980TI Asic Quality

[Kryptyx]

My 2 new cards (Asus Strix 980TI OC) have an ASIC quality score of 83% and 87%. How does this compare with what some of you guys have? I did some research and it seems most people are in the 70-75 range but its hard to gauge if that is for reference cards or non-reference. I noticed EVGA charges a lot more for 80+ ASIC quality on their Kingpin. Did I get lucky or is this an expected range for non-reference cards? I had 2 non factory OC cards prior with Asic of 72 and 75.

[Nacho Marco Segui]

But do they oc nicely? That is the important thing

[Lays]

Just luck really, just because they have super high ASIC doesn't necessarily mean they'll OC well.

 

Mine is %73.8 for instance, max clocks are 1550 on the core & 9108 mhz effective on the memory with a universal waterblock & a fan pointed at the VRM & memory

 

[Kryptyx]

Haven't tried yet, I installed them, ran ASIC test and then boxed up my system because I moved. I might try tonight. Previous cards would get around 1500. Hopefully these will too.

[iamdarkyoshi]

4 minutes ago, Lays said:

Just luck really, just because they have super high ASIC doesn't necessarily mean they'll OC well.

 

Mine is %73.8 for instance, max clocks are 1550 on the core & 9108 mhz effective on the memory with a universal waterblock & a fan pointed at the VRM & memory

 

I have heard that LOWER ASIC quality generally means better overvolting and therefore overclocking headroom

[David_K]

My GTX 970 G1 has asic score of 64% , also eplida memory which just seals the deal.. the luck is amazing with this one, friend of mine got his G1 970 few months before and has a score higher than 80 and samsung memory  

[CarterDavison]

3 minutes ago, iamdarkyoshi said:

I have heard that LOWER ASIC quality generally means better overvolting and therefore overclocking headroom

That's not true at all.

[Lays]

Just now, iamdarkyoshi said:

I have heard that LOWER ASIC quality generally means better overvolting and therefore overclocking headroom

 

Not so much.  Maxwell really doesn't like volts much, a lot of the time 1.25v is where the 980 ti will stop scaling, so higher ASIC is usually better, because it will usually reach higher clocks @ low volts. (Some cards will reach max clocks at 1.18-1.22v on air even.)

[CarterDavison]

Just now, Lays said:

 

Not so much.  Maxwell really doesn't like volts much, a lot of the time 1.25v is where the 980 ti will stop scaling, so higher ASIC is usually better, because it will usually reach higher clocks @ low volts. (Some cards will reach max clocks at 1.18-1.22v on air even.)

Ty on mobile lol

[iamdarkyoshi]

2 minutes ago, CarterDavison said:

That's horse shit./That's not true at all.

I forgot where I read this, but it was a while ago.

 

2 minutes ago, Lays said:

 

Not so much.  Maxwell really doesn't like volts much, a lot of the time 1.25v is where the 980 ti will stop scaling, so higher ASIC is usually better, because it will usually reach higher clocks @ low volts. (Some cards will reach max clocks at 1.18-1.22v on air even.)

What about for older AMD cards? That is where I heard this from

[David_K]

People usually say that lower asic's score is better for LN2 and extreme overclocking, while higher asic's are more efficient so you get to the higher clocks on lower voltages and this is better for air/water cooling. 

[Lays]

8 minutes ago, David_K said:

People usually say that lower asic's score is better for LN2 and extreme overclocking, while higher asic's are more efficient so you get to the higher clocks on lower voltages and this is better for air/water cooling. 

 

On older generations this may have been true, but when it comes to scaling on cold, it's somewhat of a tossup it seems right now on Maxwell GPU's.

 

A buddy of mine has a 90.1% ASIC 2173 mhz K|ngp|n card (on Ln2 completely maxed out.) He said he thinks it'll do 2200 next time, just needs to figure out some things with it.

I believe the card Vince (K|NGP|N) used to set the records on Firestrike recently was also a card that was over 85%.  (He mentioned having a card that was over 85% that did 1630 on AIR, so I assume that's most likely the card he used to set the records with)

I haven't really seen anyone on LN2 mentioning ASIC values from 980 TI's, I've seen a few 70% cards that were doing 2055-2070 on Ln2, but haven't really seen much else.

[Kryptyx]

3 minutes ago, Lays said:

 

On older generations this may have been true, but when it comes to scaling on cold, it's somewhat of a tossup it seems right now on Maxwell GPU's.

 

A buddy of mine has a 90.1% ASIC 2173 mhz K|ngp|n card (on Ln2 completely maxed out.) He said he thinks it'll do 2200 next time, just needs to figure out some things with it.

I believe the card Vince (K|NGP|N) used to set the records on Firestrike recently was also a card that was over 85%.  (He mentioned having a card that was over 85% that did 1630 on AIR, so I assume that's most likely the card he used to set the records with)

I haven't really seen anyone on LN2 mentioning ASIC values from 980 TI's, I've seen a few 70% cards that were doing 2055-2070 on Ln2, but haven't really seen much else.

Ah ok. Obviously the real test is to OC them. I was just curious with what others were scoring with their 980TIs.

[CarterDavison]

1 hour ago, iamdarkyoshi said:

 

What about for older AMD cards? That is where I heard this from

Ah, in that case you would have been correct. I thought you meant in Maxwell context, Maxwell is very strange.

[Kryptyx]

ASIC quality is all about voltage leakage. The higher the number, the more efficient the chip is. This is why a higher number generally means a better overclock, especially on air.