zotac 970 can't oc?

[hunkeelin]

i got 2x in sli: http://www.newegg.com/Product/Produc...Q&gclsrc=aw.ds


I saw other 970 reviews claiming they could easily up to 1300mhz+ and with good models like the one from asus/msi they can get to 1500mhz.

I however, can't even OC that well. For my zotac 970 my power limit is 106% whilist the reviews online from asus/msi have a power limit of 110%. Moreover I can barely OC to 1300mhz before nvidia display driver keep crashing.

I don't think is the driver's problem. I have a clean install. install after + completely removing etc....

My question is
1) is my card that weak and not ocable?
2) is it because i'm running sli?
3) is it because my temps are around 79~80C? Wheredoes the other high end 970 run much lower. (in afterburner i simply set the temp limit to 90C). My temps never exceed 80C btw.


Thanks

 

[skipsneeky2]

Could be VRMS,for single moderate overclocking these are supposed to be alright.But you got 2 of them and the vrms aren't really cooled off all that well.

I will Edit this post with a link to a 67 page thread based solely off this Zotac 970 for you shortly

http://www.overclock.net/t/1514089/zotac-970/170 Check pictures on post #180

 

[guskline]

hunkeelin: First EXTREMELY impressive hardware. I have found that running SLI or CF generally increases the heat in each card especially the top one. OCing really causes heat issues. I looked at the Zotac card and here is what I see.
1. Small PCB so unlikely to have extras to assist with OCing. I have 3 fans you have 2.

2. 2 6 pin connectors. On my Gigabyte GTX970 G1 Gamer Ihave a 6 and 8 pin for extra power

3. Standard core at 1076 base. My base is 1178. Likely that it is set here due to smaller pcb and power supply.

To OC at higher levels in SLI you need to dissipate the heat and perhaps better cool the VRMs. I can OC my R9 290s well in CF BUT they have full EK waterblocks which keep both the GPU and the VRAM very cool.

Personally I would keep the GPUs stock in SLI.

 

[VirtualLarry]

I don't really recall Zotac ever being considered a "top" card for OCing, through various NV generations. They tend to be a minimalist, budget card brand.

 

[skipsneeky2]

I don't really recall Zotac ever being considered a "top" card for OCing, through various NV generations. They tend to be a minimalist, budget card brand.

Yeah my 770 really can't overclock much for anything,but the AMP! cooler is pretty impressive for stock clock operation.From what i can collect this Zotac 970 card is better off then any blower either Evga or PNY which both seem common enough and cheap but i doubt the card is mean't to be taken seriously as a overclocker but more of their answer to the lack of a quality reference card.

Zotac 970 reminds me of my old MSI TF3 7850,its short and cute and can stay cool but it will make a bit more noise then anything else aftermarket but a surprise or two may be up its sleeves.Still better then blower either way which is what you would get if you decide to be picky on brand at $329.

You get 3 choices for $329 off 2 sites,Evga ACX based 970 right off Evga,or blower style EVGA/PNY or your Zotac off Newegg.I am guessing Evga might not be offering their free game promotion off their site?Need to look more into this.

 

[hunkeelin]

but my temps are below 80C in hw mon..... at most 79C is it too high?

so... the question is, if I get a custom water loop on both my gpu and keep both gpu under 70C I will be able to OC to 1500mhz ?

Also, I don't think is vram, in gpuz the benchmark I am running (metro last light) never uses vram that exceed 60%.

 

[iiiankiii]

is 1300mhz boost speed or base speed? If it's boost speed, I would say that's on the low side. If it's base speed, that's actually a pretty good overclock. I can only do around 1250mhz on my evga.

 

[hunkeelin]

^ boost of course.

 

[Face2Face]

How are you overclocking? Are you increasing the mV and Power limit? Are you seeing your cards throttling?

From what I can gather, looking at Ryan's review, the GTX 970 hits a thermal throttle point at 80c.

Owing to its open air cooler, low TDP, and TDP throttling, the load temperature for the GTX 970 FTW tops out at 63C in all scenarios. Whether it’s gaming or FurMark, 100% TDP means the card finds its cooling/temperature equilibrium at nearly 20C less than its thermal throttling point

http://www.anandtech.com/show/8568/the-geforce-gtx-970-review-feat-evga/15

 

[hunkeelin]

^I use msi afterburner. I did increase the power limit to max. I mean c mon.. I can't even do a + 50mhz. I shouldn't need to touch voltage for a mere +50mhz.

 

[Face2Face]

^I use msi afterburner. I did increase the power limit to max. I mean c mon.. I can't even do a + 50mhz. I shouldn't need to touch voltage for a mere +50mhz.

Read above - Sorry ninja edit - Cards could be throttling if they're hitting 80c.

 

[hunkeelin]

^ the problem is, temps not over 80C and it throttles because the temp limit (if you hit the arrow beside power limit) is set that way. the card will not throttle if you set the temp limit high enough.

Throttling isn't the issue. If it throttles then my benchmark should just run with a lower score. The problem is I keep getting "nvidia display driver stopped responding and has recovered" which crashes my metro last light benchmark.

After reading my previous paragraph some might think is a software issue. I assure you it is not. When I simply set my clock to 1075mhz(the lowest possible), my metro bench mark ran just fine.

 

[hunkeelin]

Could be VRMS,for single moderate overclocking these are supposed to be alright.But you got 2 of them and the vrms aren't really cooled off all that well.

I will Edit this post with a link to a 67 page thread based solely off this Zotac 970 for you shortly

http://www.overclock.net/t/1514089/zotac-970/170 Check pictures on post #180

I never open up any gpu before but in the pictures which one of them are VRMs? If possible can you circle it for me? thanks!

 

[Face2Face]

I never open up any gpu before but in the pictures which one of them are VRMs? If possible can you circle it for me? thanks!

There is no reason to believe it's your card's VRM thats overheating. I would try adding some voltage to see what happens. Plus try Unigine Heaven and see if it still crashes.

You could just have picky cards?

 

[hunkeelin]

^ for realbench's openCL x10 runs I can oc my sli card up to 1460mhz np. Just somehow metro's benchmark is not liking it. Software issue or OC issue? btw my temps always below 80C. My unigine heaven 4.0 score is 2709

 

[cmdrdredd]

Probably just an issue with the cards not taking that OC well for that game. My MSI 970s can go to 1580Mhz before they crash like that. They get obscenely hot or just throttle when I do that though because I need to voltage up. So I settled on stock voltage and 1430Mhz. They would do more but I'm temperature limited because of my particular motherboard. It's an mATX board which places the cards very close together and starves the first card for air just a bit. The top card in my system will go to around 85c at a max load while the bottom card hovers around 63c. This happens because the hot air is rising to hit the first card and the top card's fans are blowing down right onto the back of the bottom card. if I set my temp target to 91c and tell it to prioritize that they can go to above 1500Mhz easy. The problem is my top card will hit 90c and the bottom will go to 70c. If I had a different motherboard that spaced the PCIe slots more to allow better airflow this wouldn't be a problem. I'm not willing to let my card run that high during games. So I temp target to 85c and let it settle around 1430. That's plenty, even if it throttles back a bit more which it may at times, I'm still hitting 17k in Firestrike which is 7k points higher than my 670 SLI setup was.

 

[Alpha0mega]

Nothing being able to OC at all is odd. My GPU setup is similar to yours. Two Zotac 970s in SLI. In my case I am able to OC them fairly well, much better than I had hoped. Using Afterburner, I had gotten them to +200 on core and +300 on memory, no overvolting, stable for 2 hours of Heaven, but playing Far Cry 4 maxed out, I had driver crashes every 2-3 hours.

I tried upping the voltage by +15 mV, though I don’t think it does anything. The max voltage never goes above 1.212v no matter what I set in Afterburner. The overclock.net thread linked in an above post also mentions that the only way voltage increase works is with Firestorm.

So I have been reducing the OC little-by-little. At +190 core, +200 memory, FC4 had one driver crash in ~10 hours of playing. So have dropped the memory to +185. Interestingly, it’s the GPU I bought later that always crashes. It’s on top, but I have 2 slots of spacing between the cards and a fan blowing between them. The max temp difference is 6 C, with the top one hotter of course. I tried changing the SLI master card to the lower one too, but it’s always the top one that had a driver crash.

It could be the temps, or it could be the silicon lottery. Since I bought them about a month separated, they are a little different. The earlier bought card has Samsung memory, the later one has Hynix. Their coolers are also subtly different, the older one’s heat pipes are a closer to the edge of the card.

 

[hunkeelin]

Then what is stopping my card to OC to 1500mhz? When I run heaven, during the most demanding scene it throttles to 1366mhz. I am perflex because i did set the temp limit to 90C. GPUZ was reading 79C when my gpu sli throttles. I don't get it.

I wish someone can help me answer this
if I go kraken g10 + h50 will it let me achieve results better than G1/msi/strix?
or
since g10+h50 doesn't cool vrm, I will need to go the custom water cooling route. If I do go that route can I achieve OC that surpass G1/msi/strix?

Lastly, what is the reason why I cna't achieve the OC results as compare to msi/strix/g1? is it because they have better quality pcb or it's mostly because of cooling?

Thanks

 

[cmdrdredd]

Did you set priority to temp or power target? When I set mine to power they throttle early and often. Temp they don't as much.

Also according to an article here on AT, they have a built in tdp limiter and they will never go above that with a stock bios. They will throttle and drop voltage to keep under the limit.

Despite not being temperature limited, what we can see right away is that regardless of the clock speed settings it uses, the GTX 970 FTW is TDP limited under all scenarios. At no point in time are we able to maintain the card’s top boost bin, and instead the card spends its time fluctuating between the boost bins it can hold while maintaining power consumption of 145W. The actual drop off from the maximum boost bin depends heavily on the game; some games average clock speeds close to the maximum, while others have to pull way back.

http://www.anandtech.com/show/8568/the-geforce-gtx-970-review-feat-evga/15

 

[hunkeelin]

^i priority the temp.