|
Before I start ripping everything out of this machine one piece at a time, I figured I'd run it past some other geeks, to see if there are any suggestions. Problem description: Video stutters and becomes choppy; desktop effects, like window openings and closings, do the same. Attempted fixes: Replaced graphics card, and reinstalled Windows Recent changes: This is a brand new rig I just built from scratch. -- Operating system: Windows 7 Ultimate System specs: ASUS ROG Maximus VIII Ranger motherboard Intel i7 6700K CPU Thermaltake 750 watt 80+ bronze power supply 2x Samsung EVO 850 500 gig drives Two Western Digital platter drives 16 gig Corsair Vengeance LPX 2x8GB (CMK16GX4M2B3000C15) EVGA GTX 960 SSC Location: USA I have Googled and read the FAQ: Yes My last computer took a poo poo, and I just built this one last week. After firing it up, I noticed terrible desktop graphics performance, and Windows kept opening the "Windows has detected performance is slow" prompt, asking me if I wanted to disable Aero. A Windows Experience refresh looked OK, showing a 7.9 for graphics, but performance was definitely poo poo. I may not have done an Experience refresh at the height of the problem. The graphics card I initially bought was a Zotac GTX 960. I was skeptical about Zotac to start with, so that's the first thing I replaced. I am no longer getting the Windows warning, but performance is still degraded. It seems OK when I first boot up, but it gets worse over time. Additional information: Once the problem shows up, it stays until I reboot. I don't see anything unusual going on in the Task Manager. The guy at the store accidentally brought out Corsair 2x4s, and I did not notice until I had installed them. When I first replaced them with the 2x8s, the machine would not boot. It booted when I moved them into the second pair of slots, and now they are back in the first pair of slots, and the machine boots fine. There may be some connection to hard drive activity. I did not notice this before replacing the graphics card, but the problem in question definitely showed up a few minutes ago while I was deleting a ton of information from one of the Samsung EVOs. The problem was more severe than I would expect just from disk activity, and the video I was playing was not on the same drive I was deleting files from. Also, the problem remained after the deletion was finished. I have Samsung Magician installed, and it was installed on the last machine, too. Crystal Mark doesn't show any significant difference between having Samsung Magician and not having Samsung Magician. I haven't run a Crystal Mark test during an incident of video choppiness, yet. My Windows Experience Index showed 6.0 for Aero desktop performance, and for 3D performance, after I replaced the Zotac with the EVGA. I uninstalled/reinstalled the drivers, and now it's showing 7.9, the same as the Zotac. I really, really don't want to have to tear this thing down and start showing up at the store to exchange new stuff every day. Any ideas? I guess I am going to start by uninstalling Samsung Magician, since there might be a link between the problem and hard drive activity, but I wanted to get this question out there in case that turns up a zero. Edit: Oops. poo poo-post tag. Edit II: Probably not Samsung Magician. The problem is starting to creep up, and I haven't done anything except watch a video. Edit III: The loving Aero Performance warning is back. Centripetal Horse fucked around with this message at 02:25 on Jun 27, 2016 |
|
#
?
Jun 27, 2016 01:44
|
|
|
|
| # ? Apr 27, 2024 21:41 |
|
|
Just to be clear, you did reinstall Windows from scratch on this machine?
|
|
#
?
Jun 27, 2016 02:37
|
|
Alereon posted:Just to be clear, you did reinstall Windows from scratch on this machine? Yes. I've reinstalled Windows, reinstalled all drivers, swapped out the RAM, moved the RAM into different slot pairs, swapped out the graphics card, run the machine with Samsung Magician installed and uninstalled, installed up-to-date video codecs, installed up-to-date chipset drivers, and installed the latest BIOS. It's stuttering and hosed up, right now, but I just ran the Windows Experience Index, and the scores didn't change. I just had a hunch. I'm looking at CPU temp, now, and it maxed out at 64 degrees Celsius, and one core hit 79 on a benchmark test. It's waffling between 33 and 60ish a lot. I think something's up.
|
|
|
#
?
Jun 27, 2016 04:24
|
|
|
Those temperatures aren't concerning at all. Did you make sure the power cables were connected to the graphics card?
|
|
#
?
Jun 27, 2016 05:55
|
|
Alereon posted:Those temperatures aren't concerning at all. Did you make sure the power cables were connected to the graphics card? Sixty-four at idle is not normal. I don't think 79 after a five-second benchmark from the sensor software is great, either. Fake edit: I went and looked it up, in case I was wrong. Intel doesn't seem to provide temperature ranges (I didn't see them in the data sheets), but some geek sites with actual test ranges showed numbers like these: Idle/Normal/Max Core i7-6700K* 26 to 35°C/53 to 70°C/72°C I didn't look at the temperature while anything was running / during the worst of the stuttering. I'll do that tomorrow after work. The graphics card is powered, and in a 16x slot that is running at 16x according to BIOS.
|
|
|
#
?
Jun 27, 2016 06:21
|
|
|
What program are you using to monitor temperatures, and what are you using to test? The temperatures you've described (mid-30C idle, 60C range under load, 70C range under benchmark) are completely normal. Your CPU will probably hit the 90C range under an AVX2 benchmark like Prime95 and won't throttle until 100C. You are certainly nowhere near the range where you'd be seeing problems with your system due to the CPU temperatures.
|
|
#
?
Jun 27, 2016 06:59
|
|
Alereon posted:What program are you using to monitor temperatures, and what are you using to test? The temperatures you've described (mid-30C idle, 60C range under load, 70C range under benchmark) are completely normal. Your CPU will probably hit the 90C range under an AVX2 benchmark like Prime95 and won't throttle until 100C. You are certainly nowhere near the range where you'd be seeing problems with your system due to the CPU temperatures. I used RealTemp, and another that I can't recall the name of. I know the i7s can be stable (supposedly) up to 100c for periods of time. The figures in the mid-30s were reference ranges. My CPU hit the 60s while idling. The "benchmark" that caused the reading of 79 was just the quicky built into RealTemp. It's idling at 45ish, right now. I hit low 70s by turning on the computer and opening this browser. It's definitely running warm, but it doesn't seem to be getting out of hand. I'm still looking for the problem. I've already had to make two trips to the store to return and/or exchange things. This is getting annoying. Right now, there is a visible delay when I type, and my Aero effects are poo poo. I don't suppose there's anything in the BIOS settings of this board that I should be considering, is there?
|
|
|
#
?
Jun 28, 2016 00:01
|
|
|
I'd use onboard video temporarily and see if the issues continue like that. Since you have four HDs it wouldn't hurt to run the portable zip edition of CDI to check HD health too: http://crystalmark.info/download/index-e.html
|
|
#
?
Jun 28, 2016 22:10
|
|
Zogo posted:I'd use onboard video temporarily and see if the issues continue like that. It would appear that the issue goes away when using onboard video. Is my PCIE x16 slot hosed? Is it a chipset issue? I've already replaced the graphics card. Damnit. This is starting to smell like a complete teardown.
|
|
|
#
?
Jul 1, 2016 03:32
|
|
|
Centripetal Horse posted:It would appear that the issue goes away when using onboard video. Is my PCIE x16 slot hosed? Is it a chipset issue? I've already replaced the graphics card. Damnit. This is starting to smell like a complete teardown. I'd run DDU http://www.wagnardmobile.com/?q=display-driver-uninstaller-ddu- and then install older drivers like http://www.nvidia.com/download/driverResults.aspx/96073/en-us In recent months there's been some bad video drivers released that've caused people some issues. If older drivers don't help it could be a motherboard/PSU/GPU issue. Another thing to try would be using the GPU and just having the OS HD connected.
|
|
#
?
Jul 1, 2016 04:45
|
|
Zogo posted:I'd run DDU http://www.wagnardmobile.com/?q=display-driver-uninstaller-ddu- and then install older drivers like http://www.nvidia.com/download/driverResults.aspx/96073/en-us I don't think it's a driver issue. The drivers on the CD were definitely not new, and I had problems from the start. I guess it could be a power issue, but that seems unlikely. Two Samsung SSDs, and two platter drives, make up the entirety of the accessories I have connected. The platter drives get virtually no use. Those are both still worth a shot, I guess. I wanted to also try moving the graphics card to the other x16 slot, but it's too long, and butts up against the 2.5" enclosure where the SSDs are mounted.
|
|
|
#
?
Jul 1, 2016 05:44
|
|
|
I didn't see any evidence that the issue was power-related. To be safe, I purchased a new motherboard, processor, and power supply. The motherboard was the prime suspect, so I swapped that, first. I immediately noticed a drop in processor temperature. Instead of up to 60-something, the processor is running at 29-34 degrees Celsius. Some of the time. I went to a site that uses a tiny bit of Flash, and I thought I noticed the problem come back a little bit. I went to Kongregate, and the problem came back nearly full force almost immediately. I noticed that dwm.exe went from its usual 0% processor usage to 12-15%. I also saw that processor temperature jumped into the high-50s, and stayed there, occasionally fluctuating upward. Disk usage was also spiking randomly into multi-megabyte territory, even with nothing (visible) going on. Killing dwm did nothing; it immediately respawned with the higher CPU usage. This was definitely not a problem before I built the new system. I'm not sure whether to bother swapping in the new processor and power supply, or not. I can't say it's 100% a Flash/browser plugin issue, but I can definitely reproduce the problem by visiting Flash-heavy sites. Edit: Yep. I accidentally visited a site that used a tiny bit of Flash, and it happened, again. Even busting out a code:Edit II: Logging out doesn't fix the issue, either. I'd try switching users, but I only have one account set up. Centripetal Horse fucked around with this message at 05:38 on Jul 2, 2016 |
|
|
#
?
Jul 2, 2016 02:17
|
|
|
|
| # ? Apr 27, 2024 21:41 |
|
|
Were both video cards made by EVGA? I've noticed more bad EVGA cards in here over the years than others. Did you try just having the OS drive connected during these tests? I would guess bad PSU before bad processor but both are possible. I'd also install Windows 10 unless you had a reason for not doing so.
|
|
#
?
Jul 5, 2016 22:44
|
|