Discussion in 'News' started by btarunr, Jan 21, 2012.
So higher or lower ASIC a good thing or a bad thing? What does the % actually represent?
I got only 47.1% on ASIC quality is my video card defective? a lower % means more likey to die? or less OC? or less reliability?????????
I've an ASIC quality of 102,6%. Can it be higher than 100% or is this an error?
Nice job. Again you forgot to put the names of those who translated gpu-z
Not sure whether this is accurate or not..
The ASIC quality of my MSI GTX 460 Hawk is 87.4% whereas the quality of my Talon Attack version is 64.6%. If one isn't familiar with MSI cards, see here.
Surely the card with nearly 23% higher quality would qualify for a higher spec SKU.
Maybe it's reading the cards wrong when in SLI..
Here's mine. 84.3%.
GALAXY GTX 560Ti GC.
VDDC stock is 1.012 V.
Thanks for the update.
I think we should all calm down on the ASIC thing here for a minute and realize it is a new feature and things might either need to be ironed out first..... or that we all don't fully understand everything. I am seeing a lot of questions by people almost panicking saying "my one stock reference video card is 93% while my other high end non reference card is 43%". Just calm down....I think it is just a great little new tool we all can use and it shouldn't be made into more then that. And it defiantly shouldn't be made into to the new metric on which we sell second hand video cards. Because as we all can see from this thread.... some people with 63% ASIC can bump up their voltage and get a killer overclock. Sometimes higher then people who have a ASIC of 80% or more.
For example.... I have a ASIC of 84% on one card and 82.7 on my other HD7970. I can overclock these cards very nicely without voltage increase. So I think it would be pointless to us it as the new metric for everything second hand. People might be screwing themselves over!
Wafer quality refers to leakage. A low-leakage part is going to be the most efficient, but is more likely to either coldbug or not overclock very far(pseaking in terms of many samples form the same process)
Higher leakage parts(with lower quality), when cooled properly, should clock better when temps are decreased.
A part with zero leakage is NOT what an overclocker wants...it's what a person who wants to save on power consumption wants.
Basically, if a silicon part does not leak, it doesn't give off as much excess power as heat.
20mmrain, is the lower "quality card" slightly hotter? I'd like to see how each clocks under water or LN2. finding the sweetspot for each process would be interesting.
My gigabyte windforce OC is only 66.6% im guessing these GTX570s are mostly bad leakage. I haven't tried any OC but im sure it wont be too bad
780 core is stock on this
just a quick question
If you card is producing artifacts can this been read from these ASIC scores? as in a bad core will have <20% or something?
The next logical step seems to check for correlations between ASIC, power consumption and overclockability, in-between a fixed model (say GTX 560Ti), of course.
Then, we may see, if this is going to be the "next big thing", whilst the last one was the CPU batch...
my GTX460 TOP that runs all day 870mhz gets rated for 40%
my GTX560 gets 71%
edit just noticed, the pixel fill rate for my 460 is wrong.
my 560 ssc 2gb asic quality is 103% is that good or.....
Well this just sucks ass . Yet another old ass card of mine that is not supported ! So I have no clue what mine is or even would be !
Sparkle Calibre GTX 550 Ti 1GB: 71.1% (only OCs about 50MHz to 1GHz core)
ECS GT 440 512MB GDDR5: 80.6% (OCs about 80MHz to 890 core.)
How long does the test take?
81.8% on my sapphire 7970.
Which one? The ASIC percentage is simply read off the chip.
If I understand this all correctly, Higher the % just means more likely to have a higher OC, It's doesn't mean your card is going to die.
Great job as always W1zzard, I'm starting to think your a mad scientist of GPU's.
Also I know my 4850 isn't currently supported (which is perfectly ok because I already know my card can't be upped by 20mhz), but just for reference where do you click to do this test?
Neat, nice set of improvements.
Only if you can use them all . Mine is so OLD it can not take advantage of them . GOD it sucks being POOR !
Every thing I own is so old it all belongs in a museum !
Whoa there. Your stuff looks pretty damn good to me. Appreciate what you have, life is easier that way.
Awe man, I've just checked your specs, it's not that old. That Q9650 at 4GHz is still very respectable.
I don't blame you for wanting the latest though and I hope you're in a position to get some new hardware soon.
EDIT: Dammit erocker, you beat me to it.
<---- Not as crappy as my stuff, but then again I can still play all my games at descent quality. I just want to have new stuff!!
I could buy a New Intel 2011 a couple 7970 and stuff, but I'm super tight with my money I always find a reason to not buy something.
I just built a Pentium E6600 (Wolfdale 2MB) system with an ASRock G41M-S3 and overclocked it to 3.8GHz and was mighty surprised how well that system runs. It was actually able to hold 50-60FPS in TF2 with a GTX 550 Ti installed (CPU was maxed out though) but I give it credit for running TF2 much better than my laptop's quad core Phenom II 2.1GHz which goes down to 30 and less all the time. This setup is just another folder/cruncher now (I needed another place to put the GTX 550) and it does excellent. Your Q9650 is probably better than many of the Phenom IIs out there, and people still use them everywhere.
Hmm. On my notebook with ATI Mobility FireGL V5000 (mobility x700) it always shows PCI-E 1.1 x16@x16 1.1 no matter if I have powerplay set to optimal performance or optimal battery life.
According to my knowledge M26 (mobility x700) supports dynamic lane count switching that should be adjusting number of PCI-E lines that are used.
Separate names with a comma.