Why is everyone so obsessed with Classifieds from EVGA?

Exyia said:
nothing but the generic pictures/descriptions

question still remains unanswered....wtf does he actually do other than throw LN2 on it, turn up voltage, and increase clock offset....all of which we can do on any other card

again - CPU I can respect - so many more variables to play with...but GPU I just don't understand. I want to be proven wrong by ignorance, but nobody who actually knows anything is speaking up - just fans drinking the kool aid of the name brands
Click to expand...

Have you even tried reading any of the OC forums? There is tons of information on OCing GPU cards out there. And I am not sure what you want as a difference in OCing EVGA vs OCing other cards? They use the same GPU, and you use the same methods to OC them. The only difference between different resellers is the components they use and the PCB makeup. And even the PCB isn't necessarily that different since they are all based off the reference PCB that Nvidia puts out.

The main differences you will get with EVGA vs ASUS vs Whoever is the components they use and the coolers they use. Some of which you can replace yourself if you know what you are doing.

As for what Kingpin actually does, do you think he would list everything? He has once or twice in the past explained some of the things he does to get his crazy overclocks, but he never reveals all of his secrets, i mean when your job is to compete for the top OCs, why would you reveal all your tricks? Some of the things he has done is replace certain components on the cards, reduce resistance in certain areas to allow for higher voltage to parts on the board, flash the firmware with his own custom setups, etc.

As with all overclocking, one of the keys is to be methodical and gradual. Boosting the voltage up too quickly can adversely affect the card and limit your peak. You are far more likely to get a higher ceiling if you move slowly and methodically rather than just jump to the first 'safe' OC settings.
 
bpizzle1 said:
I loved my 780 classy. It felt very solid, and OCd like a beast. That said, I've been reading that since they launched the Kingpin line, the regular classified cards aren't what they used to be. Granted, this generation of classified is a low enough premium to justify over a reference card imo. You still get the better cooler and power delivery.

I just ordered a Zotac 980ti AMP! Extreme this morning. I've never had a Zotac card, but the reviews combined with the new lifetime warranty swayed my wallet this time. It's the same price as the classified and has a better cooler and seems to be in the same league as the G1 Gaming.
Click to expand...

I bought two Zotac GTX 980s when they were released.
Zotac has been a pleasure to deal with.
My cards overclocked like crazy.
Their CS sent me several emails after I registered the cards, just to make sure everything was going fine.

You won't be disappointed.:D
 
sk3tch said:
There was a Kingpin 580. So it's 4 series.

The overall point is - the money spent to "upgrade" to a higher end, same GPU card is typically not worth it - but it is fun. And that's what the hobby is about. So if you have the coin - go for it. Just don't expect your Classified 980 Ti to blow you away versus a vanilla 980 Ti. You will not be able to tell a difference - except with things like temperature and also the fun of the additional tweaking/tuning you may be able to do. On the other hand - stock OCing is what OCing is all about. Getting more bang for your buck. So the whole crazy card trend is kind of funny to me. I fall for it too...don't get me wrong. :)
Click to expand...

Ah, thanks for the info. I think I'm going to shy away from evga and try for a HOF down the line. Like on of the previous said about their 980 ti classi, that is a poor clock. Hey, it's only 50 bucks more.....

I'm on my phone at the moment, but did you adjust the voltages?
 
sk3tch said:
There was a Kingpin 580. So it's 4 series.

The overall point is - the money spent to "upgrade" to a higher end, same GPU card is typically not worth it - but it is fun. And that's what the hobby is about. So if you have the coin - go for it. Just don't expect your Classified 980 Ti to blow you away versus a vanilla 980 Ti. You will not be able to tell a difference - except with things like temperature and also the fun of the additional tweaking/tuning you may be able to do. On the other hand - stock OCing is what OCing is all about. Getting more bang for your buck. So the whole crazy card trend is kind of funny to me. I fall for it too...don't get me wrong. :)
Click to expand...

Except stock cards are not guaranteed to hit speeds above the base rating whereas some of the overclocked cards are. By buying an already overclocked card, you are hedging your bets a little bit that it will have a higher ceiling.

But personally I usually stick with a stock card and do my own OCs unless the OC'd card comes with a nice quality cooling system with it that costs less or the same amount as it would cost me to get my own aftermarket system and time to install.
 
NoOther said:
Except stock cards are not guaranteed to hit speeds above the base rating whereas some of the overclocked cards are. By buying an already overclocked card, you are hedging your bets a little bit that it will have a higher ceiling.
Click to expand...

I would guess that the hedge would be negligible in the grand scheme of things -- about as infinitesimal as the proportion of GTX 980 Tis that would not be able to hit the SKU's specified clocks within the default voltage table.

Speaking of which there is a correlation between the operating voltage of the card and the ASIC rating. Lower % cards typically have a higher operating voltage bin defined by the BIOS. This is why we see a fair number of non-reference cards with relatively low ASIC scores (in the 60s or less) -- they are simply able to stably* reach the specified SKU clock speeds with the higher operating voltage. Some of these low-ASIC cards have high ceilings but that's the story as to why SKU selection is really still a crapshoot.

*Some people might construe this as "binning" but really 1291MHz isn't awfully aggressive to achieve on stock voltage with almost every GTX 980 Ti out there. Stability testing can be pretty loose as demonstrated in some of ASUS's earlier offerings (where they might test samples from batches rather than each individual GPU). The only cards that might be of value here are those with stipulated boost clocks that are beyond the reach of some proportion of the overall population for stock voltage tables.

At best custom PCBs will net you a couple of additional MHz (on air/water) over the same exact chip on a different PCB. Any difference in this PCB-derived OC advantage is usually overwhelmed by the tremendous variability in the silicon lottery itself.

Moral of the story is with rare exceptions (i.e. notable factory OCs) just buy the SKU that fits your desired mix of cooling/noise/aesthetics.
 
KickAssCop said:
I lost twice this time. Both my MSI and this classified didn't break 1500. That sucks ass. Even though in terms of real world performance it is meaningless but given how much you pay for these puppies, you expect better.
Click to expand...

As you have the card, how does the fan size on the Classified compared to a normal gtx 980 Ti with the ACX 2.0+ cooler? Same size or bigger? It looks bigger in the pics with the raised shroud on the top.
 
Direfox said:
As you have the card, how does the fan size on the Classified compared to a normal gtx 980 Ti with the ACX 2.0+ cooler? Same size or bigger? It looks bigger in the pics with the raised shroud on the top.
Click to expand...
It is bigger. However, MSI fans are similar with higher size than normal backplate on a graphics card.
 
MSI has 100mm fans and I assume the Classified does to from the card height. The gigabyte uses small 80mm fans which create more more noise and higher RPMs to stay cool, hence one of the reasons I sent mine back. ASUS STRIX used 95mm like my 980.

I'm guessing the standard evga cards use 90mm judging from the pci standard card hight they are
 
BlueSaber80 said:
MSI has 100mm fans and I assume the Classified does to from the card height. The gigabyte uses small 80mm fans which create more more noise and higher RPMs to stay cool, hence one of the reasons I sent mine back. ASUS STRIX used 95mm like my 980.

I'm guessing the standard evga cards use 90mm judging from the pci standard card hight they are
Click to expand...

Evga site has been down all day, but my short phone search found the 980 ti kingpin has 100mm, which looks larger than the classified since the shrouds are similar/identical. Also to note, the classified states having the acx 2.0+ which i think is 92mm
 
KickAssCop said:
I lost twice this time. Both my MSI and this classified didn't break 1500. That sucks ass. Even though in terms of real world performance it is meaningless but given how much you pay for these puppies, you expect better.
Click to expand...

Did you mess with the voltages at all? For instance using the Classified Tool for the EVGA card and bumping up core voltage, etc. That is unfortunate though as even my Titan X with a 1.255v bios can hit 1531 MHz quite easily. Then again it's all silicon lottery. My 780ti Kingpin card was crap barely doing above stock (1215 MHz or so) and the RMA of that card could do only 1345 MHz with 1.34v.
 
On a related note, Kingpin actually posted a statement about the Maxwell architecture and it's lack of clock gain to voltage increases. He basically said that, beyond 1.25v, Maxwell will see very very little boost in core clock unless your load temps are at least below 10-15c (maybe 25c, can't really remember right now.)

Basically, it seems to barely matter whether you purchase a vanilla card or a classy or Kingpin variant if you're only going to be air or even water cooling. Your performance will hardly change beyond 1.25v.

This explains why my Nvidia branded 980 actually overclocks higher than a buddy of mine's 980 Classified. Even though he can pump way more voltage into it, his ASIC score is like 68% whereas mine is 73.8%. While not a huge difference, my vanilla card does 1560 core and +500 memory on 1.25v whereas his maxes out around 1500 even with +300 on memory. Even under water, it doesn't look like it would matter much at all :(

Also, his statement is for ALL vendors, not just EVGA cards. It's all about the silicon lottery.
 
Last edited:
MaelstromOC said:
On a related note, Kingpin actually posted a statement about the Maxwell architecture and it's lack of clock gain to voltage increases. He basically said that, beyond 1.25v, Maxwell will see very very little boost in core clock unless your load temps are at least below 10-15c (maybe 25c, can't really remember right now.)

Basically, it seems to barely matter whether you purchase a vanilla card or a classy or Kingpin variant if you're only going to be air or even water cooling. Your performance will hardly change beyond 1.25v.

This explains why my Nvidia branded 980 actually overclocks higher than a buddy of mine's 980 Classified. Even though he can pump way more voltage into it, his ASIC score is like 68% whereas mine is 73.8%. While not a huge difference, my vanilla card does 1560 core and +500 memory on 1.25v whereas his maxes out around 1500 even with +300 on memory. Even under water, it doesn't look like it would matter much at all :(

Also, his statement is for ALL vendors, not just EVGA cards. It's all about the silicon lottery.
Click to expand...
Sorry but I call BS on a reference 980 being stable under full load at 1560/8000. Not to mention you would hit the power limit looong before those clocks anyway. The highest I saw any review site get was under 1520 with 100% fan going and that still throttled to under 1500 at times. Many 980 reference cards could not even do 1500 at all.
 
misterbobby said:
Sorry but I call BS on a reference 980 being stable under full load at 1560/8000. Not to mention you would hit the power limit looong before those clocks anyway. The highest I saw any review site get was under 1520 with 100% fan going and that still throttled to under 1500 at times. Many 980 reference cards could not even do 1500 at all.
Click to expand...

Might be possible with a modified BIOS. I've had more golden reference than non-reference cards and they really stretched their legs with BIOS edits to disable silly things like boost and to expand the power limit.

It seems that overclocking might be limited by the following factors (in order of importance):
1. Silicon lottery
2. BIOS on video card
3. Cooling
4. The rest (e.g. PCB)
 
misterbobby said:
Sorry but I call BS on a reference 980 being stable under full load at 1560/8000. Not to mention you would hit the power limit looong before those clocks anyway. The highest I saw any review site get was under 1520 with 100% fan going and that still throttled to under 1500 at times. Many 980 reference cards could not even do 1500 at all.
Click to expand...

Sorry, the 1560/8000 was with a bios I modded myself to increase the power limit. Did it get toasty with the reference cooler? Sure did! Did it run those clocks? Sure did!

To clarify, without the bios being modified, I do hit power target and downclock if I increase the memory overclock while also boosting the core clock as well. Basically, I can run 1560/~7500 (rough estimate on memory) with the stock configuration before hitting the power limit. Raising the power limit let's me hit 1560/8000, but I can do no more. If I don't push the ram up that high, I can hit 1580 on the core, but the memory overclock gives a higher gain in FPS than a measly 20mhz on the core.

Improved cooling would help keep the temps down, but it definitely just shows how limited the cards are beyond their factory set limits.

Also, as the guy above me stated, I've actually had better luck with vanilla cards myself. I used to always buy a Classy or FTW card (I generally purchase from EVGA), but after having had a few vanilla cards that outperformed them, I just wrote them off as a lazy tax. Why pay more for something I can do myself with software? Cooling not considered, of course.
 
Shogon said:
Did you mess with the voltages at all? For instance using the Classified Tool for the EVGA card and bumping up core voltage, etc. That is unfortunate though as even my Titan X with a 1.255v bios can hit 1531 MHz quite easily. Then again it's all silicon lottery. My 780ti Kingpin card was crap barely doing above stock (1215 MHz or so) and the RMA of that card could do only 1345 MHz with 1.34v.
Click to expand...
I have been using after burner. Will try precision and see what's up.
 
I looked at my ASIC score - 63%. Yikes.

Been playing with a modded BIOS though and I can get up to around 1440MHz now at 1.21v.
 
To anybody with this card, does it stay reasonably quiet? If you aren't overclocking it further?

I have my computer purchase ready but I can't decide on which 980 Ti to get!
 
Seyba said:
To anybody with this card, does it stay reasonably quiet? If you aren't overclocking it further?

I have my computer purchase ready but I can't decide on which 980 Ti to get!
Click to expand...
It is quieter than my MSI 980 Ti gaming and also quieter than my 970 WF3 cards.
 
You must log in or register to reply here.
Back
Top