G73JH Repaste Collective Summary

Here's the results after a repaste.

Before:


Immediately After


And Two weeks later.


C02 card revision, vBios 9.3, 11.1a drivers ...

Can't say I'm impressed, but I'm happy.

My paste looked to be fairly well applied ... over 3/4ths of the chip. Since I eventually had to go through Gary/Mason/Chastity to get my RMA done properly, I expect that someone did take their time to make sure it was done right.

I did a long re-read of the g73 repaste-related threads, and although there are some low temps reported, they're not the norm - generally speaking, people end up with 85-90 core temp (tss0), and ~5-10 more on the memIO temp (tss1). Shader temps (tss2) are almost always within 2 degrees of the core, and most people idle between 55-60.

Here's a few links of other other post-repaste temps. (1), (2a), (2b), (3), (4), (5), (6), (7), (8), (9), (10).

These generalizations are also generally for ICD, though there were a few people who used other types of paste (mx-2, for example) that ended up with similar temps. Since most thermal pastes are within 2 degrees of each other when applied properly, it's application method and not material that counts here.

NOTE: You'll notice Chastity has crazy low temps. You'll also note she's got a new card revision (How do I know? She needed a vbios update to fix her GSODs). Also, Quadzilla has crazy low temps. Ken Lee has crazy low temps. Something else to note? Quad and Ken are both quite experienced when it comes to repasting. Personally, I look at all those examples I linked, and think : "Hmm, is another repaste worth trying to get my card down another 2-3 degrees, when I'm not even sure that I've got a GPU that can run that cool?". Also, I was hoping for a crazy low temp because there are a few vocal people who have crazy low temps. But they're the vocal minority, I believe. (A very helpful minority, I might add).

Anyways, I believe the take home message is ATI cards run hot. It's always been the case. As chastity said - if your temps are under 98 degrees for ALL of your sensors (which are all on the GPU die), you're fine.

Edit: More people with high temps while gaming (higher with furmark, I'm sure): (1), (2), (3), (4), (5), (more as I get around to updating...)

 

Temps would be much better, if Asus would have let the vBIOS open for changing voltage (undervolting). Like MSI.
The Asus-5870 also isn't running 700/[email protected].

 

There are a variety of end user custom vBios available, or you can roll your own with the radeon bios editor. I've chosen to stick with the one that asus is using on their new units, because ... for some reason, people are generally happy to change things they can touch and control (like disassembly), but don't want to risk something they can't physically alter (like the vbios).

But yes, you'd get an additional drop if you set it to 0.95, which you can do. You just have to roll your own bios. Not hard, honestly.

 

I modded such 5870m-vBIOS since mid last year.

But my G73 won't run with this setting. I have tested it.
Maybe 1.00V would do the trick, but... (damn proprietary Asus)
My X7811-5870m (MSI GX740 barebone) runs with 0.95V (stable).

 

I havn't bothered doing it either, but I've read a bunch of reports that others have set their gpu to .95v using RBE - again, I've read that it's hardwired into the card to have only two options.

 

[edit for later]

 

Hi there DCx. I ran the dreaded furry donut test just for you. I didn't want to run it after my RMA. They replaced my motherboard out there in Cali so they obviously had to repaste my GPU. My ICD7 is gone now I guess. It's been idling and gaming with temperatures as good or better than when it left me. If it was bad, I guess I didn't want to know because I would be tearing the whole thing down again. Anyway, my fear was all for naught. It seems to be just fine. I assumed your tests were 1280X1024 with Extreme Burn Mode, Post FX, and 4X MSAA so that's what I am running as I type this.

I forgot to download AMD Clock Tool before I started and the version I downloaded while the test was running reported that it didn't find a compatible card. I grabbed GPUz instead. It's been running almost exactly fifteen minutes and it is bouncing between 77 and 78C. GPUz reports the three temperatures but doesn't name them... Temp1 is 78C, Temp2 is 81.5C, and Temp3 is 78C. It doesn't look like it's going climb anymore. It's freaking cold in my house right now which is very odd for Texas. I would say the temperature is about 69F or so. My CPU temps look good I think. The one core that is maxed out for this test is at 60C. The other cores are reporting 53C.

It's been running for twenty minutes by the time I finish posting and the temperatures haven't climbed at all. 78.0, 81.5, and 78.0. I guess the California RMA facility used darn good paste and they do an excellent job of applying it.

Edit: I run Chastity's modified vBios and my idle temp was bouncing between 49C and 50C all night before I ran the furry donut test. It idles at 405/1000 @.95V. When I run the test it goes to 700/1000 @ 1.15V. P4G High Performance setting plugged in, of course.

 

Yesterday i did it too. My book was always cool but now it freezes Got idle temps of 43 - 48 and under stress 68 C max!! These temps are real, i will make screenshots. Before the repasteing i had 46 - 51 @ idle and max 74 C @ performance. I used IC7 paste and spend a bit more than tow hours on it.

Now i can take the 209er Bios
VBios was my own.

Edit: Roomtemp while i was testig was 21,8 C

 

I've done some tests too. The G73 was placed on a Zalman NC-3000U.
My stock tim and paste job wasn't bad. But i repasted with MX-3.
Downgraded to BIOS 209.

Test with Furmark (1280X1024, Extreme Burn Mode, Post FX, 4X MSAA):

After 15 minutes (NC3000 off):


Now turned on NC3000.
After ~24 minutes (NC3000 on Fullspeed):


Quit the Donut.
3 Minutes cooldown (CPU + GPU fan off!):


Now the NC3000 is set to lowest speed and there are no fans on in the G73 for almost 5:30 min!
If the gpu hits 47°C, the fan starts and runs for only 20 seconds. Nice.
Ambient temp. ~21°C

Interesting is also the gpu fan rpm while stressing. You can see, it is just 3700rpm. Fullspeed is at 4500-4600!

 

Same here

This is why i love and bougth this notebook, the potential!!

A little bit (lol a little bit XDD ) tuning and its perfect.

flash the HDD, the bios, the vbios and repaste the chips.....

 

What did you paste with? I couldn't find what you used in your post.

 

I sent them ICD to use.

 

Dammm you guys are making me get the repaste itch lol =/ (my computer just froze while typing that sentence, not a BSOD, simply locked up...odd thats never happened before, had to hold the power button down to get out of it)

@Jody, lucky you, they stole my ICD7, didnt use it to repaste my machine, and my temps are still sitting at 94C in furmark....perhaps they used my ICD7 on your machine...

 

LOL! Thank you for donating a tube of ICD7 to the cause!

 

Maybe you should try lapping your GPU heatsink.
I also read somewhere that adding a minute coating of ICD7 to the heatsink plate may help insure an even spread with no air pockets.
But this is just what I read.

 

94 is good, if it's memIO, that'd mean your core is 5-10 lower, right?

 

Im actually watching a video and its sooooooo quiet!! Really awsome.

I took much of the paste and i connected the pipes with it.

 

Hi guys, can't find IC Diamond thermal paste here at Germany. Do you guys think Arctic Silver 5 will do the job?

* Auto shutdown twice today while I'm playing FIFA. Repaste time.

 

AS5 should do the job too.
I've used AC MX-3. Those tims have pretty much the same result.

Gruß

 

Correct

I understand that 94C is pretty good for the stock paste, I was just a little sour that I was without my laptop for a couple of weeks while it was in RMA, paid for the ICD, and didnt get its benefits or have it sent back to me, you know?

Granted others have 5870's that run much hotter, and have had far worse RMA experiences so I guess in that sense it wasnt terrible.

@Jody; oh anytime lol

 

I've got ICD7 and that's my temps too, fwiw.

It's likely because we've got earlier revision cards (c02 for me, the other option was c07 or c09 ... the c12/14 cards were GSODing badly). Earlier revisions run hotter ... but hold a much better overclock. I think we're the only ones that can run 870 / 1150 stable here. You can check by opening your bottom panel up and looking for the green-ish sticker.

Still, with my multitudes of RMAs, I wish I'd have ended up with a later revision of card.

Anyways, I made this thread for people like you and me, who have hot cards (or badly deformed heatsinks, or whatever the reason), because there are about 10 links of other people reporting about the same temps as us - but they're not regulars and only posted once or twice.

People like Chas or Ken who are regulars here make us all feel bad by having cool running chips and great paste jobs. I know I've got a great paste job, and a relatively flush heat sink (I lapped it, though I didn't mention that) and I still get WAY hotter than I think I should. But, I guess that's just how it goes.

TBH at this point, now that I've done all I can, I'm happy. I can't "fix" it because it's not broken.

Besides, I am now gaming at a temp that is really near my old idle temps (dead space 2 at 64/67/65, used to idle at 62/65/63). I know it's not particularly hard on the gpu, but DS1 prior to my repaste was giving me 75-80, and DS2 was 80ish. Also, my temps are quite sensitive to my ambient temps - 2 degree change in ambient gives a 2-5 degree change in furmark at load.

 

Ya I saw your 93C or so after* paste temps

Your probably right, and Ill admit reading this has made me feel a bit better about it lol

DS2 gives me low 70's, SCII gives me low to mid 70's (rarely low 80's), L4D2 gives high 70's, AOE and SupCom also give me low temps (similar to DS2), its only furmark that I ever hit 80/90s (to be expected), so I know I shouldnt have anything to worry about, but like you I strive for perfection/greatest efficiency/temps ect

Ambient has a great affect on mine too, im coming up on 7 months here with little cleaning so ill be sure to open it up soon to blow out the vents, while im at it ill check my card revision, thanks DCx

 

For the first time in the 8 or so months that I've had my G73JH it simply shut down due to what I think was GPU temperature being too high.

I just started playing Crysis recently and a few days ago the CPU just shut off in the middle of the game. I figured it was just a one time thing because I've been playing Crysis several hours in a row the past couple days and it didn't happen again, but then I went to play CSS today which should be much less intensive and surprisingly it happened again, CPU just shut down without warning. Then every time I've played CSS since then after about 20-30 minutes it just shuts down. This has happened like 4 times in a row now today.

So because of this I DLed CPUID to check temps (I never did in the past) and I ran CSS and spectated a game to see how high it got and it got up to 99C in like less than 10 minutes! The game was getting like 300 FPS sometimes though! That has never happened before either, what the hell is going on? I thought games were supposed to get slower as GPUs got hotter. No sign of slowdown at all, it just crashes. I'm not a computer expert at all but I definitely need to fix this, so what's the best thing I can do? I can't have CSS crashing my G73 lol... btw my temp. while typing this up is 74, seems way too high too.

 

Is the ICD7 the best thermal paste in the market? Has anyone used the shin-etsu paste on their gpu? I've been hearing good comments about it.

 

You should know the routine:

1) Blow out your vents with compressed air, check readings.
2) If still too high, your paste has degraded, and needs to be redone.

 

I havn't used it, but I've read good things too. The question though, is "is it worth it?" If you've got some paste at home already, use that. The 1-2 degree difference between amazing paste and good paste isn't worth the hassle of changing it. From what I've read, ICD7 and SHin Etsu and MX-4 are all in the same ballpark.

 

The only thing I recommend for the repaste is to avoid using the really thin TIMs. Try to use something that is viscous, since the heatsink is not perfect.

 

ok I figured that was the normal thing to do, but it seems like my gpu is getting a sort of allergic reaction now as well. Way overpowering for games that don't even need it. I still only have version 10.1 of catalyst control center and everything else is just stock, don't know why the sudden "allergic reaction". Just did a stress test on CSS and got 215 FPS with everything very high, no AA, way higher than it used to be.

 

Maybe it's time you updated your vBIOS and drivers

 

Thanks for the suggestions. Nice threads in your sig btw.

Wow I just updated graphics drivers in device manager and updated catalyst to 11.2 and rebooted (not sure if these are specifically the reason at all) but now my temp while typing this is 57 and CSS just maxed out at 87 and averaged like low 80s all while getting 150-300 fps. don't know what the hell happened but it's been mitigated drastically and I didn't even update vBIOS or BIOS (still have old versions), spray an air canister, let alone repaste. Can't wait to update BIOS to 211, from what I read it supposedly blows the fan harder for lower temps?

 

Don't stop now, finish all the updates....

 

You really should get the new vBios. As in, it's the single most important update you can do to your g73. Everything else is pointless if you don't update the vBios.

 

What program are you using to monitor your temps? If you are referring to the temp-drop in the MSI afterburner, its cause the new CCC shows the DispIO in Afterburner (which is always lower than the MemIO) instead of the MemIO temperature. I strongly suggest you update your vbios, as it will save your graphics card on the long term.

 

I'm using the CPUID hardware monitor. After using a duster on the vents it went down even further from 58ish degrees to 51-52 while watching youtube videos, but then I just updated the driver again in device manager and it's jumped back up to 65ish now! I have no idea what the hell is going on. I wish I could undo that update somehow. And yes I still have to update vBIOS, it's definitely outdated.

Wow, so I realized CCC was crashing every time I tried opening it so I reinstalled it and then temps dropped back down to 50ish again, thankfully. Now I just updated vBIOS to Chastity's version (the first one, only realized there was a second while updating, is it that much of a difference?) and idle temps are at 47-48! I'm a little too afraid to mess with the GPU clock settings though. Now about to update BIOS. Thanks everyone.

 

It's not just different revisions that may have different temps. When a piece of silicon is fabricated, each is unique like a fingerprint. They all have different electrical and thermal properties. They get tested and if they fit within specs they get used, otherwise they get discarded.

There will be some GPUs on the bottom end of the power usage and some on the hot end and some in the middle. Who knows what the variance is but it could be significant. Mobile chips are probably already the better part of the batch though.

I'm also sure that each of the cooling assemblies varies as well. So you could get a hot GPU with a less efficient heatsink and be much hotter than somebody else.

The paste thing is only going to make a noticeable difference if they didn't cover the whole GPU die and that is so incompetent as to be highly unlikely. Or if when you reassemble it you manage to squish those pesky thermal pads down more because as you can see in the first photo of this thread they push up against the heatsink and reduce contact on that side of the GPU die.

 

AMEN. 10char.

 

So many variables that makes our temps the way they are. cleaning the dust out of the fan vents will give you about 8* drop if your temps have crept up in the last few months and you haven't cleaned yet. When I opened it up there was a lot of dust in the vent fins. Had to take a qtip to scrape it all off. lot's of dust on the fan blades too. Not sure if compressed air blowing from the back will cut it. May have to take it apart if you want to do a good cleaning. CPU fan was hardly dirty.


I say under 90* is good. when I game I hit about 85* max and idle about 58*
that's with ICD. I don't prop up my laptop and my rubber feet are no longer attached so I'm probably about 5* hotter than it would be with good airflow.

 

For gaming, under 90 is about where you want to be. I get around 85, as well for Crysis 2. Something in the 82-86 range.

 

ok so I updated to BIOS 213 and now the keyboard lag is 100% gone, couldn't be happier about that. My touchbad disable function still doesn't work though for some reason, which bothers me a bit.

I just have a couple questions about overclocking with Chastity's vBIOS. How much overclocking is safe and how much isn't? Does it depend on gpu temps? After all the updates the highest my gpu has gotten is 77 degrees while playing Crysis Warhead for 1-2 hours.

 

Overclocking is highly dependant on temps. Under 100 is probably a good target.

 

ok then if my gpu's temps are maxing out at 77 degrees with one of the most demanding games at 700/1000, what is a safe number to overclock to?

 

Most can do 800/1100 pretty easily, Id start there and slowly work your way up until you either dont want to go any further, become concerned with your temps, or your computer crashes lol

 

I don't really know how many can do 800 core. I've had a couple of G73s and one could but one can barely manage 750 without artifacts.

 

^^

Well it was just based off what Ive seen in this forum. Perhaps I spoke too soon.

I will say however that AW and Sager/Clevo certainly seem have higher avg OC's *shrugs*

 

What do you mean "flash the HDD"? You mean format and rebuild?

 

No, the new hybrid hard drives actually need firmware updates.

 

Very true 800/1100 is the most common stable O/c I have seen from people on this forum it also works very well with most games (never crashed for me) although before I vbios'd and drivers updated 1000 was my max.

I can push 840/1140 but some games just dont like it.

Repaste FTW cant believe I was working with normal gaming temps of 102oC at 700/1000 and now have 800/1100 88oC at extreme burn.