Overclocked 980 vs 980Ti vs 1080 = Potential is there, but not with FE

[Arct1c0n]

2 minutes ago, App4that said:

Have to say my Hybrid with a reference PCB has been more stable than all my friends who have custom. And it overclocks with way less voltage.

 

Im really curiuos to see my 79% ASIC rated 980 Ti kingpin I got for $450 in action vs the new 980 TI upgraded Hybrid model from EVGA I snagged new in the box for $500 and how they compare this week.  

 

Old design 

 

 

 

Updated design

 

[App4that]

1 minute ago, Arct1c0n said:

 

Im really curiuos to see my 79% ASIC rated 980 Ti kingpin in action vs the new 980 TI upgraded Hybrid model from EVGA and how they compare this week. 

My Hybrid has a 78% ASIC rating and overclocks to 1500mhz just by increasing the power limit. It boosts on its own to 1366mhz and that's where I leave it as I've never needed more. And I run a Predator.

[Kryptyx]

 

 

I guess I hit the lotto with my Asus cards. I would be upset if I had a Kingpin less than 80% though.

[Energycore]

3 minutes ago, Kryptyx said:

 

How does the ASIC Quality measurement work?

[App4that]

22 minutes ago, Arct1c0n said:

 

Im really curiuos to see my 79% ASIC rated 980 Ti kingpin I got for $450 in action vs the new 980 TI upgraded Hybrid model from EVGA I snagged new in the box for $500 and how they compare this week.  

 

Old design 

 

 

 

Updated design

 

 

That updated cooler is fugly, wonder what they changed it. Might be to have it quieter as mine is anything but quiet. It's louder at idle than my 390 was gaming.

[Arct1c0n]

6 minutes ago, Kryptyx said:

 

 

I guess I hit the lotto with my Asus cards. I would be upset if I had a Kingpin less than 80% though.

 

No shit you did. You only got above 80% with Kingpin for paying $1250

[Arct1c0n]

3 minutes ago, Energycore said:

How does the ASIC Quality measurement work?

 

[Energycore]

Just now, Arct1c0n said:

 

Appreciate it.

[Kryptyx]

1 minute ago, Energycore said:

How does the ASIC Quality measurement work?

Basically it measures the quality of the chip. In theory, the higher the number the better the card will overclock. This is because it is more efficient. Most of the time this holds up but not always.

[App4that]

Remember that ASIC quality doesn't mean shit. There are other confounding variables that impact your ability to overclock you can't measure.

[Kryptyx]

1 minute ago, App4that said:

Remember that ASIC quality doesn't mean shit. There are other confounding variables that impact your ability to overclock you can't measure.

I wouldn't say doesn't mean shit, but I agree there are other factors that weigh in more. ASIC is just a good starting point.

[Energycore]

So in the context of Graphics Processors, it's a measure of how much voltage is needed to drive the stock GPU speed, from what I could gather.

[App4that]

Just now, Kryptyx said:

I wouldn't say doesn't mean shit, but I agree there are other factors that weigh in more. ASIC is just a good starting point.

Take this as an example. Remember that I hit 1500mhz stable in games with my Hybrid with a reference PCB and ASIC rating of 78.1%

 

Cut to the 20min mark if you don't want to watch the whole thing, though I suggest you do.

 

 

[Arct1c0n]

5 minutes ago, App4that said:

Remember that ASIC quality doesn't mean shit. There are other confounding variables that impact your ability to overclock you can't measure.

 

Uhhh no,, it doesn't not mean "jack shit"

[Arct1c0n]

2 minutes ago, Energycore said:

So in the context of Graphics Processors, it's a measure of how much voltage is needed to drive the stock GPU speed, from what I could gather.

 

Rather how much you can push a GPU on its stock voltage without crashing

[App4that]

1 minute ago, Arct1c0n said:

 

Uhhh no,, it doesn't not mean "jack shit"

Actually, yeah, it kinda does.

 

Take horsepower. Go to a dealership and if you had 30 of the exact same model of card with the same equipment you would get 30 different horsepower reading on the dyno. There is a lottery there too. Taking the car with the most hp WILL NOT give you the fastest car, because of the other things that go into what makes a car, fast.

 

Everyone wants a simple metric to gauge performance, hell so do I. But life isn't that easy.

[Energycore]

1 minute ago, Arct1c0n said:

 

Rather how much you can push a GPU on its stock voltage without crashing

So you could literally get to the number by dividing max stable clock by voltage?

 

Also, my 560 Ti has 84.6% ASIC quality

 

I can't compare it to other 560 Tis though. I'm running 900 MHz (80 over stock) at 1000mV (62 below stock). But I don't know if that's any good.

[Kryptyx]

1 minute ago, App4that said:

Take this as an example. Remember that I hit 1500mhz stable in games with my Hybrid with a reference PCB and ASIC rating of 78.1%

 

Cut to the 20min mark if you don't want to watch the whole thing, though I suggest you do.

 

 

I've watched it, I'm a big fan of Jay's videos. But keep in mind, 78.1% is pretty high for a 980TI. I'm able to push almost 1600mhz (1579mhz) on my cards on air.

[Kryptyx]

4 minutes ago, Energycore said:

So you could literally get to the number by dividing max stable clock by voltage?

 

Also, my 560 Ti has 84.6% ASIC quality

 

I can't compare it to other 560 Tis though. I'm running 900 MHz (80 over stock) at 1000mV (62 below stock). But I don't know if that's any good.

Older generations have higher ASIC quality as well.

[App4that]

1 minute ago, Kryptyx said:

I've watched it, I'm a big fan of Jay's videos. But keep in mind, 78.1% is pretty high for a 980TI. I'm able to push almost 1600mhz (1579mhz) on my cards on air.

If you're pushing 1579mhz on a 980ti that's the highest I've ever heard of on air. The Lightning Jay tested only hit 1535mhz. That's the most he's ever gotten.

[App4that]

This is another great resource on what ASIC is.

 

 

[Arct1c0n]

9 minutes ago, App4that said:

Actually, yeah, it kinda does.

 

Take horsepower. Go to a dealership and if you had 30 of the exact same model of card with the same equipment you would get 30 different horsepower reading on the dyno. There is a lottery there too. Taking the car with the most hp WILL NOT give you the fastest car, because of the other things that go into what makes a car, fast.

 

 

That is the shittiest parallel comparison i've ever heard

[App4that]

2 minutes ago, Arct1c0n said:

 

That is the shittiest parallel comparison i've ever heard

Absolutely, but it was from the hip. And not entirely inaccurate. You're talking to a former drag racer and professional mechanic. The parallels between performance PCs and motorsports is striking.

[Kryptyx]

12 minutes ago, App4that said:

If you're pushing 1579mhz on a 980ti that's the highest I've ever heard of on air. The Lightning Jay tested only hit 1535mhz. That's the most he's ever gotten.

Yeah it showed 1579 in GPU Tweak monitor but 3d Mark read it as 1537 so idk. Here is the result: http://www.3dmark.com/fs/7715568

Keep in mind, my CPU was not overclocked at the time.

[App4that]

2 minutes ago, Kryptyx said:

Yeah it showed 1579 in GPU Tweak monitor but 3d Mark read it as 1537 so idk. Here is the result: http://www.3dmark.com/fs/7715568

Keep in mind, my CPU was not overclocked at the time.

I'd say trust 3D mark but I could be wrong. Ether way on point.