• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

ASIC Quality - Here is the truth

Kaapstad said:
+1

It was only the last few seconds of his video that was of any value.

Lower ASIC, Higher Volts, More Heat, Less Boost.

Instead of doing a video about ASICs before reviewing the EVGA Kingpin 980 Ti it would have been more useful to do a review with 4 EVGA Kingpins each with a different ASIC to see if there was any real world differences.
Click to expand...

Yer, show what each one does would be cool.
 
Ash0787 said:
I dont get why having a higher asic quality lowers the voltage scaling though
Click to expand...

GPU VID is calculated by ASIC value to compensate for current leakage, this leakage is power being wasted due to imperfections within the GPU die.

Lower ASIC cards will boost lower and use more power doing so than higher ASIC cards. If you have a 980Ti with a 73% ASIC, for 1400 core, the VID will be around say 1.16v, and for a 76% card it might be around 1.14v. This is why higher ASIC boost higher depending on the quality as predetermined within the BIOS

On that 73% ASIC card, the other 27% imagine is current being wasted, and not used efficiently
 
MMMmmmm

I did like Jay but he has become more and more waffle and less substance. However his builds are pretty awesome.

Now I really don't understand what the reality of ASIC is.

Mine is a lowly 69 yet still overclocks very well with some fantastic Heaven and Fire strike scores all on air.....

It also holds a 1500 boost all day long, should I choose to keep it there.

Do I need, as a normal consumer, anymore??

I really think it's only an enthusiast worry TBH. Also does it really matter?

Queue Kaapstad to include ASIC in the benchmark ROH requirements.... (might be interesting TBH)
 
Still waiting to find some serious/independent data points like this. Get 12 identical NVidia/model cards. 4 grouped pretty close with low ASIC, 4 with middle ASIC, 4 with high ASIC. Push them all until artifacts/stability issues occur.

How is the ASIC/stability/overclock correlation? Any? Random? :confused:

BTW: People trying to find RMA reasons because of ASIC is ridiculous. Card has a spec. If it does not crash/artifact within those specs, it is not faulty. Overclocking potential on top of that is bonus/lottery, not something you are entitled to. Yes the manufacturers are good at salesmanship pages with fancy graphics of overclocking tools and so on. Read the small text though ;)
 
Just watch the video. It makes sense. Better Asic, less voltage leak, which equals to better voltage/performance ratio.

People posts on this forum are enough to understand most of it, no need for any serious data :P
 
razorpakk said:
Just watch the video. It makes sense. Better Asic, less voltage leak, which equals to better voltage/performance ratio.

People posts on this forum are enough to understand most of it, no need for any serious data :P
Click to expand...
I have watched the video. My basic question is still: Independent of ASIC quality, all else being equal, how many other factors of the GPU die itself are there to stability/scaling? If any?

If it's already is an established fact that ASIC quality is the only factor of the GPU die for volt/leakage/freq/scaling, I stand corrected. If not, I am still interested in good research on how much weight other factors have vs ASIC. Still read posts about high ASIC that crash/artifact quickly, and low ASIC that scale nicely. Mostly other way around though. All this could be external factors of GPU die. I.e. cooling type, thermal paste appliance, different voltage regulators, PSU, and so on.

I'll partly agree with you on the "no need for serious data" :p I am asking for info that probably has not much practical usage. Sometimes it's better to just adjust the levers and accept the result you're getting with the card you have :cool:
 
Without watching the vid and being too much of a nerd, I think the differences are minimal. Lower ASIC means less boost, more heat....but just how less or more? Minimal.

I reckon it's a marketing gimmick to get people to buy more cards to look for the golden ASIC number of 100% :)
 
I'm interested in seeing the actual range, % implies 0-100 but I've yet to see anyone report anything outside about 55-84.
 
I'd be interested in seeing the highest ASIC gpu ever. Highest I've had is 83 and that was an excellent clocker.

Someone should start an ASIC thread and people post their ASIC but only if it's higher than the post above them.
 
At least the guy makes an interesting watch. Better than someone just reading off script and monotone.

As mentioned, I took the key points from the video which were about 15 seconds worth in total, but I enjoyed listening to it whilst I worked.
 
I should really clear up what I was trying to say in my earlier post in this thread. ASIC quality is an attempt by GPUZ to read a number put there by AMD/Nvidia without any idea as to what criteria AMD/Nvidia used to arrive at that number.

Looking around the web there seems to be no correlation between the ASIC quality and overclocking or how good the card is. There doesn't seem to be any consistency at all.

Basically it's a meaningless number to the end user, useful only to AMD or Nvidia.
 
At one point some GPU's were getting more than 100% and ASIC quality has occasionally change between different versions of GPU-Z.

I have no doubt there are different ASIC qualities I just doubt GPU-Z's validity as it's clearly using some sort of algorithm, it's not just reading a number put there by the GPU manufacturers.
 
Yes I recall some people were reporting that they were getting scores in excess of 120% at one point, GPU-Z was later patched and scores changed on identical hardware.

I'm not sure how they get their reading but when stuff like that happens it's probably safe to assume that it's not totally accurate.
 
You must log in or register to reply here.
Back
Top Bottom