7970 Lightning, worth the premium?

[pgi947]

1950 1.65v, could probably have ran the volts a little lower, had plenty of room to go higher but memory correction crippled anything higher.

Correction limit is somewhere between 1800-2000, so aiming for 1800 would be a good start.

 

[hughythomas]

Does overclocking the memory yield much of an overall increase in performance?

I've only been focusing on the core - is it worth giving the memory a boost too?

 

[theCrim]

What overclocks are people getting without increasing the voltage and what are you getting when increasing the voltage?

Mine doesn't seem to want to go much further on stock voltage.

1095/1450 are the clocks I get at stock voltage without getting artifacts and 1125/1450 stable through heaven with the occasional artifact.

For my clocks with increased voltage I used to hit 1300-1320 on the core and 1800-1850 for the memory on air a few drivers back through heaven when it was a little warmer. That was at 1.362v but with the most recent beta drivers I've barely been able to hit 1285/1825 in both heaven and 3dmark11.

Does overclocking the memory yield much of an overall increase in performance?

I've only been focusing on the core - is it worth giving the memory a boost too?

Try increasing the memory clock and see for yourself. For me anything above 1750 yielded very little performance increases. I can't comment on the difference between 1850(highest my card reaches on air) and something like pgi's crazy 1950-2000 mem clock.

 

[pgi947]

You won't get massive gains in games from memory increases, for benching you might.

 

[G-Sharp]

I took the plunge and went for one of these. I just hope that it fits my case as it is rated to fit 12" cards while this one is a fraction of a mm over. Should be OK I'd assume.

Will be getting another one in CF if this one turns out good. Should be here tomorrow so I'll post some results when it comes.

 

[theCrim]

Any reason you chose the Lightning over the Matrix? Do you have an MDP to DP cable for your monitor?

 

[G-Sharp]

Any reason you chose the Lightning over the Matrix? Do you have an MDP to DP cable for your monitor?

I have a DP on my monitor (ASUS PB278Q) and the card comes with a mDP to DP adapter.

The ASUS is not ideal for my case (FT02) due to its orientation. I would've certainly considered it if it didn't have that problem.

You need to take into account that the Matrix takes up 3 slots, while the Lightning only 2, so you may get better breathing space with the Lightning should you choose to go CF.

 

[~Mike~]

Managed to get the core upto 1210, not sure mine will go much higher unless I up the voltage.

Will try upping the memory tomorrow.

 

[Louis_Y]

I like quiet so I overclocked to a fan profile of 46%. That gives me 1175 core and 1775 mem on 1.23v stable at around 65C core temp in Heaven, and around 55C in EvE which is what I'm mostly playing at the moment. Fairly happy with that.

I may have to claim a record for the worst ASIC ever though: 60.3%

 

[pgi947]

That is low, my last lightning was 62.4%, to get the very best clocks from these cards you need water

 

[DK06CTR]

only just my ASIC is only 61.1% on my lightning. ive managed to get 1200 core 1665 memory stable at 1.256v. think thats as much as i need for now. temps arent above 60c on the memory,vrms or core. and volts i think are safe at that on air.

 

[G-Sharp]

That is low, my last lightning was 62.4%, to get the very best clocks from these cards you need water

Have there been any blocks released for these? Thinking about going water after I get my second.

 

[pgi947]

Volts are fine, could manage a touch higher if needed

Yes ek have full cover blocks & ocuk sell them

 

[~Mike~]

I may have to claim a record for the worst ASIC ever though: 60.3%

Only just, mine is 60.7%

Is a low ASIC normal for the Lightning?

 

[pgi947]

Very very normal, if ASIC theory holds any truth at all then the chips are binned for Lightnings to have low ASIC.

 

[~Mike~]

Very very normal, if ASIC theory holds any truth at all then the chips are binned for Lightnings to have low ASIC.

Can you explain plz?

 

[pgi947]

ASIC theory (I call it theory as its neither proven nor disproven):

High ASIC: Usually low stock voltage, will hit reasonably high clocks on air, wouldn't benefit much from water/sub-zero

Low ASIC: Pretty much the opposite of above, higher stock voltage, won't shape up too well on air, but would fly with water/sub-zero.

iirc ASIC relates to how concentrated the voltage is on the core, high ASIC is supposed to be highly concentrated, so runs cooler = better for air, low ASIC isn't, so voltage is spread throughout the chip giving off more heat, but benefits from cooler temps, as I think the more voltage is spread the better it may perform.

^The above is my assumption anyhow, most including my self don't really buy into it. Though if it holds any truth this explains why all lightnings have low ASIC (usually below 70%) as the cards are aimed at extreme overclocking in sub-zero temps.

 

[Louis_Y]

If anyone has a high ASIC Lightning that clocks well on air but would like a low one to clock under water, I might be persuaded to swap with them then

 

[theCrim]

If anyone has a high ASIC Lightning that clocks well on air but would like a low one to clock under water, I might be persuaded to swap with them then

It's doesn't mean anything...I have a lower ASIC than you and i get similar clocks to those on water

 

[Louis_Y]

'twas in jest, hence the