Pages:
Author

Topic: Any news about nvidia 3060ti hash rates? - page 4. (Read 5712 times)

legendary
Activity: 4256
Merit: 8551
'The right to privacy matters'
December 22, 2020, 09:53:15 PM
Anyone have an EVGA 3060 Ti XC Gaming card? I have an opportunity to purchase a few of these at close to MSRP. I’m curious what kind of experience people have had mining ETH (or other coins) with this card.

Buy them. If you do not want them I will take at least one of them maybe 2 of them.

I would pay a small  markup and shipping .

All assuming you are usa based.
TGJ
jr. member
Activity: 173
Merit: 5
December 22, 2020, 08:48:48 PM
Anyone have an EVGA 3060 Ti XC Gaming card? I have an opportunity to purchase a few of these at close to MSRP. I’m curious what kind of experience people have had mining ETH (or other coins) with this card.
legendary
Activity: 4256
Merit: 8551
'The right to privacy matters'
December 22, 2020, 04:49:05 PM
also, what I'm seeing here is that the mobo itself only draws about 20 watts, is that correct?  I was wondering how much the motherboards would add to the equation, seems to be that it's very small in comparison.  

the cpu if it is a t variant is lowest power of the cpus.

intel g3900t

https://ark.intel.com/content/www/us/en/ark/products/90738/intel-celeron-processor-g3900t-2m-cache-2-60-ghz.html
tdp is 35watts

intel g3900
https://www.intel.com/content/www/us/en/products/processors/celeron/g3900.html
has a tdp of 51 watts

intel g4400t tdp of 35 watts
https://ark.intel.com/content/www/us/en/ark/products/90614/intel-pentium-processor-g4400t-3m-cache-2-90-ghz.html

intel 4900t tdp of 35 watts
https://ark.intel.com/content/www/us/en/ark/products/129952/intel-celeron-g4900t-processor-2m-cache-2-90-ghz.html


so using a t version saves watts ,drawback is if you want to run a huge rig say 8 3060ti's

you may need a better cpu.

Also bronze, silver, gold platinum, titanium psu.,

at a 500 watt build drawing a constant 500 watts.

bronze 625 at the kwatt

titanium 525 at the kwatt

also mobo efficiency

also the card itself and I am back to topic I have a 3 fan aorus 3060ti smos says I do

60.84 mh at 119 watts the 3060ti aorus
94.29 mh at 219 watts the 3080 evga

that adds to 318 watts for gpu

it  estimates my total consumption is 378 watts

I have an FSP 1200 watt platinum
I have a intel g4400t

I use a stick of samung ram with this board

https://www.amazon.com/ONDA-B250BTC-D8P-Mining-Motherboard-LGA1151/dp/B0791Z44YX

now I run on the kwatt meter at 397-401 watts. i also know this meter has always been 5% hot. I have other kwatt meters that real less.

for fun I will run this 2 card rig with just the 3080 and see how much we drop.

So on the reboot.
I read 219 for the gpu via smos
I read 259 for the rig   via smos
I read 269 for the rig via kwatt

so I dropped 378 to 259 = 119 watts via smos reading
I dropped 398 to 269 = 129 watts via kwatt reading

since the meter on the house counts I would use 129 watts for 60.84 for my card. hoping that the hot meter is maybe 5 or 6 watts higher then my house meter.

so the house meter is charging me maybe 123 to 130 watts for the card. it is hard to know how accurate the house meter really is compared to the kwatt meter.  For me to do it correctly I need to have my wife out of the house for at least 4 days and only run 1 circuit in the whole house to test properly.  Which is not going to happen.
sr. member
Activity: 2142
Merit: 353
Xtreme Monster
December 22, 2020, 04:19:03 PM
Sure, I can make a new thread if needed.  My questions are specifically about the 3060Ti, as that is the board I am focusing all my studying on right now.  But if talking about building a rig with these cards is off topic, I'm find to make a new one.  I just wanted to clarify that when I'm asking these questions, they are specific to the board in question.   

3 fans will always have a larger heatsink because it uses more space, so reason 3 fans are better = larger heatsink, however need to check the heatsink weight, example, the first design of sapphire rx 580 had a somewhat average heatsink, while it was good, summer time had reports it sucked, third design came with a heavy heatsink, even in the sahara desert that thing worked without a problem hehe, 2 fans, it did not need a third fan. Yeah a third fan uses more electricity, 6 watts average, 12v 0.5 amperes.
member
Activity: 115
Merit: 16
December 22, 2020, 04:01:54 PM
Sure, I can make a new thread if needed.  My questions are specifically about the 3060Ti, as that is the board I am focusing all my studying on right now.  But if talking about building a rig with these cards is off topic, I'm find to make a new one.  I just wanted to clarify that when I'm asking these questions, they are specific to the board in question.   
legendary
Activity: 2366
Merit: 1408
December 22, 2020, 03:46:58 PM
Yes I have a Killawatt, have had it for years. Was excited to see that I "already have a piece of mining gear" LOL.  I will be using them on the rigs I build for sure.

so I'm still not sure if it's a big deal to go with a 3 fan model over a 2 fan.  I think i asked this question, but I don't see that I got a solid response.  In my mind, I'm thinking 3 fans = more electricity consumption, but better cooling, longer life of the board.  Or is all of that mitigated by backing off the gas pedal so to speak like you guys are saying - not running the GPUs at full speed...?



Hey, better to create a new thread for you and you can ask anything you want, because it's becoming off-topic of this thread about 3060tis, ok?

About 3 or 2 fans, you can watch some reviews on youtube and foruns, or ask here about the model you want, but usually 3 fans will be better than 2, and the power consumption of 1 more fan is extremely low, nothing to worry about
There are brands that have better colling performance in general


member
Activity: 115
Merit: 16
December 22, 2020, 11:47:34 AM
Yes I have a Killawatt, have had it for years. Was excited to see that I "already have a piece of mining gear" LOL.  I will be using them on the rigs I build for sure.

so I'm still not sure if it's a big deal to go with a 3 fan model over a 2 fan.  I think i asked this question, but I don't see that I got a solid response.  In my mind, I'm thinking 3 fans = more electricity consumption, but better cooling, longer life of the board.  Or is all of that mitigated by backing off the gas pedal so to speak like you guys are saying - not running the GPUs at full speed...?

legendary
Activity: 2366
Merit: 1408
December 21, 2020, 08:01:41 PM
Thanks Phil,

and these power settings are adjusted through some GUI for the cards, or through BIOS settings, or something similar?  This will be something I can learn to adjust and manage, correct? 



Nvidia cards are pretty easy to manage and  tweak, you can use MSI Afterburner to do that
In 5 minutes you can adjust Nvidias, most are: reduce TDP to 60% or less, reduce Core Clock, set fans to 60% or 70% and push your memories to something like + 1000 or +1500mhz

AMDs are harder to tweak, but Nvidias you'll be able to, easy

To see about power consumption of your system, the only valid method is to have a Kill A Watt to measure power on the wall
legendary
Activity: 3248
Merit: 1070
December 21, 2020, 06:09:44 PM
is the 3080 woth it over the little 3060 ti?

It depends, 3080 average is around 92mhs, 3060ti average is 60mhs, 3080 240 watts, 3060ti 140 watts. 2 x 3080 = 3 x 3060ti.

it come all down about density i guess, if you want it betterto go with the 3080 which also cost more than 50%, otherwise for small setup the 3060 ti is better
sr. member
Activity: 2142
Merit: 353
Xtreme Monster
December 21, 2020, 04:55:48 PM
is the 3080 woth it over the little 3060 ti?

It depends, 3080 average is around 92mhs, 3060ti average is 60mhs, 3080 240 watts, 3060ti 140 watts. 2 x 3080 = 3 x 3060ti.
legendary
Activity: 3248
Merit: 1070
December 21, 2020, 04:21:25 PM
is the 3080 woth it over the little 3060 ti?
member
Activity: 115
Merit: 16
December 21, 2020, 11:37:22 AM
also, what I'm seeing here is that the mobo itself only draws about 20 watts, is that correct?  I was wondering how much the motherboards would add to the equation, seems to be that it's very small in comparison. 
member
Activity: 115
Merit: 16
December 21, 2020, 11:35:03 AM
Thanks Phil,

and these power settings are adjusted through some GUI for the cards, or through BIOS settings, or something similar?  This will be something I can learn to adjust and manage, correct? 

legendary
Activity: 4256
Merit: 8551
'The right to privacy matters'
December 21, 2020, 11:31:12 AM
TDP is balls to wall pedal to the metal spec for your card.

MOST cards dies if run at tdp of 100% under a year.

also you hardly get any more hash maxing to 100%
 of tdp.

here is a test rig

Video link
https://youtu.be/aie0LlO8Q5w

this rig using a titanium psu pulls 282-285 watts

gpus are set to

1) 95 watts - Tdp is 120
2) 85 watts - Tdp is 120
3) 85 watts - Tdp is 120

that is 265 watts that is 360

setting closer to the tdp for these cards in particular does not get more hash.  So basically I don't do it.

Mining is a grind.

 Don't burn out the gear by setting to the tdp. =  "this is mostly always true"
member
Activity: 115
Merit: 16
December 21, 2020, 11:25:18 AM
Gotcha, thanks for the clarification on that. I'll read that wiki an see if it helps clear that up for me.

As for the moving target - yes.  I understand this.  This is why I'm focused more on how much "coin" can be mined, rather that how much "profit" can be mined.  My intent is to mine at a level that I can cover the power bill out of pocket, and not need to sell coins to pay the utilities.  I recognized that mining "profitably" will go up and down quite a bit it seems, but since I'm not planning to mine for cash flow, I feel like i will be less worried about the day to day performance of the coins, but more interested in what they will look like in 20 years.  So my focus is on economy of scale and how I can maximize my efforts without spending a ton of money (adding infrastructure for power, etc...) It's the way I think through projects at the day job, where every new level of efficiency attained is critical to the project.  Now, once I start doing this for reals next month, I may have a completely different tune, and if that ends up being the case, I will be certain to share my findings and experiences here too to help others deciding to dive in or not. 
full member
Activity: 1275
Merit: 141
December 21, 2020, 11:07:37 AM
Can someone explain this acronym - TDP?  I'm not following how that plays into the card specs.

Basically, I'm getting started early next month with a pre-built rig to cut my teeth on, but then I'm looking at building an array of rigs in the coming year.  as I plan to mine and hold for long term, the math I am doing based on posted MH/s and power consumption rates, has me looking at the "watts per MH" and "cost per MH".  I don't know if these are ideal metrics to track, but if they are, and if the 3060 Ti's become available within reasonably close proximity to their MSRP, then it appears to me the 3060 Ti's are good value for my purposes. I won't ever truly know first hand until I get my hand one one or two, I'm hoping to do that early next year too and drop them into the rig I'm buying and see for myself.  I know I do not have a good handle on the information I'm seeing here about tweaking the power consumption on the boards, but I do like what I'm seeing about underclocking them a little bit to keep the heat and power down, and extend their life.  It's clear to me that this is a healthy mix of art and science to get this to work efficiently.  Clearly there's a lot more to it then just dropping cards on a board and pushing a button.  I hope I am able to pick this all up quickly.  

Thermal Design Power - Ie how much power it is thermally designed to handle.

https://en.wikipedia.org/wiki/Thermal_design_power

And realize for your second part..its hitting a moving target.  Every calculation you make is subject to change almost immediately. 
member
Activity: 115
Merit: 16
December 21, 2020, 10:58:59 AM
Can someone explain this acronym - TDP?  I'm not following how that plays into the card specs.

Basically, I'm getting started early next month with a pre-built rig to cut my teeth on, but then I'm looking at building an array of rigs in the coming year.  as I plan to mine and hold for long term, the math I am doing based on posted MH/s and power consumption rates, has me looking at the "watts per MH" and "cost per MH".  I don't know if these are ideal metrics to track, but if they are, and if the 3060 Ti's become available within reasonably close proximity to their MSRP, then it appears to me the 3060 Ti's are good value for my purposes. I won't ever truly know first hand until I get my hand one one or two, I'm hoping to do that early next year too and drop them into the rig I'm buying and see for myself.  I know I do not have a good handle on the information I'm seeing here about tweaking the power consumption on the boards, but I do like what I'm seeing about underclocking them a little bit to keep the heat and power down, and extend their life.  It's clear to me that this is a healthy mix of art and science to get this to work efficiently.  Clearly there's a lot more to it then just dropping cards on a board and pushing a button.  I hope I am able to pick this all up quickly. 
newbie
Activity: 64
Merit: 0
December 21, 2020, 06:35:14 AM
that´s why I too wanted to buy 3070...but I struggle with the 240 and the 280 tdp variant.
legendary
Activity: 2366
Merit: 1408
December 21, 2020, 06:18:47 AM
I think the 120 W power is a result of using 50 % of TDP (which is 220 to 240 for most 3060 cards). For my gaming PC I think about buying a MSI GeForce RTX 3070 Suprim X 8GB GDDR6 which has 280 w TDP. So 50 % would be 140w. I think with 140w it gets the same hashrate....but when I make TDP of 40 or 45 % to reach 120 W, do you think I might get less mhash? So a 3060 with a higher TDP than 240 is no good choice for mining? could that be correct?

Then I would buy the normal NON-Supreme variant of the 3070....if I can get one. 3070 supreme is in stock today Sad

I bought Asus dual oc 3070, and it's hashing 61Mhs with 52% TDP, + 1099 memory oc and it's stable
The power consumption on the wall is 135 exactly (Phoenix miner shows 114w)
If I lower to 51% or less (minimum 45% TDP) I lose hashrate, so 52% for me it's the sweet spot

3060 ti is better if you'll only mine and can buy with good price, but I choose 3070 to do some gaming and the price is only 10% more of 3060ti at the moment, and it's easy to find
newbie
Activity: 64
Merit: 0
December 21, 2020, 05:59:34 AM
I think the 120 W power is a result of using 50 % of TDP (which is 220 to 240 for most 3060 cards). For my gaming PC I think about buying a MSI GeForce RTX 3070 Suprim X 8GB GDDR6 which has 280 w TDP. So 50 % would be 140w. I think with 140w it gets the same hashrate....but when I make TDP of 40 or 45 % to reach 120 W, do you think I might get less mhash? So a 3060 with a higher TDP than 240 is no good choice for mining? could that be correct?

Then I would buy the normal NON-Supreme variant of the 3070....if I can get one. 3070 supreme is in stock today Sad
Pages:
Jump to: