtest:http://www.memtest86.com/download.htm

Seatools: http://www.seagate.com/support/downloads/item/seatools-win-master/

I already ran Memtest and it showed clean. I swapped out my surge protector and moved a small heater to another outlet and haven't frozen yet. Just logged 2 straight hours on BF4. May have been overloading surge protector/outlet. Will keep posted.

 

[Darkbreeze]

You were using a power strip? If possible, don't ever use a power strip to plug the PSU into. They are notoriously prone to being faulty, as all but about two percent of power strips are extremely cheaply made regardless what you might have paid for them. If you must have one, do some investigation into a very good, well reviewed power strip/conditioner such as those used with mid to high end studio equiptment. When possible, simply plug the monitor and tower directly into the wall socket. This is the best method of providing clean power with no intermediary devices to create additional resistance or noise on the circuit.

 

[WesPlaysGuitar]

You were using a power strip? If possible, don't ever use a power strip to plug the PSU into. They are notoriously prone to being faulty, as all but about two percent of power strips are extremely cheaply made regardless what you might have paid for them. If you must have one, do some investigation into a very good, well reviewed power strip/conditioner such as those used with mid to high end studio equiptment. When possible, simply plug the monitor and tower directly into the wall socket. This is the best method of providing clean power with no intermediary devices to create additional resistance or noise on the circuit.

I didn't even think of it. Kinda just habit as it was already there. Lol but I haven't had any problems since though so I believe it was a wiring issue instead of a hardware issue, especially if my sensors looked normal. If anything, maybe a bigger PSU would solve my problems although I'm running fine right now.

 

[Darkbreeze]

I WOULD replace that CX unit anyhow. Here's my standard statement when it comes to that series.

You need to determine the capacity of PSU required as determined by what GPU, if any, is installed (IF no GPU and using onboard graphics, a 450w unit is fine and can be a Tier 3 unit. Higher tiering is always better though.) and then find a Tier 2B or higher unit from the tier list.


PSU capacity requirements: http://www.realhardtechx.com/index_archivos/Page362.htm

PSU Tier list: http://www.tomshardware.com/forum/id-1804779/power-supp...


Don't make the mistake of thinking name brands with good reputations always make and offer good products, especially when it comes to power supplies. Corsair, Cooler Master and Thermaltake, among others, are well known for having crappy power supplies when it comes to their budget models. Typically, most units from Antec, Seasonic, XFX, EVGA and Superflower are good but you should verify any potential model purchase is verified by either matching it's model number to a ranked unit on the tier list or finding and reading a positive review of the EXACT model on a reputable review site like HardtechX or JonnyGuru, to name a couple.

Don't use any calculators. They are not accurate. Use the GPU recommendations on the page I listed above. It includes overhead for the CPU and other components, so long as no overclocking is involved. If you're overclocking the CPU or GPU, add another 50-100w depending on if you're overclocking one or both, and how much. Clearly a mild or medium overclock isn't going to increase the demand and load like a high or extreme overclock is so just use good judgement. In most cases, if you're overclocking one component I recommend an additional 50-100w. If you're overclocking two components, increase the capacity by 100-200w. If the overclocking will be extreme, an extra 200w minimum is probably a good idea. These are basically guesswork as every overclock and the derived load will be different for every system, but it's fairly close.

 

[WesPlaysGuitar]

I WOULD replace that CX unit anyhow. Here's my standard statement when it comes to that series.

You need to determine the capacity of PSU required as determined by what GPU, if any, is installed (IF no GPU and using onboard graphics, a 450w unit is fine and can be a Tier 3 unit. Higher tiering is always better though.) and then find a Tier 2B or higher unit from the tier list.


PSU capacity requirements: http://www.realhardtechx.com/index_archivos/Page362.htm

PSU Tier list: http://www.tomshardware.com/forum/id-1804779/power-supp...


Don't make the mistake of thinking name brands with good reputations always make and offer good products, especially when it comes to power supplies. Corsair, Cooler Master and Thermaltake, among others, are well known for having crappy power supplies when it comes to their budget models. Typically, most units from Antec, Seasonic, XFX, EVGA and Superflower are good but you should verify any potential model purchase is verified by either matching it's model number to a ranked unit on the tier list or finding and reading a positive review of the EXACT model on a reputable review site like HardtechX or JonnyGuru, to name a couple.

Don't use any calculators. They are not accurate. Use the GPU recommendations on the page I listed above. It includes overhead for the CPU and other components, so long as no overclocking is involved. If you're overclocking the CPU or GPU, add another 50-100w depending on if you're overclocking one or both, and how much. Clearly a mild or medium overclock isn't going to increase the demand and load like a high or extreme overclock is so just use good judgement. In most cases, if you're overclocking one component I recommend an additional 50-100w. If you're overclocking two components, increase the capacity by 100-200w. If the overclocking will be extreme, an extra 200w minimum is probably a good idea. These are basically guesswork as every overclock and the derived load will be different for every system, but it's fairly close.

Well I just had a random freeze in BF4 so I'm going to go ahead and assume it's the PSU. My computer was updating at the time and the only time I've ever frozen is under overall system load. I'll be getting some financial aid money soon so I'll upgrade to a 1000W psu probably within the next few weeks. Is there any recommendation you have?

 

[Darkbreeze]

Is anything overclocked, including the CPU? Try going into Catalyst control center and disabling Overdrive for both the CPU and GPU. No graphics or processor overdrive enabled. See if it still does it. You really don't need a 1000w PSU unless you plan to add a second GPU at some point. A good 650-750w unit would be fine. I'd highly recommend this:

PCPartPicker part list / Price breakdown by merchant

Power Supply: EVGA 750W 80+ Gold Certified Fully-Modular ATX Power Supply ($108.99 @ SuperBiiz)
Total: $108.99
Prices include shipping, taxes, and discounts when available
Generated by PCPartPicker 2015-01-10 18:45 EST-0500


OR this:

PCPartPicker part list / Price breakdown by merchant

Power Supply: EVGA 750W 80+ Bronze Certified Semi-Modular ATX Power Supply ($74.98 @ OutletPC)
Total: $74.98
Prices include shipping, taxes, and discounts when available
Generated by PCPartPicker 2015-01-10 18:45 EST-0500

 

[Rukeith]

Well I hope throwing money at a new PSU will resolve the problem... Literally tonnes of people have this problem and most are resolved by Tweaks, VBIOS Update or RMA. Did you try what I said before in MSI afterburner? Settings > General TAB > Enable both 'Force constant voltage' and 'Disable ULPS', then restart. BTW when watching YouTube videos, do you get random freezes then too? Hardware Acceleration has a hard time working well with that card. Well anyway, don't get a 1000w honestly lol you don't need that much. A high quality 750w with enough amps on the 12v rail will be more than enough for any single card.