Jump to content

P370EM, 680m SLI, Possible Temp Issue


lm1z

Recommended Posts

Hey all,

Just started posting on the forum. Lots of useful information here! Thanks for the responses thus far.

OK, so here's what's going on.

Flashed both cards with slv7's Clevo 33.10 non-OV bios.

Using Precision X to monitor and OC.

Everything was going well. +220 on core, +600 on memory.

Temps playing BF3 around 80c.

I could bench 3dMark11 showing no signs of throttling. 12.5k score. Vantage score of 35k.

This is with my 3740qm at 41x 40x 39x 39x and my memory at 9-9-9-24.

So, I decide I want to OV a bit.

I go with the 1.050 version of 33.10 and I notice that my bench numbers dropped. Now, this is probably a fault on my end, as I wasn't logging temps with Precision X during this test.

However, I switched to Nvidia Inspector, and my recorded max-temp, albeit I need to make sure that this was my primary GPU, was 78c.

Note that my bench numbers dropped with the SAME clocks, which leads me to believe that the card is throttling due to temperature. I dropped to an 11.9k on 3DMark11 during this period.

I try some other versions of the VBIOS. The 1.000 and the 1.025. No dice. Around the same scores.

I notice that the performance IN-GAME is not affected. Everything runs great with OV and higher clocks. I was able to run 1050/2400mhz in game with no issues and increased performance. Temps no higher than 80-82c in Precision X (primary card). So, even though I can't bench well, I decide let's go for it since my in-game experience is improved and I show no signs of detriment to the system.

Today, I try to give Crysis 3 a try.

Oh, no.

On max settings, I'm recording temps of 90-94c. And it's perfectly clear that the card throttles itself when it reaches 90-92c.

Runs like butter during less intense areas, but as soon as the card has to bump in performance, temps spike and card throttles.

I immediately know that this is exactly what's happening during benching, and Crysis 3 is a very intense game.

I decide to mess around with the VBIOS again to decrease temps.

I move down from the 1.050 to the 1.025 to the 1.000. No dice. Temps still hit 90c with my +220/600 settings.

THEN, I start detuning to see if I can decrease temps. Again, no dice. How?

I end up going BACK to the non-OV bios and this is what concerns me.

Crysis 3 at stock clocks is still hitting 90c on my computer...

Now, I don't think I've done any damage to the heatsinks. My card would be at much higher temps at idle if that were the case.

What concerns me is that this game is so intense it's causing my primary card to overheat and throttle.

I'd love for an experienced vet to help me out here.

Now, if running Crysis 3 at stock clocks (on very high) can cause temps to hit this, so be it, but I can't seem to think this is the case.

Hope this all made sense.

Any help would be appreciated.

L

Link to comment
Share on other sites

How hot is the slave card? Usually that runs hotter than the primary on the P370.

Well, my labeling might be wrong.

The identities of the cards are 0 and 1.

If you take this into account, I assume 0 is the primary and 1 would be the slave. #1 is running hotter.

The slave would be the one I am experiencing the issue with, then, and it's running at a little over 90c on stock clocks/voltage in Crysis 3.

I'm going to go home and make sure everything's gravy when running BF3. If not, then it's possible the connections between the heat-sink and card might be an issue.

Link to comment
Share on other sites

Check it out.

GPU-Z must be monitoring the Primary card. It's at ~50c.

But my slave is at 91c! Stock voltage and clocks...

This is playing Crysis 3.

BF3 is just fine sitting under 80c.

Weird!

kdppc3.png

Link to comment
Share on other sites

Then, here's the temps while benching at +220/600 on stock voltage.

Clearly a temperature and throttling issue.

Kind of stuck on what to do. Looks like there isn't much room to OC on this P370 due to heat! I wonder If I can apply a better heatsink and/or thermal paste to the cards?

Even Crysis 3 throttles due to heat with STOCK clocks.

2n0uc60.png

Link to comment
Share on other sites

It looks like this may be a bug with precision. The graphic reports 91C while the graph says 45C. The P370EM is known for high temps anyway though.

As for cooling improvements, you could get 375sm heatsinks, which have bigger heatpipes and radiators for the GPUs, or you could work on die contact with your current heatsinks.

Link to comment
Share on other sites

It looks like this may be a bug with precision. The graphic reports 91C while the graph says 45C. The P370EM is known for high temps anyway though.

As for cooling improvements, you could get 375sm heatsinks, which have bigger heatpipes and radiators for the GPUs, or you could work on die contact with your current heatsinks.

GPU-Z is actually showing you GPU1's temp (Primary),vs. GPU 2 (slave). I verified that there is no bug with Precision. They match up when I have them aligned on the same card.

It just looks like Crysis at Ultra settings is very intense and the cooling on the Sagers, like people have mentioned, isn't very good :(

I logged 3DMark11 on stock temps and voltage and it maxes out at 84c.

Thanks for the replies, man.

I'd love additional members to chime in if they have any insight!

Link to comment
Share on other sites

Yeah unfortunately the slave card on the P370EM is always going to be an issue due to the design they went with. The slave card heatpipes for the sink are just too long and have to travel too far to be efficient. The main GPU is better but still not optimum. best you can do is re paste and make sure for a good contact, there are a few mods and tricks on here you can do. And if you want to go further lapping the sink surface also helps but the problem will always be there, just with less effect. ;)

Here is an e.g where you see the 2nd GPU and how far the pipes have to travel..

ju9vnc.jpg

Link to comment
Share on other sites

Yeah unfortunately the slave card on the P370EM is always going to be an issue due to the design they went with. The slave card heatpipes for the sink are just too long and have to travel too far to be efficient. The main GPU is better but still not optimum. best you can do is re paste and make sure for a good contact, there are a few mods and tricks on here you can do. And if you want to go further lapping the sink surface also helps but the problem will always be there, just with less effect. ;)

Here is an e.g where you see the 2nd GPU and how far the pipes have to travel..

ju9vnc.jpg

This is great info. Thank you! Kind of a bummer with the design. Maybe I can go with some 375EM heat-sinks. That could always help. I could always repaste, too!

Link to comment
Share on other sites

  • 1 month later...

I have the same configuration and i was having heat problems even without overclock due to the very high ambient temperature where i live.

One solution is removing the exhaust grilles, it's very easy, no cutting required at all. Or if you want it to look nicer you can do what i did in the pic.

Another trick is making a more open fan intake. With those 2 mods i can play crysis at about 85C without putting the fans at full speed. Before the mod i would reach 95C with normal fan speed and about 87C with full speed. With ambient temps at about 35C. It makes a few degrees difference

post-11025-14494995763933_thumb.jpg

post-11025-14494995763223_thumb.jpg

Link to comment
Share on other sites

I have the same configuration and i was having heat problems even without overclock due to the very high ambient temperature where i live.

One solution is removing the exhaust grilles, it's very easy, no cutting required at all. Or if you want it to look nicer you can do what i did in the pic.

Another trick is making a more open fan intake. With those 2 mods i can play crysis at about 85C without putting the fans at full speed. Before the mod i would reach 95C with normal fan speed and about 87C with full speed. With ambient temps at about 35C. It makes a few degrees difference

[ATTACH=CONFIG]8410[/ATTACH][ATTACH=CONFIG]8411[/ATTACH]

I did something similar on my p150EM. I used a small file that could fit into the individual slits and I made them just over 2x as large. It dropped temps probably 1-2C on the CPU and 3-4 on the GPU. GPU temps do not improve at all with taking the bottom plate off anymore, while CPU seems to improve by 1 degree.

Exhaust vent mods didn't seem to really do anything for me, but I think my exhaust vents were less restrictive than yours.

Link to comment
Share on other sites

  • 3 weeks later...

I have a P370em with two 670mx. I am getting 60 degrees primary card and 71 degree slave on furmark. This is with +135 +800 oc stock volts in ambient 25 degree room. Is this larger difference from the ops numbers just due to 670mx being much cooler? Or does it mean my heat sinks are ok?

Link to comment
Share on other sites

  • Moderator
I have a P370em with two 670mx. I am getting 60 degrees primary card and 71 degree slave on furmark. This is with +135 +800 oc stock volts in ambient 25 degree room. Is this larger difference from the ops numbers just due to 670mx being much cooler? Or does it mean my heat sinks are ok?

I would advice you NOT to use furmark. It's just bad for your GPU..

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.