Jump to content

M17x R3 GPU upgrade to gtx 780m


veer

Recommended Posts

Hi, just now ran 5 mins of Prime95, temperatures stabilised at 84 degC on the CPU, room temperature 22 degC. This was with a consistent 54W power consumption according to Throttlestop at 2600Mhz full turbo. CPU fans stabalised at 4000rpm or 93%. Previous to me re-pasting the CPU temperatures were 93 degC under similar conditions with 4300rpm or 100%. Previously I just had the stock thermal paste that Dell applied at the factory, gradually over a period of a year and a half the CPU temperatures rose, even if I cleaned out dust from the heatsink. I repasted with Arctic Silver 5 about 3 months ago.

Yes, your CPU temperature is a little high considering its only during gaming (not 100% load), and it's not overclocked either. What's your room temperature? (I guess in New York this time of year it will be similar to mine).

Link to comment
Share on other sites

Yea about the same I would believe. I remember doing some renders with Revit architecture for a side job I did during the summer and the CPU was at 100% for 4 hours (renders for artificial lights and natural lights KILL CPU). The temps were about 90C - 95C... That was with the 2670QM tho. and it was before I dusted the vents. They were clogged.

I have a tube of Arctic silver 5 and ICD7. The ICD seems really "thick" compared to the AS5. I got a tube included when i bought my 780m on eBay. I used the AS5 on my GPU and CPU, cuz it was easier to apply. I read a lot of forums about AS vs ICD and it stated that the AS5 has a 200 hour cure time while the ICD7 does not.

You think I should switch the CPU paste to ICD7 as opposed to AS5? The gpu temps seem normal now. BF3 campaign on ultra gets my GPU to about high 50's to low 60's with the HWiNFO fan control, but the CPU is mostly 75C - 80C with the HWiNFO

Link to comment
Share on other sites

I don't think there's much difference in performance between pastes, maybe 5 degC max between best & worst, but that's assuming they're applied correctly & accurately. Sometimes it can be difficult to apply them properly, especially with low pressure heatsinks (like we have), then it's easier to do it properly with a softer paste. With our intel CPU on the M17xR3 I've tried the spread method & the line method with AS5, and both work about the same it seems - Arctic claim that you should use the spread method on mobile CPU's. I wouldn't bother using the IC Diamond, I heard it can scratch your chip & heatsink due to the hardness of the particles, which over time if you reapply a number of times, then I guess that could reduce the performance due to the surface imperfections. IC Diamond might give you 1 or 2 degC better than AS5 if both are applied properly.

You could test if your cooling system is working properly. Like this perhaps:

1) Run Prime95 for 5 mins, while looking in Throttlestop window to see what the clocks & Watt consumption of the CPU is. Note the highest temperature in Throttlestop window.

2) If your temperatures are over 84 degC by a significant amount and your Watt consumption on the CPU was less than 54W during the test, then you know the cooling system is not working as efficiently as mine. (Assuming your room temperature is similar). If it's not working as efficiently, then you could consider a repaste. Sound good?

EDIT: cure time changes with AS5 I've not noticed any significant difference, no lowering of temperatures over 200hours for me.

I was thinking of purchasing a cooling pad, do you think that it will help? I just have my laptop on a book to keep the bottom vents a little elevated. I noticed that in your sig, you have a "Zalman NC2000 notebook cooler". I checked up the reviews from the people who purchased it from amazon and most of them seemed quite pleased with it's performance. How will you rate it with regards to cooling?

Link to comment
Share on other sites

I was thinking of purchasing a cooling pad, do you think that it will help? I just have my laptop on a book to keep the bottom vents a little elevated. I noticed that in your sig, you have a "Zalman NC2000 notebook cooler". I checked up the reviews from the people who purchased it from amazon and most of them seemed quite pleased with it's performance. How will you rate it with regards to cooling?
Link to comment
Share on other sites

@link3rd A laptop cooler will improve temps 2-3C optimistic, vents intake distance 5cm - exhaust 20cm, we're talking about 100W cards and extreme processors so start with this CPU/GPU http://forum.techinferno.com/alienware-m17x-aw-17/24-m17x-retention-mod.html careful not to add too much pressure, use IC Diamond or Gelid Extreme thermal compounds. HWiNFO fan control will keep the laptop cooler, @Robbo default CPU tables/full blast 85C not constant - 90C, the R3 has a good cooling system but many times not well implemented thanks to 'users' loud noise levels complaints

M17xR3 GPU/CPU/PCH modded heatsinks, Gelid extreme/Fujipoly EXTREME Thermal pads, HWiNFO fan control, CPU temps probably a record for the R3, GPU all models so far...

x5v8.png

OC 1097/6180 OV 1.075v, 4.4 GHz NVIDIA GeForce GTX 780M video card benchmark result - Intel Core i7-2960XM Processor,Alienware M17xR3

Link to comment
Share on other sites

@J95, you really do like economy of words, so I don't fully understand what you're saying there. I think I get most of it, but I just can't be sure!

Interesting link for the CPU mod, reading now. Can see how that might be useful for XM users.

Link to comment
Share on other sites

@J95, you really do like economy of words, so I don't fully understand what you're saying there. I think I get most of it, but I just can't be sure!

Working now, (Latin studies) self-explanatory :D , start here :59: 150,452 views http://forum.techinferno.com/hwinfo32-64-discussion/65-alienware-fan-control.html

Interesting link for the CPU mod, reading now. Can see how that might be useful for XM users.

QM / GPU too !

Link to comment
Share on other sites

@link3rd A laptop cooler will improve temps 2-3C optimistic, vents intake distance 5cm - exhaust 20cm, we're talking about 100W cards and extreme processors so start with this CPU/GPU http://forum.techinferno.com/alienware-m17x-aw-17/24-m17x-retention-mod.html careful not to add too much pressure, use IC Diamond or Gelid Extreme thermal compounds. HWiNFO fan control will keep the laptop cooler, @Robbo default CPU tables/full blast 85C not constant - 90C, the R3 has a good cooling system but many times not well implemented thanks to 'users' loud noise levels complaints

M17xR3 GPU/CPU/PCH modded heatsinks, Gelid extreme/Fujipoly EXTREME Thermal pads, HWiNFO fan control, CPU temps probably a record for the R3, GPU all models so far...

x5v8.png

OC 1097/6180 OV 1.075v, 4.4 GHz NVIDIA GeForce GTX 780M video card benchmark result - Intel Core i7-2960XM Processor,Alienware M17xR3

Very interesting link... =)

EDIT: Come to think of it, I remember that whenever I screwed on my CPU heat sink, it wasn't tight at all. It was due to those clips. I feel I'll try that mod in the VERY near future....

Link to comment
Share on other sites

@Robbo and J95, Would you all mind sharing with me the max temperatures your 670XM and 780M both get at stock voltage and overvolted? I play hitman absolution, and I noticed that with the autofan option, I get about 65-70C at stock voltage and default clocks, 70-75C stock voltage @ 985/3000 MHz and 75-80-ish (+/- about 2C) at overvolt 12.5 mV @ 1007/3000 MHz...

I also screwed my CPU heatsink tighter after reading the link J95 posted above. Previously, I remembered not putting a lot of torque on the screws as I tightened them after pasting the CPU. I noticed better temperatures on the CPU while playing Hitman. I only played a little while, so I have yet to see the temps with CPU intensive games.

Link to comment
Share on other sites

I also screwed my CPU heatsink tighter after reading the link J95 posted above. Previously, I remembered not putting a lot of torque on the screws as I tightened them after pasting the CPU. I noticed better temperatures on the CPU while playing Hitman. I only played a little while, so I have yet to see the temps with CPU intensive games.

Retention mod (c-clips removed) very careful when screwing, don't add unnecessary pressure...you can break the die.

Link to comment
Share on other sites

@Robbo and J95, Would you all mind sharing with me the max temperatures your 670XM and 780M both get at stock voltage and overvolted? I play hitman absolution, and I noticed that with the autofan option, I get about 65-70C at stock voltage and default clocks, 70-75C stock voltage @ 985/3000 MHz and 75-80-ish (+/- about 2C) at overvolt 12.5 mV @ 1007/3000 MHz...

I also screwed my CPU heatsink tighter after reading the link J95 posted above. Previously, I remembered not putting a lot of torque on the screws as I tightened them after pasting the CPU. I noticed better temperatures on the CPU while playing Hitman. I only played a little while, so I have yet to see the temps with CPU intensive games.

Hi, here are my maximum GPU temperatures (670MX) while running 10mins of Heaven Benchmark (room temp = 20 degC, standard laptop automatic fan control (no HWInfo), Zalman NC2000 laptop cooler fans on max):

Stock: 0.925V 600Mhz core / 700Mhz memory = 56 degC (overall system Watt Consumption: 95W)

Overvolt #1: 0.975V, 1006Mhz core / 1100Mhz memory = 63 degC (overall system Watt Consumption: 125W)

Overvolt #2: 1.05V, 1124Mhz core / 1100Mhz memory = 66 degC (overall system Watt Consumption: 145W)

Laptop GPU fans only sped up to max rpm during periods of time during 'Overvolt #2'. (65degC is when max GPU fan RPM automatically kicks in)

Link to comment
Share on other sites

Hi, here are my maximum GPU temperatures (670MX) while running 10mins of Heaven Benchmark (room temp = 20 degC, standard laptop automatic fan control (no HWInfo), Zalman NC2000 laptop cooler fans on max):

Stock: 0.925V 600Mhz core / 700Mhz memory = 56 degC (overall system Watt Consumption: 95W)

Overvolt #1: 0.975V, 1006Mhz core / 1100Mhz memory = 63 degC (overall system Watt Consumption: 125W)

Overvolt #2: 1.05V, 1124Mhz core / 1100Mhz memory = 66 degC (overall system Watt Consumption: 145W)

Laptop GPU fans only sped up to max rpm during periods of time during 'Overvolt #2'. (65degC is when max GPU fan RPM automatically kicks in)

WOW! how is it that the temps are so low? I really doubt that hitman absolution is that graphics intensive... I'm running it on ultra. Where can I get this Heaven Benchmark program?

Retention mod (c-clips removed) very careful when screwing, don't add unnecessary pressure...you can break the die.

I left the c-clips in. I haven't removed them yet. I just applied more pressure to it as it currently is, because i remember that i intentionally left them "hand-tight". Taking the c-clip retention into account, threw my paranoia out the window and decided to screw it up tight and let the c-clips do it's job.

Link to comment
Share on other sites

WOW! how is it that the temps are so low? I really doubt that hitman absolution is that graphics intensive... I'm running it on ultra. Where can I get this Heaven Benchmark program?

Here's a good quality link for download of Heaven Benchmark:

Unigine Heaven Benchmark 4.0 Windows Download

I ran it at Quality Ultra, Tessellation Maximum, no Anti-aliasing, resolution set to my native screen size (yours probably 1920x1080).

Well, my temperatures should be a bit lower than yours because your 780M has a higher TDP (Watts) than my 670MX, at least when comparing stock settings - your card consumes a lot more Watts/more power. If you've got a Watt Meter you could plug it into your electical outlet so you could measure your overall system Watt consumption and compare that to mine to get a better comparison, although your CPU would use more power too, but it would give you an idea. As I mentioned in one of my posts a few days ago, the quality of the paste job on the GPU can make a big difference, if not done right then you can see large increases in temperature. If yours is getting really hot, then that could be the cause; I know j95 has showed you the c-clip mod, but it's probably easier & safer to make sure you've got a good basic paste job first, at which point you may not even need the c-clip mod which carries those extra risks.

EDIT: I don't think Hitman Absolution is a particularly good benchmark for testing max GPU temperature, because when I've run the Hitman Benchmark GPU usage never remained stable at 100%, there's quite a bit of CPU or platform limited behaviour in that one so GPU is not always at 100% usage, whereas the Heaven Benchmark keeps your GPU at 100% constant GPU usage (resulting in near worst case temperature scenarios while gaming). What's the maximum temperature you get during a 10 minute Heaven Benchmark run? (You can just let that benchmark run on a loop forever until you click on the Quit button, so useful for stability testing).

Link to comment
Share on other sites

Here's a good quality link for download of Heaven Benchmark:

Unigine Heaven Benchmark 4.0 Windows Download

I ran it at Quality Ultra, Tessellation Maximum, no Anti-aliasing, resolution set to my native screen size (yours probably 1920x1080).

Well, my temperatures should be a bit lower than yours because your 780M has a higher TDP (Watts) than my 670MX, at least when comparing stock settings - your card consumes a lot more Watts/more power. If you've got a Watt Meter you could plug it into your electical outlet so you could measure your overall system Watt consumption and compare that to mine to get a better comparison, although your CPU would use more power too, but it would give you an idea. As I mentioned in one of my posts a few days ago, the quality of the paste job on the GPU can make a big difference, if not done right then you can see large increases in temperature. If yours is getting really hot, then that could be the cause; I know j95 has showed you the c-clip mod, but it's probably easier & safer to make sure you've got a good basic paste job first, at which point you may not even need the c-clip mod which carries those extra risks.

EDIT: I don't think Hitman Absolution is a particularly good benchmark for testing max GPU temperature, because when I've run the Hitman Benchmark GPU usage never remained stable at 100%, there's quite a bit of CPU or platform limited behaviour in that one so GPU is not always at 100% usage, whereas the Heaven Benchmark keeps your GPU at 100% constant GPU usage (resulting in near worst case temperature scenarios while gaming). What's the maximum temperature you get during a 10 minute Heaven Benchmark run? (You can just let that benchmark run on a loop forever until you click on the Quit button, so useful for stability testing).

Thank you, I will download it later when I get home from work. I am really not looking forward to removing my gpu again tho. However, if I have to, I will. =(

P.S. My screen resolution is 1600 x 900

Link to comment
Share on other sites

Thank you, I will download it later when I get home from work. I am really not looking forward to removing my gpu again tho. However, if I have to, I will. =(

P.S. My screen resolution is 1600 x 900

Haha, 1600x900 with a 780M, that will be some awesome framerates you'll see! That's because 1920x1080 screen has 1.4 times more pixels that need rendering in comparison to the 1600x900 screen - your games will fly! I'm pleased I have the lower res screen, for the same reason (even if the quality is not as good as the 1080 screen).

Yes, if your temperatures are really high during the heaven benchmark (at stock settings), then it could well indicate that the pasting & padding of the heatsink may need to be improved. It's important to get the heatsink padded up accurately so that the heatsink is able to sit nice & flat on the GPU core (making good & full contact with it). This was the hardest & most time consuming part of the process for me, to make sure all the pads on the heat sink were in the right place with the right height so that they make contact with all the chips on the card, while still allowing the heatsink to sit flat against the actual GPU core. I think that's the most crucial part of the installation process, alongside making sure that the GPU core is pasted up properly. I used AS5, and the spread method on the GPU core. Did you have the 460M/560M installed before? If so, that's the same heatsink as mine, so I should think that the X-bracket from your 460M/560M combined with your 460M/560M heatsink should work fine for the 780M (after moving the pads around, and adding/subtracting in some areas). I think that because I have the same heatsink & X-bracket, and my GPU core is GK104, which is the same physical chip as the 780M chip, so the physical parameters of the chip should be the same as your 780M in terms of contact height etc. (It's important to use the correct X-bracket, because the height of the posts are different, which helps determine the contact point of the heatsink to the GPU core. (So, I think you'd be OK in terms of that coming from a 460M, based on my experience with that particular setup).

Link to comment
Share on other sites

Haha, 1600x900 with a 780M, that will be some awesome framerates you'll see! That's because 1920x1080 screen has 1.4 times more pixels that need rendering in comparison to the 1600x900 screen - your games will fly! I'm pleased I have the lower res screen, for the same reason (even if the quality is not as good as the 1080 screen).

Yes, if your temperatures are really high during the heaven benchmark (at stock settings), then it could well indicate that the pasting & padding of the heatsink may need to be improved. It's important to get the heatsink padded up accurately so that the heatsink is able to sit nice & flat on the GPU core (making good & full contact with it). This was the hardest & most time consuming part of the process for me, to make sure all the pads on the heat sink were in the right place with the right height so that they make contact with all the chips on the card, while still allowing the heatsink to sit flat against the actual GPU core. I think that's the most crucial part of the installation process, alongside making sure that the GPU core is pasted up properly. I used AS5, and the spread method on the GPU core. Did you have the 460M/560M installed before? If so, that's the same heatsink as mine, so I should think that the X-bracket from your 460M/560M combined with your 460M/560M heatsink should work fine for the 780M (after moving the pads around, and adding/subtracting in some areas). I think that because I have the same heatsink & X-bracket, and my GPU core is GK104, which is the same physical chip as the 780M chip, so the physical parameters of the chip should be the same as your 780M in terms of contact height etc. (It's important to use the correct X-bracket, because the height of the posts are different, which helps determine the contact point of the heatsink to the GPU core. (So, I think you'd be OK in terms of that coming from a 460M, based on my experience with that particular setup).

Yea, I intentionally got the 1600x900 screen for that reason. I had the 6990m before, lol. Check my sig. I got the x-bracket with the card and I also had a big sheet of fuji poly thermal pad/ putty that I used on my 6990m. I had some remaining so I used those to pad my 780m.

I spent a LOOONG while cutting and matching the pads over the chips and letting everything line up with the heatsink. I also got a 2nd hand 100 watt heatsink on eBay. Took me hours to install, because the pipes were slightly bent. I was never so stressed out in my life! Imagine that the card was lining up with the PCI slot and because the pipe is bent, the grills were not lining up with the vent slot in the laptop chassis. vice versa. OMG! i literally had to screw up, test, unscrew, massage the pipes approximately into place, screw up, test, unscrew play with the pipes................. about 10 times!!! *SIGH*

when i screwed up, to know if the pads sit flush with the chips and the back plate for the heatsink; they stuck to the plate. If they didnt stick, i knew it was too thin and added more to suit. Then took them off the heatsink back plate and put the back on the card and screwed back up and if they stuck when i pulled off the heatsink, i knew it lined up. I did this over and over and over and over till ALL the pads stuck to the back plate EXACTLY. After all this testing and matching headache, I realised that it wasnt fitting in my computer because of the bent pipe. OMG!!!!! THAT WAS A LOOOOONG NIGHT!

If i have to take out the gpu again to repaste I will, once i ABSOLUTELY have to. However, I really am not looking forward to it.....

Link to comment
Share on other sites

Yea, I intentionally got the 1600x900 screen for that reason. I had the 6990m before, lol. Check my sig. I got the x-bracket with the card and I also had a big sheet of fuji poly thermal pad/ putty that I used on my 6990m. I had some remaining so I used those to pad my 780m.

I spent a LOOONG while cutting and matching the pads over the chips and letting everything line up with the heatsink. I also got a 2nd hand 100 watt heatsink on eBay. Took me hours to install, because the pipes were slightly bent. I was never so stressed out in my life! Imagine that the card was lining up with the PCI slot and because the pipe is bent, the grills were not lining up with the vent slot in the laptop chassis. vice versa. OMG! i literally had to screw up, test, unscrew, massage the pipes approximately into place, screw up, test, unscrew play with the pipes................. about 10 times!!! *SIGH*

when i screwed up, to know if the pads sit flush with the chips and the back plate for the heatsink; they stuck to the plate. If they didnt stick, i knew it was too thin and added more to suit. Then took them off the heatsink back plate and put the back on the card and screwed back up and if they stuck when i pulled off the heatsink, i knew it lined up. I did this over and over and over and over till ALL the pads stuck to the back plate EXACTLY. After all this testing and matching headache, I realised that it wasnt fitting in my computer because of the bent pipe. OMG!!!!! THAT WAS A LOOOOONG NIGHT!

If i have to take out the gpu again to repaste I will, once i ABSOLUTELY have to. However, I really am not looking forward to it.....

Did you run the 10 minute Heaven Benchmark test, this will show you if your temperatures are ok? Then you'll know if you've got a problem that needs addressing.

Well it sounds like you took your time padding up the heatsink, and sounds like you did a good job with that. What a pain you had with the bent heat sink pipes too! When the whole heatsink & card assembly is installed are any of the parts under stress - like maybe you had to force it into place in your laptop to fit due to the bent pipes. If the heatsink is under stress (forced into place) when installed in the laptop is it possible that this stress might be causing a force on the card to lift the heatsink away from the GPU core? That may effect cooling.

Is it worth reusing your X-bracket from your 460M? Is it possible that this matches with your heatsink better, I'm not sure if X-brackets are matched to particular GPUs or particular heatsink designs or both? Is it worth trying your old 460M heatsink? (I'm not sure there's a difference between M17xR3 heatsinks in terms of cooling capacity - mine has 3 heat pipes and that was installed on my 560M). Or maybe it's just not pasted up optimally? If you dissemble your heatsink from your chip you should see the paste evenly distributed over the whole of the GPU core. Also, if the layer of paste on the GPU looks quite thick when removing the heatsink, then it could prove that the heatsink is not sitting close enough to the GPU core (affected by X-bracket). But really, you don't need to do any of this stuff if your temps are ok when checked at 100% GPU usage for an extended period, so check with the Heaven Benchmark & post back with your temperatures, ran on a loop for at least 10 mins.

Link to comment
Share on other sites

Did you run the 10 minute Heaven Benchmark test, this will show you if your temperatures are ok? Then you'll know if you've got a problem that needs addressing.

Well it sounds like you took your time padding up the heatsink, and sounds like you did a good job with that. What a pain you had with the bent heat sink pipes too! When the whole heatsink & card assembly is installed are any of the parts under stress - like maybe you had to force it into place in your laptop to fit due to the bent pipes. If the heatsink is under stress (forced into place) when installed in the laptop is it possible that this stress might be causing a force on the card to lift the heatsink away from the GPU core? That may effect cooling.

Is it worth reusing your X-bracket from your 460M? Is it possible that this matches with your heatsink better, I'm not sure if X-brackets are matched to particular GPUs or particular heatsink designs or both? Is it worth trying your old 460M heatsink? (I'm not sure there's a difference between M17xR3 heatsinks in terms of cooling capacity - mine has 3 heat pipes and that was installed on my 560M). Or maybe it's just not pasted up optimally? If you dissemble your heatsink from your chip you should see the paste evenly distributed over the whole of the GPU core. Also, if the layer of paste on the GPU looks quite thick when removing the heatsink, then it could prove that the heatsink is not sitting close enough to the GPU core (affected by X-bracket). But really, you don't need to do any of this stuff if your temps are ok when checked at 100% GPU usage for an extended period, so check with the Heaven Benchmark & post back with your temperatures, ran on a loop for at least 10 mins.

I had a reallllly long night last night........ :72:

I run the test on stock. I was getting 70+C... i went against my better judgement and removed the card to do a repaste. I installed it. green screen flashes... *SIGH* i tried everything. i removed the card and cleaned the inserts and air blown the pci slot... about 5-8 times. i even flashed bios back to stock, then back to svl7. i got it running. I run the benchmark and it was still glitching.

I even run it stock, overclock at stock voltage and over-volt +0.0125mV and overclock... it still glitched with the green lines.... sigh... temps were good however- 70C max on the overvolt and overclock. the others were 60'sC. I pasted it good. the paste job i did before was a mess..... :upset:

I turned on hitman absolution. screen went dead! :06::33_002:

i force powered off and powered on.... only back light..... not video feed.... :72::23_002:

I tried re-seating it countless times.. nothing...

I looked at the card and noticed that they were indents in the card where the restraints are to hold the card in place... near the sides.... right in front the PCI slot..... like the holder actually scrapped about a millimeter off the surface of the card... REALLY small! :62: can this render the card useless???

i put back in my 6990m and it worked flawlessly, so i know its the card. SIGH! i don't know what to do...

I couldn't sleep last night.... can you advise anything? sigh! I'm really regretting buying that card now..:too_sad:

EDIT: If it is that the card is really destroyed and 100% non-repairable. I strongly believe that I will give up gaming as a whole....

Link to comment
Share on other sites

I had a reallllly long night last night........ :72:

I run the test on stock. I was getting 70+C... i went against my better judgement and removed the card to do a repaste. I installed it. green screen flashes... *SIGH* i tried everything. i removed the card and cleaned the inserts and air blown the pci slot... about 5-8 times. i even flashed bios back to stock, then back to svl7. i got it running. I run the benchmark and it was still glitching.

I even run it stock, overclock at stock voltage and over-volt +0.0125mV and overclock... it still glitched with the green lines.... sigh... temps were good however- 70C max on the overvolt and overclock. the others were 60'sC. I pasted it good. the paste job i did before was a mess..... :upset:

I turned on hitman absolution. screen went dead! :06::33_002:

i force powered off and powered on.... only back light..... not video feed.... :72::23_002:

I tried re-seating it countless times.. nothing...

I looked at the card and noticed that they were indents in the card where the restraints are to hold the card in place... near the sides.... right in front the PCI slot..... like the holder actually scrapped about a millimeter off the surface of the card... REALLY small! :62: can this render the card useless???

i put back in my 6990m and it worked flawlessly, so i know its the card. SIGH! i don't know what to do...

I couldn't sleep last night.... can you advise anything? sigh! I'm really regretting buying that card now..:too_sad:

EDIT: If it is that the card is really destroyed and 100% non-repairable. I strongly believe that I will give up gaming as a whole....

Oh man, I'm really sorry to hear all of that! 70 degC at stock on Heaven Benchmark would have been OK I think, although you did find out the paste job was a bit of a mess from before and you got better temperatures by repasting. I'm really sorry to hear your card is glitching, I truly am, I know how much effort & resources can go into something like that. First thing I would say is don't do anymore work on it until you've kind of calmed down & got some sleep, because working on something detailed like that just doesn't really end well when you're knackered.

After doing a quick google search on 'Green Lines' and 'GPU', it seems that there are quite a few references to potential GPU failure, specifically the RAM of the card. What does Device Manager in Windows say about your GPU, is it flagging up any faults there? I think you did the right thing by trying to re-seat the card a few times & to blow the air out of the PCI slot, and cleaning the contacts of the card. The green lines could point to RAM issues on the card, I'm wondering if one of your RAM chips on the card might be getting really hot after your repaste. Perhaps if the heatsink is not sitting flat on all the RAM chips, then one of them might be getting really hot & temporarily glitching as it gets hotter. There's always a risk of accidentally damaging hardware when you do these things yourself (& also if you let Dell Technicians do them it seems!), I'm pretty sure I fried my system RAM once by vacuum cleaning my laptop - STATIC! Always worth taking static precautions when doing electrical work on your laptop/PC, wearing a anti-static wrist strap or grounding yourself by touching some bare metal before touching internal components. Once again, really sorry to hear about your problems with the card, and let us know what you decide to do.

EDIT: not sure what you mean by the scrapes on your card, I couldn't picture what you were describing - a photo?

EDIT 2: wonder if it would be worth uinstalling NVidia video drivers & then doing a clean NVidia driver reinstall, just as a quick check that there are no software driver problems causing the issue - quick & easy to check. (Especially as you talk about having reinstalled the 6990M temporarily, so I'd check to make sure you've uninstalled your AMD graphics drivers too).

Link to comment
Share on other sites

I had a reallllly long night last night........ :72:

I run the test on stock. I was getting 70+C... i went against my better judgement and removed the card to do a repaste. I installed it. green screen flashes... *SIGH* i tried everything. i removed the card and cleaned the inserts and air blown the pci slot... about 5-8 times. i even flashed bios back to stock, then back to svl7. i got it running. I run the benchmark and it was still glitching.

I even run it stock, overclock at stock voltage and over-volt +0.0125mV and overclock... it still glitched with the green lines.... sigh... temps were good however- 70C max on the overvolt and overclock. the others were 60'sC. I pasted it good. the paste job i did before was a mess..... :upset:

I turned on hitman absolution. screen went dead! :06::33_002:

i force powered off and powered on.... only back light..... not video feed.... :72::23_002:

I tried re-seating it countless times.. nothing...

I looked at the card and noticed that they were indents in the card where the restraints are to hold the card in place... near the sides.... right in front the PCI slot..... like the holder actually scrapped about a millimeter off the surface of the card... REALLY small! :62: can this render the card useless???

i put back in my 6990m and it worked flawlessly, so i know its the card. SIGH! i don't know what to do...

I couldn't sleep last night.... can you advise anything? sigh! I'm really regretting buying that card now..:too_sad:

EDIT: If it is that the card is really destroyed and 100% non-repairable. I strongly believe that I will give up gaming as a whole....

Did you do anything to increase the heatsink die pressure? If so you may have damaged the GPU BGA. It is possible to fix a damaged BGA by baking the whole card in the toaster oven.

Link to comment
Share on other sites

Did you do anything to increase the heatsink die pressure? If so you may have damaged the GPU BGA. It is possible to fix a damaged BGA by baking the whole card in the toaster oven.

I think that's a good point, but I'm thinking he should probably do that as a last resort because baking could be risky too. Do you think it could be the heatsink not sitting flat on all the VRAM chips?

Link to comment
Share on other sites

EDIT: not sure what you mean by the scrapes on your card, I couldn't picture what you were describing - a photo?

I cant take any images now, but if you remember, when you insert the card at an angle, into the pci slot... ALL the way in the slot. after, you have to rotate the card into place on top of the motherboard right. now, there are 2 pins. one on either side of the gpu. The gpu has slots to accommodate these pins. the pins are meant for keeping the card properly fastened and keeping it inside the pci slot. on hindsight, i remember applying more pressure to the screws when i screwed down the heatsink. I'm in work and all i can do is think about that are the possibilities for solutions or problems i may have done accidentally.

Link to comment
Share on other sites

I cant take any images now, but if you remember, when you insert the card at an angle, into the pci slot... ALL the way in the slot. after, you have to rotate the card into place on top of the motherboard right. now, there are 2 pins. one on either side of the gpu. The gpu has slots to accommodate these pins. the pins are meant for keeping the card properly fastened and keeping it inside the pci slot. on hindsight, i remember applying more pressure to the screws when i screwed down the heatsink. I'm in work and all i can do is think about that are the possibilities for solutions or problems i may have done accidentally.

Ah, ok, so I can visualise the pins you're talking about. So the pins have cut a 1mm deep grove somewhere on the bottom surface of the card while you were installing it one time? 1mm is pretty deep, if that's cut through one of the circuits? I guess it depends where the 1mm deep groove is located, and if there are circuits in the area.

EDIT: when screwing the card down to the motherboard, I don't think that's a particularly sensitive area, so the extra force you used would be OK I imagine, as long as the card was located into the slot properly as you were screwing it down. (Are you using some reasonable quality electronics type screwdrivers, because I've found using those that you don't need to press down as hard to screw it down firmly? These are the ones I use: http://www.amazon.co.uk/gp/product/B000NRU88E/ref=oh_details_o08_s00_i00?ie=UTF8&psc=1

Link to comment
Share on other sites

Ah, ok, so I can visualise the pins you're talking about. So the pins have cut a 1mm deep grove somewhere on the bottom surface of the card while you were installing it one time? 1mm is pretty deep, if that's cut through one of the circuits? I guess it depends where the 1mm deep groove is located, and if there are circuits in the area.

EDIT: when screwing the card down to the motherboard, I don't think that's a particularly sensitive area, so the extra force you used would be OK I imagine, as long as the card was located into the slot properly as you were screwing it down. (Are you using some reasonable quality electronics type screwdrivers, because I've found using those that you don't need to press down as hard to screw it down firmly? These are the ones I use: Hama Mini Screwdriver Kit: Amazon.co.uk: Computers & Accessories

not screwing down the card to the motherboard. i meant screwing down the heatsink to the video card. *SIGH* :30_002:

i swear this computer will give me a heart attack... i just feeling to run outside in the street and scream at the top of my voice, LOL! wow! right now, im just looking busy in work. i swear i cant think straight :45:

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...

Important Information

By using this site, you agree to our Terms of Use. We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue.