Originally Posted by eduardorz80
This is my first time posting to a forum so I apologize in advance if I am not following etiquette, just tell what the rules are and I'll follow.
I have two Gigabyte Gtx 980 G1 on water. Asic: 69.2% / 74.7%
I installed the latest bios from Laithian, Thank you very much, V5Beta.
The new bios works great, +60 on the core and memory. So my clocks sit at 1591 (boost) and Mem at 1753.
My question is regarding Power Limit, it seems that as long as my power limit at least 100% anything higher has no effect on performance, my TDP does not go past 90%, Avg is 80%.
Is this normal? or could I possibly get more performance out these cards, maybe I am doing something wrong? I
came from a 780ti and TDP was 120% when overclocked.
Here are my system specs:
5960x @ 4.4
16gb 2800mhz ram
Gtx 980 G1 x2
I agree with cookieboyeli, you have nice hardware that is a great overclock! Thanks for sharing the results with us!
To be clear, we are referring to the Power TDP% in software. Here is an example:I've been wanting to post about this "TDP% software reading" so thanks for asking. In my BIOS 100% power is higher than stock power, EDIT: Not with the final release version. which is why you are seeing that. If you take a look at the video I posted on perfcaps, you can actually see how every last bit of the power % slider is fully utilized and needed to remove perfcaps at extreme GPU loads :For you and others,
let's dig in a bit as to some of the findings through proven testing. To answer your question quickly, you cannot compare the TDP% value of a GTX980 custom BIOS to a 780Ti value.
As a matter of fact, you can't even compare the TDP% value of a stock BIOS
of the same make/model with a modified BIOS
(assuming the power settings were not
left stock).The easy (I hope) explanation is:
There are tweaks in the BIOS that allow you to adjust the range of the power % slider
in your fav overclocking program. This is why with my BIOS you will notice the power % slider can go up to 140%
. I could actually make the range on the power % slider any value that I wanted and the ACTUAL max power, when the slider is maxxed all the way out, actually never changes. The actual TOTAL POWER allowed is IDENTICAL in both cases when the slider is set to MAX.
PS. The reason I chose 140% in V5BETA because it gave you almost double the "tweak points" for power %, for those who want the choice
but it is mostly cosmetic because we typically just max it out anyway
. Compared with the stock OEM BIOS, 122% was 22 tweak points from 100%-122% max. V5BETA is 140% with 40 tweak points from 100%-140% max. It is simply how the MAX power configured in the BIOS is distributed across the total RANGE of the power % slider.Another example:
Let's use the default power % range with the G1 (maximum of 122%) as an example. What does 122% mean? People tend to think that raising it to 122% power would be adding 22% more power to the card than stock
right? NOPE. That is NOT the case necessarily.
. It would ONLY be true if the total max power allowed (value) in the BIOS was mathematically 22% more than stock wattage. It all depends on how the BIOS is configured (what max power value was used) and what the default value is. IE: 22% of 5 dollars isn't the same as 22% of 1,000 dollars but they are both 22% right? How I figured this out:
I started to play around with the TDP value (top section in power tab). I increased it to multiple values (even an insanely high one) over and over and re-flashed.. There was NO difference in the TDP% software reading under load. I knew right away at this point that the TDP% calculation had NOTHING TO DO with the ACTUAL TDP value in the BIOS if modifying the TDP value in the BIOS made no difference with the calculation of the power TDP% in software. I started to play around with the increase of power on the PCI-e rails..(slightly lower and raise) and again, NO difference in the power TDP% reading either! I was like HUH?!? I figured, OK it must be the relationship between the MAX POWER value and the MAX TDP value...
.. NOPE AGAIN!
There was NO difference in the TDP% reading with this change either! I tested if they matched or one was slightly lower/higher also.
Scratching my head, I needed to seek an understanding as to what DOES impact the TDP% reading (what changes CAN I make that affect the reading/value) and it basically came down to 2 things:(1) Memory Overclocking
- UNDER LOAD If you set your memory to +0 and monitor power TDP% and then set it to +500 (or highest you can) and monitor TDP% again you'll notice overclocking ONLY the memory made the TDP% jump considerably
.(2) MAX POWER section of the BIOS
- Adjusting the difference
between the default
power value and the maximum
power value. The same adjustment that allows you to specify the range of the power slider.
The thing is, adjusting the mathematical DIFFERENCE between DEFAULT power and MAX POWER doesn't ever actually CHANGE THE MAXIMUM POWER ALLOWED it just changes the calculation of the power TDP% shown in software... which at that point to me, I don't even bother looking at it. It has no actual relationship to the power being DRAWN or any relationship to the actual HEAT OUTPUT of the card as far as I can tell.