31 replies to this topic

[NeoAres]

Okay, I have been working on this problem for a couple weeks now, and up until today it's been bewildering the heck out of me. This morning I finally reached a shred of consistency with the problem, which is why I feel comfortable bringing it forward.

First, my specs, bc I know you'll ask:
Mobo: Biostar TA970 (new)
CPU: AMD FX6300 (new)
GPU: XFX (Radeon) R7 260X (1 y/o)
RAM: 16GB Dual Channel (1 y/o)
Power: Antec 900w (new)
OS: Windows 7 Ultimate
Other attached stuff includes 2 HDDs, 2 DVD-RWs (all very old), PCI-E wireless card (1 y/o), Bluetooth mouse (1 y/o).

The reason I mention the other stuff is because I would like to rule out power drain as a culprit. The power supply is brand new and I'd rather not believe I didn't get a big enough one.

Anyhow, the problem (well, more like the interim solution): if I unplug both my DVD-RWs and my second HDD from the system, MWO runs as smooth as water, even on "very high" setting. If I plug even one of these items back in, the game becomes choppy on all but the "low" setting. However, here's the part that has really boggled my mind and made me think this isn't a simple power drain issue: it has been smooth while all devices have been plugged in (while I was troubleshooting with other changes), but none of my previous fixes have worked consistently.

For example, one time I put my mobo on pegs to test GPU contacts as a culprit. The fix worked. I then put the mobo on washers instead so I could screw the GPU into the case. The game got choppy again. I put the mobo back on the pegs to repeat the "fix", and the game was still choppy. This is just one of a half-dozen instances where I was able to get the performance I desired once, but was then unable to repeat the feat under identical conditions later. Until today I've felt like I was dealing with a human baby instead of a machine because it's been acting totally irrationally.

Anyhow, today I managed to get consistency. Every time I unplug my non-essential SATAs from the system, I get the performance I want. Every time I plug one or more back in, it reverts to being choppy. It's a fix to be sure, but I certainly don't want to have to do this every time I want to play MWO, so I'd like some opinions on why this is happening, particularly verification or ruling out of power as an issue while I still have time to exchange the power supply.

Things I've already done:
1. Updated BIOS
2. Updated drivers for Mobo, CPU, GPU, and Wireless card
3. Mobo is on washers
4. Checked temperature (CPU peaks at 18C and system running at 26C)
5. Fully updated Windows
6. Virus/Malware check
7. Defragged HDD

Things I've already tried to no effect:
1. Overclocking GPU and/or CPU
2. Turning off all nonessential programs
3. Unplugging power cord for long period of time
4. Plugging in just the second HDD, or just one of the DVD-RWs. All three must be unplugged to get desired effect.

Does this issue sound familiar to anyone? Anyone have any novel ideas for me to try? Next on my list of fixes is nuking my HDD, which I'd rather not do if there's an easier way.

[Flapdrol]

Set the windows power plan to "high performance" if it's not already.

[Oderint dum Metuant]

That actually doesn't make a whole lot of sense, i could understand if the problem disappeared with just the HDD removed, as it could then be that something (program/malware/virus) eating clock cycles of that AMD CPU (which is weak for MWO by the way)

Try flaps idea and see what happens.

The board and CPU you list as new, what was the old stuff? do you still have it?
And did you do a fresh install of windows after your latest swap of parts?

Edited by DV McKenna, 30 December 2014 - 11:18 AM.

[Thorqemada]

The PCI Bus Driver from AMD does usually not Install - in the Device Manager go to the PCI Bus and install the Driver from there - it should be in the AMD/Catalyst Omega Directory - see if that helps.

Do you run your HDD/DVD RWs in IDE comaptible mode (Bios Setting) ?
AFAIK the AHCI Driver is a bugged...
(which is why i am running my PC in the IDE-comaptible mode)

Power should not be a problem but could if to many devices hang on only 1 powerline - once a single Floppy Drive brought my own PC to the knees

Good Luck!

Edited by Thorqemada, 30 December 2014 - 12:13 PM.

[xWiredx]

Another possibility is that the CPU is waiting while attempting to read/write/otherwise engage one of the other drives and it is throwing errors (ie one of those devices could be experiencing issues). A similar possibility is the mobo's SATA controller or one or all of the SATA cables. It could be a short somewhere, that is definitely a possibility I suppose, but there are a few things here to rule out, and each one is going to be kind of weird.

After making sure the proper software is in place as mentioned above (drivers, BIOS settings)... Optical drives are cheap, it would be easy to get a new one and plug it in for testing. HDDs... a little more expensive. If you have another HDD you can plug in to test with, I would. Eliminating your specific drives would be a a good first test if possible. If the issue still happens with a newer, known good drive of any kind, then we can move onto the cables. Try a new SATA cable. If the problem persists, we're looking at the SATA plugs and then the SATA controller on the mobo.

[NeoAres]

Okay, thank you all for your feedback. Here is how it has gone so far:

Flapdrol: it was already at High Performance
McKenna: I did not try a fresh install of Windows. My CD is elsewhere. I may try that if the problem crops up again.
Wired: A cable issue seems unlikely since I've tried multiple configurations. The mobo's sata controller is a possibility; is that something I can fix or is it just faulty?
Thorquemada: I tried implementing both of your solutions and this is what happened.
1. PCI Bus Driver switched, but sadly no effect.
2. I went into BIOS to discover that system was already in Native IDE mode. I switched it to Legacy IDE mode and guess what? It worked! Performance spiked.
3. I can't explain why, but out of curiosity I switched to AHCI mode. The computer crashed on bootup. I guess that's out of the question...
4. I went back to Legacy IDE mode with all drives attached and performance was back to crap. Nice going, me.
5. I unplugged all nonessential SATAs again. Performance was still crap. I got worried my control wasn't actually a control after all.
6. I went back into BIOS and switched back to Native IDE, the true control. Performance went back up (phew!)
7. I plugged the rest of the SATAs in again, remaining in Native IDE form. Performance stayed up. (WTFyay!)
I don't know why it's working now in a configuration that it previously did not work in, but all my stuff is plugged in and the game is working at proper performance so I am not going to touch it anymore. Curiosity is not a worthy reason to continue this headache.

Once again, thank you all for your help! Even though I don't think we found a logical solution, at least we found a complete and, hopefully, permanent one.

[NeoAres]

Never mind. I rebooted and performance went back down. I've been unable to repeat the success since except in the control state, so whatever success there was is the perfect example of the inconsistent success I've been achieving.

One clue that's driving me away from the SATA-specific explanation and more towards a windows explanation is that I have a mp3 player that I can attach to the computer via usb. When I booted the computer in control state with it attached, performance went down. When I unplugged it and rebooted, performance went back up. This leads me to believe that the problem probably isn't with any specific piece of hardware, but with windows' interaction with it's various pieces of hardware. Seems like I should hurry up and try that windows reinstallation.

[Soul Tribunal]

NeoAres, on 30 December 2014 - 01:11 PM, said:

Okay, thank you all for your feedback. Here is how it has gone so far:

Flapdrol: it was already at High Performance
McKenna: I did not try a fresh install of Windows. My CD is elsewhere. I may try that if the problem crops up again.
Wired: A cable issue seems unlikely since I've tried multiple configurations. The mobo's sata controller is a possibility; is that something I can fix or is it just faulty?
Thorquemada: I tried implementing both of your solutions and this is what happened.
1. PCI Bus Driver switched, but sadly no effect.
2. I went into BIOS to discover that system was already in Native IDE mode. I switched it to Legacy IDE mode and guess what? It worked! Performance spiked.
3. I can't explain why, but out of curiosity I switched to AHCI mode. The computer crashed on bootup. I guess that's out of the question...
4. I went back to Legacy IDE mode with all drives attached and performance was back to crap. Nice going, me.
5. I unplugged all nonessential SATAs again. Performance was still crap. I got worried my control wasn't actually a control after all.
6. I went back into BIOS and switched back to Native IDE, the true control. Performance went back up (phew!)
7. I plugged the rest of the SATAs in again, remaining in Native IDE form. Performance stayed up. (WTFyay!)
I don't know why it's working now in a configuration that it previously did not work in, but all my stuff is plugged in and the game is working at proper performance so I am not going to touch it anymore. Curiosity is not a worthy reason to continue this headache.

Once again, thank you all for your help! Even though I don't think we found a logical solution, at least we found a complete and, hopefully, permanent one.

FYI you only really need to worry about AHCI mode when using SSD's, and it needs to be set before the first installation of the OS or , like what happened with you, a BSOD awaits.

That being said this isn't a power issue. If your CPU or GPU didn't get enough power, they simply wouldn't work. They have a min-threshold they have to meet and a 900w PSU is about 2x what you need on that rig (approx). I run 3x the HW and I use at peak 400W.
My only two thoughts are, and this will take some tweaking in windows, is either a conflict of drivers that is eating CPU cycles *(legacy HW can do that from time to time)* or windows isn't setting the correct affinity or priority for MWO when other devices are running in the background. That you can adjust. I know MS tweaked some settings in their scheduler for AMD's CPU 'modules'. I'd try changing the priority (which I believe you can do in Task Manager , its been a while since I've had to do that).

What you should do to, is with the legacy HW unplugged see what processes are running in the Background, and they with the HW plugged in see what changes. If you notice something eating a lot of CPU cycles once the old HW Is plugged in, then you know where the problem is.

-ST

[Thorqemada]

Your Bios is up to date?
http://www.biostar.c...php?S_ID=623#dl
(only do this if you have experience with it)


Unlucky that it did not help permanently to switch the modes...

Edited by Thorqemada, 30 December 2014 - 02:11 PM.

[NeoAres]

Thorqemada, on 30 December 2014 - 02:07 PM, said:

Your Bios is up to date?
http://www.biostar.c...php?S_ID=623#dl
(only do this if you have experience with it)


Unlucky that it did not help permanently to switch the modes...

yes, I downloaded the update utility and it downloaded and installed the latest bios

Anyway, I just booted up again and performance is high now without any tinkering, but I'll try setting the priority higher next time it goes back down and see if that brings it back up.

Edited by NeoAres, 30 December 2014 - 05:52 PM.

[Golrar]

Ok, this is a far stretch and I am at work and my PC is admin locked down so I cannot check myself here. That being said - if I remember correctly, if you have additional drives in your Windows path, they can slow performance. Maybe these added devices are in the path command and it is searching through that extended path for libraries, drivers, cache, etc. Disconnecting those devices makes those entries int he path invalid and increases performance.

If it isn't an intermittent power issue, that's the only thing I can think of. I haven't seen this kind of issue since the older Windows iterations, but that is not to say it isn't possible in 7 or 8. You know what Spock said about the impossible and the improbable.

[MechWarrior4172571]

NeoAres, on 30 December 2014 - 11:05 AM, said:

Okay, I have been working on this problem for a couple weeks now, and up until today it's been bewildering the heck out of me. This morning I finally reached a shred of consistency with the problem, which is why I feel comfortable bringing it forward.

First, my specs, bc I know you'll ask:
Mobo: Biostar TA970 (new)
CPU: AMD FX6300 (new)
GPU: XFX (Radeon) R7 260X (1 y/o)
RAM: 16GB Dual Channel (1 y/o)
Power: Antec 900w (new)
OS: Windows 7 Ultimate
Other attached stuff includes 2 HDDs, 2 DVD-RWs (all very old), PCI-E wireless card (1 y/o), Bluetooth mouse (1 y/o).

The reason I mention the other stuff is because I would like to rule out power drain as a culprit. The power supply is brand new and I'd rather not believe I didn't get a big enough one.

Anyhow, the problem (well, more like the interim solution): if I unplug both my DVD-RWs and my second HDD from the system, MWO runs as smooth as water, even on "very high" setting. If I plug even one of these items back in, the game becomes choppy on all but the "low" setting. However, here's the part that has really boggled my mind and made me think this isn't a simple power drain issue: it has been smooth while all devices have been plugged in (while I was troubleshooting with other changes), but none of my previous fixes have worked consistently.

For example, one time I put my mobo on pegs to test GPU contacts as a culprit. The fix worked. I then put the mobo on washers instead so I could screw the GPU into the case. The game got choppy again. I put the mobo back on the pegs to repeat the "fix", and the game was still choppy. This is just one of a half-dozen instances where I was able to get the performance I desired once, but was then unable to repeat the feat under identical conditions later. Until today I've felt like I was dealing with a human baby instead of a machine because it's been acting totally irrationally.

Anyhow, today I managed to get consistency. Every time I unplug my non-essential SATAs from the system, I get the performance I want. Every time I plug one or more back in, it reverts to being choppy. It's a fix to be sure, but I certainly don't want to have to do this every time I want to play MWO, so I'd like some opinions on why this is happening, particularly verification or ruling out of power as an issue while I still have time to exchange the power supply.

Things I've already done:
1. Updated BIOS
2. Updated drivers for Mobo, CPU, GPU, and Wireless card
3. Mobo is on washers
4. Checked temperature (CPU peaks at 18C and system running at 26C)
5. Fully updated Windows
6. Virus/Malware check
7. Defragged HDD

Things I've already tried to no effect:
1. Overclocking GPU and/or CPU
2. Turning off all nonessential programs
3. Unplugging power cord for long period of time
4. Plugging in just the second HDD, or just one of the DVD-RWs. All three must be unplugged to get desired effect.

Does this issue sound familiar to anyone? Anyone have any novel ideas for me to try? Next on my list of fixes is nuking my HDD, which I'd rather not do if there's an easier way.

Mobo is supposed to make ground contact at the stand-off areas--it's how it is supposed to be. Don't need washers at all, in fact, it's SUPPOSED to be grounded with screws to the chassis. BUT, it has to be on stands, or you could short out the mobo's other components that are sticking out on the back against the ground of chassis. What you are describing thou, is some intermittent problem, and if I have to guess, I'd say your mobo's sata connectors are going bad. Try to install new cables or swap them with other compoents and see if that helps. Visually inspect the insides of the sata connectors on the mobo too.

[NeoAres]

Jesus DIED for me, on 31 December 2014 - 12:13 AM, said:

Mobo is supposed to make ground contact at the stand-off areas--it's how it is supposed to be. Don't need washers at all, in fact, it's SUPPOSED to be grounded with screws to the chassis. BUT, it has to be on stands, or you could short out the mobo's other components that are sticking out on the back against the ground of chassis. What you are describing thou, is some intermittent problem, and if I have to guess, I'd say your mobo's sata connectors are going bad. Try to install new cables or swap them with other compoents and see if that helps. Visually inspect the insides of the sata connectors on the mobo too.

It wasn't a ground issue, it was an imperfect fit issue. I noticed that my PCI cards weren't perfectly seated in their slots because the back panel edge was too "high up" (read: too far from mobo surface). Raising the mobo up on washers helped eliminate that imperfection so the cards are perfectly seated now (though, as I stated previously, that wasn't the problem in hindsight)

Well, I think I found the problem, and this one actually makes some sort of sense (though everything else leading up to it makes none). I moved my video card to the lower PCIx16 slot and the problems went away. I moved it back up to the upper x16 slot and the problems returned. I moved it back down to the lower slot and the problems went away again. I ran GPU-Z with the video card in each, including once when the GPU was running in the upper slot and performance was high, and here's what I discovered, regarding the Bus Interface:

1. Upper slot-high performance: PCI-E 3.0x16 @ x16 1.1 idle, 2.0 under load (which makes sense as the mobo's slots are 2.0 unfortunately)
2. Lower slot-high performance: PCI-E 3.0x16 @ x4 1.1 idle, 2.0 under load, which means that despite the lower x# rating, performance was still higher than the following...
3. Upper slot-low performance: PCI-E 3.0x16 @ x16 1.1 idle, 1.1 under load. Regardless of the game or the render test, the bus interface refuses to go up to 2.0 and I think this is the symptom that is cause the game's choppiness. So now that I have a symptom, I need to discover its source and, equally importantly, why that source is intermittent.

Edited by NeoAres, 02 January 2015 - 11:54 AM.

[MechWarrior4172571]

NeoAres, on 02 January 2015 - 08:27 AM, said:

It wasn't a ground issue, it was an imperfect fit issue. I noticed that my PCI cards weren't perfectly seated in their slots because the back panel edge was too "high up" (read: too far from mobo surface). Raising the mobo up on washers helped eliminate that imperfection so the cards are perfectly seated now (though, as I stated previously, that wasn't the problem in hindsight)

Well, I think I found the problem, and this one actually makes some sort of sense (though everything else leading up to it makes none). I moved my video card to the lower PCIx16 slot and the problems went away. I moved it back up to the upper x16 slot and the problems returned. I moved it back down to the lower slot and the problems went away again. I ran GPU-Z with the video card in each, including once when the GPU was running in the upper slot and performance was high, and here's what I discovered, regarding the Bus Interface:

1. Upper slot-high performance: PCI-E 3.0x16 @ x16 1.1 idle, 2.0 under load (which makes sense as the mobo's slots are 2.0 unfortunately)
2. Lower slot-high performance: PCI-E 3.0x16 @ x4 1.1 idle, 2.0 under load, which means that despite the lower x# rating, performance was still higher than the following...
3. Upper slot-low performance: PCI-E 3.0x16 @ x16 1.1 idle, 1.1 under load. Regardless of the game or the render test, the bus interface refuses to go up to 2.0 and I think this is the symptom that is cause the game's choppiness. So now that I have a symptom, I need to discover its source and, equally importantly, why that source is intermittent.

Ouch. I assume your upper slot is rated at 2.0? x4 PCIe is less than x16--this could account for extra power to the video card and/or motherboard, possibly, such that it works fine at a higher state. I suspect, based on what you described, either of these two things--1. Something got stuck in the slot (visually inspect) and/or some pin bent or worn out pad on the video card, or, 2. (this one is more likely) you don't have enough power coming to the video card and/or mobo. Check your power connections, rating of your power supply, and make sure all the extra power connectors on the motherboard are getting their juice--the 8 pin power connector (and/or 4 pin power connector(s)) by the CPU are plugged in. Reseat those power connectors if you would too, to make sure there is no oxidation issue at work. My theory is that, Not enough power supplied = lower speed state @ x16 slot, BUT, when you move it to x4 slot suddenly there is less of the bus lanes to feed juice to and so it senses sufficient power supply and jumps back to where it should be. My 2 cents. It definetely sounds like power issue in the motherboard.

[NeoAres]

Jesus DIED for me, on 02 January 2015 - 05:04 PM, said:

Ouch. I assume your upper slot is rated at 2.0? x4 PCIe is less than x16--this could account for extra power to the video card and/or motherboard, possibly, such that it works fine at a higher state. I suspect, based on what you described, either of these two things--1. Something got stuck in the slot (visually inspect) and/or some pin bent or worn out pad on the video card, or, 2. (this one is more likely) you don't have enough power coming to the video card and/or mobo. Check your power connections, rating of your power supply, and make sure all the extra power connectors on the motherboard are getting their juice--the 8 pin power connector (and/or 4 pin power connector(s)) by the CPU are plugged in. Reseat those power connectors if you would too, to make sure there is no oxidation issue at work. My theory is that, Not enough power supplied = lower speed state @ x16 slot, BUT, when you move it to x4 slot suddenly there is less of the bus lanes to feed juice to and so it senses sufficient power supply and jumps back to where it should be. My 2 cents. It definetely sounds like power issue in the motherboard.

I switched out the mobo, cpu, and power supply and the problem remains. The only remaining possibilities are the video card and the windows installation. The video card pins show no signs of physical damage, and neither does its auxiliary power connector, so I still hold out hope that it's not defective.

[Golrar]

Neo, I am still kind of skittish when I hear you say washers for mobo mounting. What exactly do you mean? Do you not have ANY posts (standoffs) under the board and previously had it mounted directly to the chassis? Or did you have the little standoffs underneath the board that you screwed into? I have never seen a case fit so bad that you had to raise up the mobo more to get the PCI cards to seat correctly. Did your rear ports and backplate not fit either?

[MechWarrior4172571]

NeoAres, on 02 January 2015 - 07:41 PM, said:

I switched out the mobo, cpu, and power supply and the problem remains. The only remaining possibilities are the video card and the windows installation. The video card pins show no signs of physical damage, and neither does its auxiliary power connector, so I still hold out hope that it's not defective.

That leaves GPU to try to switch with another good GPU as a test, and possibly power connectors going to the GPU. Try using different ones (if your power supply has multiple VGA power connectors). If you can borrow one from someone or have another reliable GPU--try using that. If you switched mobo and PSU then there is little else left to try. By the way, if you mount your mobo to the chassis on washers but WITHOUT stands you could get a short somewhere and that could account for CPU and/or VGA throttling and giving you bad performance that way too. Be careful. Make sure your motherboard is ALWAYS mounted on metal stubbies/stand-offs. Never use washers, unless they are made of metal like the stand-offs because, like I said earlier, motherboards are supposed to make ground connection at screw points through the stand-offs with the chassis.

[MechWarrior4172571]

Golrar, on 02 January 2015 - 10:55 PM, said:

Neo, I am still kind of skittish when I hear you say washers for mobo mounting. What exactly do you mean? Do you not have ANY posts (standoffs) under the board and previously had it mounted directly to the chassis? Or did you have the little standoffs underneath the board that you screwed into? I have never seen a case fit so bad that you had to raise up the mobo more to get the PCI cards to seat correctly. Did your rear ports and backplate not fit either?

That is a scary point but a valid question. Also some whacky PC cases are off a bit and you have to bend things a little to get the slots aligned/tightened in them--in this case it's probably a good (and a relatively cheap) idea to get a new case for $30 or so. Rosewill sells good cases that don't cost alot (including rebate.)

Edited by Jesus DIED for me, 03 January 2015 - 01:51 AM.

[NeoAres]

Regarding the washers thing, just for you guys I removed them. There are posts and they were sitting on top of them. Now they are not.

Anyhow, the troubleshooting continues with a different system. I would replace the GPU if I could, but sadly it is the only PCI-E GPU I have.

New equipment specs:
Mobo: Biostar Hi-Fi A85W
CPU: AMD A-10 5800k Trinity
Power: Antec 550w
In addition, it's running in a different case with a different DVD-ROM and a different Wireless PCI card. The only things that are the same are the RAM, HDD, and GPU (and keyboard, mouse, and speakers).

With this configuration, not only does the issue still exist with the main PCIx16 slot, but it even carries over to the secondary slot (giving me x4 1.1 instead of x4 2.0). I'm going to try putting the new power supply (the 900w one) in this case and see if it helps the problem any. I'll let you know how it goes.

[MechWarrior4172571]

NeoAres, on 03 January 2015 - 12:02 PM, said:

Regarding the washers thing, just for you guys I removed them. There are posts and they were sitting on top of them. Now they are not.

Anyhow, the troubleshooting continues with a different system. I would replace the GPU if I could, but sadly it is the only PCI-E GPU I have.

New equipment specs:
Mobo: Biostar Hi-Fi A85W
CPU: AMD A-10 5800k Trinity
Power: Antec 550w
In addition, it's running in a different case with a different DVD-ROM and a different Wireless PCI card. The only things that are the same are the RAM, HDD, and GPU (and keyboard, mouse, and speakers).

With this configuration, not only does the issue still exist with the main PCIx16 slot, but it even carries over to the secondary slot (giving me x4 1.1 instead of x4 2.0). I'm going to try putting the new power supply (the 900w one) in this case and see if it helps the problem any. I'll let you know how it goes.

Trinity has a built in APU, correct? See how that behaves in MWO on everything low.. just for the interest's sake. Yeah, it could be anything, GPU, USB short in any of the USB peripherals, some short in a hard drive connector, DVD-ROM shorting out, etc... its hard to tell, most likely, though, I would say it's GPU itself that is bad. How is the GPU temperature? Maybe thermal paste in it needs to be reapplied? Maybe it's throttling because it's overheating? Check the back of the GPU for any visible damage on the component level.. really stretching here.