GeForce 6100 Crashing

[MarioStarz]

For whatever reason my pc keeps crashing due the GeForce 6100 (its a chipset) in my pc, I checked my temperatures and this what I saw:

This on idle by the way:
CPU: 124-126 °F
Ambient: 82 °F
Remote: 79 °F
HD0: 84 ° F
Temp1: 104 °F
Core: 102 °F
GPU: 32°F

I'm pretty sure my GPU isn't 32 °F, I clocked it down with ATITool and that makes my pc last longer.

When it crashes its a bunch of think colored horizontal lines going down the screen, showing colors of whatever was on the screen at that time.

Also I get this error signature from Windows XP:

BCCode: 9c BCP1:00000004 BCP2: 80546DF0 BCP3: B20000000 BCP4SP00070F0F OSVer: 5_1_2600 SP: 3_0 Product: 256_1

and I limit the ram to 512mb.

It seems to only crash when World of Warcraft: The Burning Crusade.

Any help would be strongly appreciated.

 

[VirtualLarry]

No offense, but that's some pretty creaky kit. XP still? On the internet? Good luck.

Anyways, list your COMPLETE PC specs, including Power Supply (PSU) brand and the amps on each rail. (You have to take the side off of your PC and actually look at the PSU, system diag / info tools won't be able to tell you that info.)

 

[MarioStarz]

No offense, but that's some pretty creaky kit. XP still? On the internet? Good luck.

Anyways, list your COMPLETE PC specs, including Power Supply (PSU) brand and the amps on each rail. (You have to take the side off of your PC and actually look at the PSU, system diag / info tools won't be able to tell you that info.)

This computer has been with my family for years, it was my grandpa's, and then given to my parents, then lost for a while until I found last year at my father's house.

I believe my PSU amps on each rail is: 25 amps or 300? (Not quite sure never heard this term before)



Here are the specs:
PSU:

Brand: ISO
Model: ISO-400 4
Power Capacity: 300W
AC Input: 115V/230V,8/5A,60/50HZ
Cooling Fan: 80MM Rear Fan

CPU: AMD Sempron 3300+ Processor (64-bit) (2GHz, , 128KB L2 cache)

Chipset: NVIDIA nForce 410

Memory: 768 MB (I only use 512 MB)
Expandable to 2GB

Hard Drive: 100GB (7200rpm, 2MB cache)32

Optical Drive: 48x CD-RW/DVD combo drive

Media Reader: 8-in-1 digital media manager (Secure Digital (SD), smart media, micro drive, memory stick, memory stick pro, compact flash, mulitimedia card, USB 2.0)

Video: NVIDIA GeForce 6100 GPU
up to 128MB of shared video memory

PCI-Express (PCI-E x16) slot available

Sound: 6-channel (5.1) AC97 audio

Network: 10/100Mbps integrated Ethernet LAN (RJ-45 port) (I took off the network card because I didn't need because I got Belkin USB wireless receiver)

Modem: 56K ITU V.92-ready fax/modem (RJ-11 port)

Ports/Other: 5 USB 2.0 (4 in back, 1 in media reader), 1 VGA external connector, 1 serial port, 1 parallel port, 2 PS/2 ports (keyboard and mouse), 5 audio ports (2 in front, 3 in back)

 

[VirtualLarry]

Hmm, that's pretty old. Do you know if you have DDR or DDR2 memory? That would tell us if you have a Socket 939 CPU (DDR) or an AM2 CPU (DDR2). (Edit: CPU-Z can tell you, look under the two memory tabs, and see if it lists DDR or DDR2.)

Anyways, is this a branded OEM rig, or a custom-built one?

A rig that old, is probably close to the end of it's usable lifespan, or possibly at least the PSU.

Have you checked your board for bad capacitors? They may have a bulge in them, be discolored, or leaking something. They are the little "tin cans" on the mobo.

It's unlikely that the IGP in the 6100 is overheating, although the thermal paste used to affix the heatsink to the chipset could be dried out too.

I mean, is it feasable to upgrade (budget-wise)? Or do you need to make that machine last?

It's really not worth putting much into, other than perhaps a cheaper 300-400W PSU, if the flakiness is caused by the PSU.

You can run OCCT 4.x, and do the Linpack test first, and check the CPU temps, make sure that they don't go much above 55-58C. (Not F, C) See if it crashed the PC. If it does, then you might have either temp-related or PSU-related problems.

Btw, has the CPU heatsink been cleaned out?

If it has been well cleaned out, consider new thermal paste for the CPU, if it's still overheating. Otherwise, consider replacing the PSU.

Edit: Is there a particular reason that you are limiting your RAM to 512MB, when you have 768MB? In XP, that seems unnecessary, unless that is helping with the crashes? If so, you should do some memory testing, with a bootable Memtest86+ (the plus is important) CD, and test with all sticks in first, then if you get errors, test each stick individually. You might need to test overnight to get a good test.

 

[VirtualLarry]

Network: 10/100Mbps integrated Ethernet LAN (RJ-45 port) (I took off the network card because I didn't need because I got Belkin USB wireless receiver)

It's remotely possible that your crashed could be caused by the wireless driver.

There's a few programs that I've read recommended on here, called "WhoCrashed" and nirsoft's "BlueScreen Viewer", that will read memory dumps from blue-screen crashes, and tell you the faulting driver or piece of kernel code that faulted. If it's all the same driver, then it's probably a driver bug. If it's random, then it's probably hardware issues, which given the age of the machine, wouldn't surprise me all that much.

 

[MarioStarz]

I have DDR2.
It's prebuild.
I checked for bad capacitors, they all appear to be in working order.
I wasn't aware that there was a heatsink for the GPU, the paste is probably as old as the pc itself.
I need to make it last.
The previous PSU, I broke (it shocked me a little, and the memory was damaged but fixed since then) because I forgot to unplug my computer. *Ultimate Facepalm*.
I've replaced cpu thermal paste twice, the first time there wasn't enough.
When I limit the RAM it lasts longer, and I have used Memtest86+ (not overnight though), and nothing was wrong with RAM.

 

[MarioStarz]

It's remotely possible that your crashed could be caused by the wireless driver.

There's a few programs that I've read recommended on here, called "WhoCrashed" and nirsoft's "BlueScreen Viewer", that will read memory dumps from blue-screen crashes, and tell you the faulting driver or piece of kernel code that faulted. If it's all the same driver, then it's probably a driver bug. If it's random, then it's probably hardware issues, which given the age of the machine, wouldn't surprise me all that much.

Hmmm.. Never thought about that, but I never got and bluescreen crashes.

 

[VirtualLarry]

Sorry for the slow reply, got busy downloading Ubuntu 15.10 and derivatives.

Anyways, good that you've tried Memtest86+.

I had forgotten that you said that your screen going into a striped screen. Does the PC itself freeze up when it does that? Do you still hear music playing from games, and can you hit NumLock and ScrLock and CapsLock and see the lights on the keyboards blink on and off when you hit those keys?

I'm thinking, the easiest thing that you could do, is throw in a cheap GPU. Something like a GT720 or GT730 (or GT630 if you can find one cheaply). Ideally, you would be able to borrow one from a friend, before committing to spend anything.

Those lower-end GPUs, are: 1) the most inexpensive, for performance, 2) faster than your GF 6100 IGP, by far, although in the grand scheme of cards, they're not even really gaming cards, and 3) don't require supplemental PCI-E 6-pin power connectors.

If it's purely an IGP overheating, or otherwise an IGP hardware problem, or an IGP driver problem, then a new GPU would likely fix it.

HOWEVER.

The IGP silicon is in the chipset. If it's overheating, or damaged (from voltage or heat, or from when your PSU "shocked" you and died), then your system chipset may be slightly damaged, and may not work 100% satisfactorily with a replacement PCI-Express GPU.

Edit: Still, I would like you to download OCCT, and run the "Linpack" test, and watch the graphs, especially the CPU Core Temp graph(s), and then tell me the max it gets up to, or if it gets above 58C, and / or crashes or does the screen with the lines.

http://www.ocbase.com/index.php/download

 

[MarioStarz]

My PC will completely freeze, or sometimes it will bring back up the last frame before it crashed, then immediately go back to the lines, and none of the keys do anything.
and no sound.
Even before the PSU got replaced it still crashed just the same.

Your probably right, I should probably just buy a dedicated card, I saw one on walmarts website and it looked decent. http://www.walmart.com/ip/EVGA-GeForce-210-512MB-DDR3-PCI-Express-Graphics-Card/15739049

 

[MarioStarz]

Sorry for the slow reply, got busy downloading Ubuntu 15.10 and derivatives.

Anyways, good that you've tried Memtest86+.

I had forgotten that you said that your screen going into a striped screen. Does the PC itself freeze up when it does that? Do you still hear music playing from games, and can you hit NumLock and ScrLock and CapsLock and see the lights on the keyboards blink on and off when you hit those keys?

I'm thinking, the easiest thing that you could do, is throw in a cheap GPU. Something like a GT720 or GT730 (or GT630 if you can find one cheaply). Ideally, you would be able to borrow one from a friend, before committing to spend anything.

Those lower-end GPUs, are: 1) the most inexpensive, for performance, 2) faster than your GF 6100 IGP, by far, although in the grand scheme of cards, they're not even really gaming cards, and 3) don't require supplemental PCI-E 6-pin power connectors.

If it's purely an IGP overheating, or otherwise an IGP hardware problem, or an IGP driver problem, then a new GPU would likely fix it.

HOWEVER.

The IGP silicon is in the chipset. If it's overheating, or damaged (from voltage or heat, or from when your PSU "shocked" you and died), then your system chipset may be slightly damaged, and may not work 100% satisfactorily with a replacement PCI-Express GPU.

Edit: Still, I would like you to download OCCT, and run the "Linpack" test, and watch the graphs, especially the CPU Core Temp graph(s), and then tell me the max it gets up to, or if it gets above 58C, and / or crashes or does the screen with the lines.

http://www.ocbase.com/index.php/download

Well, for my needs anyway.

 

[VirtualLarry]

wasn't aware that there was a heatsink for the GPU, the paste is probably as old as the pc itself.

It's the system chipset, the IGP is part of that silicon. Those can get quite hot, as they often have just a tiny little passive heatsink, and it's possible that if the paste dried up, then it could overheat after a time and cause the crashes.

The previous PSU, I broke (it shocked me a little, and the memory was damaged but fixed since then) because I forgot to unplug my computer. *Ultimate Facepalm*.
I have used Memtest86+ (not overnight though), and nothing was wrong with RAM.

On one hand, if the PSU damaged the RAM at one point, it may have damaged the memory-controller that is built into the CPU, that talks to the RAM too. On the other hand, if an overnight Memtest86+ didn't throw errors, then you're probably good.

One other thing to note is, Memtest86+ does NOT test memory that is "reserved" for the IGP. The only way that I know of, to test all of the memory sticks, is to swap some of the RAM slots around between overnight tests, to ensure that the memory that the IGP reserves, is taken from different sticks, between different tests.

 

[MarioStarz]

It's the system chipset, the IGP is part of that silicon. Those can get quite hot, as they often have just a tiny little passive heatsink, and it's possible that if the paste dried up, then it could overheat after a time and cause the crashes.

On one hand, if the PSU damaged the RAM at one point, it may have damaged the memory-controller that is built into the CPU, that talks to the RAM too. On the other hand, if an overnight Memtest86+ didn't throw errors, then you're probably good.

One other thing to note is, Memtest86+ does NOT test memory that is "reserved" for the IGP. The only way that I know of, to test all of the memory sticks, is to swap some of the RAM slots around between overnight tests, to ensure that the memory that the IGP reserves, is taken from different sticks, between different tests.

OK, tonight I'll test and see if the memory is bad.

 

[VirtualLarry]

Your probably right, I should probably just buy a dedicated card, I saw one on walmarts website and it looked decent. http://www.walmart.com/ip/EVGA-GeForce-210-512MB-DDR3-PCI-Express-Graphics-Card/15739049

While that is a discrete card, that's awfully slow. I would try to afford a GT430, GT620, GT630, GT720, or GT730. Bonus if it has 128-bit memory bus, and / or DDR5 instead of DDR3.

Have you checked out EVGA's B-stock video card page? That's often a good place to get cards.

http://www.evga.com/Products/ProductList.aspx?type=8

If you can afford it, I would get this card:
http://www.evga.com/Products/Product.aspx?pn=01G-P3-3736-RX

It claims to have GDDR5, and 128-bit memory bus, and 96 CUDA cores. That should be good for gaming. (Especially if you are running on a 6100 IGP.) $39.99

http://www.evga.com/Products/Product.aspx?pn=02G-P3-2639-RX
The GT630 for $29.99 at the end of the list, with 2GB of DDR3 wouldn't be a bad choice either. (Although GDDR5-based cards are generally preferred for gaming.) If that's the limit of your price range ($30), then it blows away that card from Walmart.

 

[MarioStarz]

While that is a discrete card, that's awfully slow. I would try to afford a GT430, GT620, GT630, GT720, or GT730. Bonus if it has 128-bit memory bus, and / or DDR5 instead of DDR3.

Have you checked out EVGA's B-stock video card page? That's often a good place to get cards.

I'll check out that page, real quick!

 

[MarioStarz]

While that is a discrete card, that's awfully slow. I would try to afford a GT430, GT620, GT630, GT720, or GT730. Bonus if it has 128-bit memory bus, and / or DDR5 instead of DDR3.

Have you checked out EVGA's B-stock video card page? That's often a good place to get cards.

http://www.evga.com/Products/ProductList.aspx?type=8

If you can afford it, I would get this card:
http://www.evga.com/Products/Product.aspx?pn=01G-P3-3736-RX

It claims to have GDDR5, and 128-bit memory bus, and 96 CUDA cores. That should be good for gaming. (Especially if you are running on a 6100 IGP.) $39.99

http://www.evga.com/Products/Product.aspx?pn=02G-P3-2639-RX
The GT630 for $29.99 at the end of the list, with 2GB of DDR3 wouldn't be a bad choice either. (Although GDDR5-based cards are generally preferred for gaming.) If that's the limit of your price range ($30), then it blows away that card from Walmart.

This one looks better than the Walmart one: http://www.evga.com/Products/Product.aspx?pn=02G-P3-2639-RX

 

[MarioStarz]

This one looks better than the Walmart one: http://www.evga.com/Products/Product.aspx?pn=02G-P3-2639-RX

LoL, I didn't see the rest of that post!

 

[VirtualLarry]

Your probably right, I should probably just buy a dedicated card, I saw one on walmarts website and it looked decent. http://www.walmart.com/ip/EVGA-GeForce-210-512MB-DDR3-PCI-Express-Graphics-Card/15739049

If you looked at the specs on that card, they are pretty slow. I think that they would even be slower than your IGP!

16 processing cores
Memory: 512MB, 32-bit DDR3

Your system memory bus is either 64-bit (single-channel), or 128-bit (dual-channel), and I think that the 6100 has maybe 16 cores too.

In contrast, the two cards I linked have 128-bit memory buses, and at least 96 cores. Not to mention, more RAM.

 

[VirtualLarry]

This one looks better than the Walmart one: http://www.evga.com/Products/Product.aspx?pn=02G-P3-2639-RX

That one should be fine. Not as fast as the $39.99 GDDR5 one, but it should be playable, assuming that it works in your system. (It should, I'm just referring to hoping that the chipset and RAM is OK.)

 

[VirtualLarry]

Edit: Still, I would like you to download OCCT, and run the "Linpack" test, and watch the graphs, especially the CPU Core Temp graph(s), and then tell me the max it gets up to, or if it gets above 58C, and / or crashes or does the screen with the lines.

http://www.ocbase.com/index.php/download

Do this still, because if CPU temps get too high, or it crashes / freezes, without a GPU load, then you have other issues that need to be addressed, and a new GPU may not help.

 

[MarioStarz]

Do this still, because if CPU temps get too high, or it crashes / freezes, without a GPU load, then you have other issues that need to be addressed, and a new GPU may not help.

OK, thanks man!

 

[VirtualLarry]

Just a cautionary note: Stress tests like Linpack, while safe to run on "solid, working" systems, may be dangerous to run on systems that are: 1) wildly overheating (although, by default, OCCT will stop the test if it gets to hot, so that shouldn't harm anything), 2) have weak power supplies, or 3) have weak CPU VRM circuits.

The reason for this is, they draw a lot of power and generate a lot of heat (though generally not in excess of CPU TDP).

This is the best way to determine if temps or weak / bad PSU is the cause of crashes, however.

OCCT provides graphs of voltage and temps, that you can watch, to see if the +12V line goes too low during the tests, or if temps get too high.

Edit: By default, OCCT runs the tests for an hour, but you really only need 10-15 minutes for a pass / fail.

 

[MarioStarz]

Just a cautionary note: Stress tests like Linpack, while safe to run on "solid, working" systems, may be dangerous to run on systems that are: 1) wildly overheating (although, by default, OCCT will stop the test if it gets to hot, so that shouldn't harm anything), 2) have weak power supplies, or 3) have weak CPU VRM circuits.

The reason for this is, they draw a lot of power and generate a lot of heat (though generally not in excess of CPU TDP).

This is the best way to determine if temps or weak / bad PSU is the cause of crashes, however.

OCCT provides graphs of voltage and temps, that you can watch, to see if the +12V line goes too low during the tests, or if temps get too high.

OK, thanks for the tip!

 

[VirtualLarry]

I've PMed someone else for backup, in case I'm missing something.

I then realized, that I never asked you to try newer or older drivers.

Have you ever updated the NVidia video drivers? It's possible, that if they are the original video drivers that were released when the computer was built, that they could have bugs that were addressed in newer drivers.

Anyways, if the OCCT:Linpack runs OK for 10-15 minutes, and doesn't overheat and crash, then run the OCCT video card / GPU test for 15 minutes, and see if that crashes.

If neither one of those crash, then your hardware is probably OK, and it probably is a video driver bug problem.

I will say, I'm kind of surprised that you can even play WoW on 512MB of RAM.

 

[MarioStarz]

I've PMed someone else for backup, in case I'm missing something.

I then realized, that I never asked you to try newer or older drivers.

Have you ever updated the NVidia video drivers? It's possible, that if they are the original video drivers that were released when the computer was built, that they could have bugs that were addressed in newer drivers.

I had the wrong video drivers at first, then I found the right ones by using the auto detect feature on Nvidias website.

 

[VirtualLarry]

I had the wrong video drivers at first, then I found the right ones by using the auto detect feature on Nvidias website.

Hmm, those should be the right ones, then.