crazydane
|
|
October 30, 2017, 11:20:12 AM |
|
Ah... nothing really interesting to see here people. 1080Ti is still the king. Yes, I love my 1080ti's But too expensive... 1080ti is the best balanced with the price and powers solution market has now, if you can grab one for $700 it would be basically a stealer! see for yourself : 1. performance wise one 1080ti is equivalent (rough) for two 1070 and three 1060 2. 1080ti eats like 250-260W, 1070 - 130-150W, so 2 of them will be 260-300W, 1060 - 100-120W, so 3 of them will be 300-360W - who is the king? 1080ti! 3. the price for 1080ti is $750, for 2x1070 is $850, for 3x1060 is $950. who is the king? 1080ti! How many sols does one 1080ti do compared to a 2 1070s? I'm interested in the watts and temps of running one 1080i vs 2 1070s Here's a 1070 rig on top and a 1080ti at the bottom: And here a water cooled 1080Ti rig being pushed a little harder:
|
|
|
|
RealSwissMiner
|
|
October 30, 2017, 12:35:23 PM |
|
Nice setup crazydane. Temperatues looking good and also the efficiency across the board. How much is that water cooling solution consuming energy wise?
|
|
|
|
leowonderful
Legendary
Offline
Activity: 1624
Merit: 1130
Bitcoin FTW!
|
|
October 30, 2017, 12:42:30 PM |
|
Also possible to pre-order on nvidias website for 449 i belive. (only for us/canada)
What is it that is the determinating factor of the hashrate? It has 8GB mem, CUDA cores close to an 1080, why shouldnt it be as good as a 1080? Any hardware pro's that can answer?
Memory is a big factor. GDDR5x in the 1080 compared to GDDR5 in the GTX 1070 TI. It should be what makes or breaks the GPU in terms of mining Only on the ETH algorithm, nothing else is as picky about memory - but it likely will easily beat the 1070 and MIGHT argue or even beat the 1080ti on ETH mining (the 1070 beats the 1080 on ETH due to that GDDR5X). It should fall in just a HAIR under the 1080 and easily beat the 1070 on everything else. My pair of Aorus GTX 1080ti cards (the BASE Aorus, not the higher-clocked version) pull about 780 sol/s at 250 watts (100% tdp), +100 core +150 mem, after a fair bit of tweek testing the core and memory settings seem to be optimal for it on ZCash. They still pulls 720 at 200 watts TDP (80%) and the "efficiency knee" seems to be somewhere in the 150 watt (620 sol) to 175 watt (680 sol) range. One is in my game system running Win7, the other in my experimental B250 Mining Pro riser rig running XUbuntu LINUX - but the stats are the same give or take measurement limits. Even pushing, I've never gotten an Aorus to go much over 70C on 80% fan even with a 90F ambient - and the one in my gaming machine has never gone over 66C or so since it's in an air-conditioned room at around 80F ambient in a case with good airflow. My Gigabyte Windforce GTX 1080ti pulls close to those numbers, but falls down some on the high end as it overheats faster and can't maintain boost as well - and it runs 3-4C hotter unless I pump the fan up to 90% or so. Up side, it's only 8+6 for power connection instead of 8+8 for the Aorus. Best I've seen out of a 1070 was around 450 ballpark, some folks have claimed a lot higher but they apparently have had a lot better luck on getting their memory to overclock a lot than I have. I guess it’s more or less up to what you’ll want to mine in the end, then. I’d still recommend possibly getting a 1080 over a 1070 ti if pricing stays as it is (mostly because of mining flexibility and perhaps better resale down the line) and until we actually get reliable and confirmed mining hashrate reports here. Forgot to specify Ethereum there, mining profitability has shifted a lot in the last few months
|
|
|
|
BlockMedx
Newbie
Offline
Activity: 56
Merit: 0
|
|
October 30, 2017, 12:54:19 PM |
|
I keep struggling to understand why anybody would want one of these over a 1080ti. I understand they're less expensive and have a performance boost over the 1070's....but the 1080ti is still the best price per performance tradeoff despite its higher price.
|
|
|
|
yangshenglu
Member
Offline
Activity: 161
Merit: 16
|
|
October 30, 2017, 12:58:38 PM |
|
What is this?
|
|
|
|
Dovaogedys
Newbie
Offline
Activity: 32
Merit: 0
|
|
October 30, 2017, 05:20:15 PM |
|
I already have one here. Eager to test it now.
|
|
|
|
Boathouse
|
|
October 30, 2017, 05:37:08 PM |
|
I already have one here. Eager to test it now.
That is early than expected.
|
|
|
|
crazydane
|
|
October 30, 2017, 06:00:34 PM |
|
Nice setup crazydane. Temperatues looking good and also the efficiency across the board. How much is that water cooling solution consuming energy wise?
The D5 PWM pump consumes maybe 12w at 1.5 gpm flow rate. The 120mm fans on the radiator only consume 0.6W each at 1600 RPM (Noctua NF-F12 PWM). So the water cooling add a whopping 14.4 watts to the power consumption of that rig. I'm adding a 2nd 480 rad to get the temps down in the 30's as my ambient to coolant differential is rather high right now (about 20 degrees C where it should be more like 2 degrees).
|
|
|
|
P4ndoraBox
|
|
October 30, 2017, 06:17:44 PM |
|
I already have one here. Eager to test it now.
You already have it ? Share your result.
|
|
|
|
RealSwissMiner
|
|
October 30, 2017, 07:23:54 PM |
|
Nice setup crazydane. Temperatues looking good and also the efficiency across the board. How much is that water cooling solution consuming energy wise?
The D5 PWM pump consumes maybe 12w at 1.5 gpm flow rate. The 120mm fans on the radiator only consume 0.6W each at 1600 RPM (Noctua NF-F12 PWM). So the water cooling add a whopping 14.4 watts to the power consumption of that rig. I'm adding a 2nd 480 rad to get the temps down in the 30's as my ambient to coolant differential is rather high right now (about 20 degrees C where it should be more like 2 degrees). Cool, sounds interesting. Thanks for sharing, appreciated!
|
|
|
|
pfft
|
|
October 31, 2017, 07:22:12 AM |
|
I already have one here. Eager to test it now.
Any update?
|
|
|
|
aws-premium
Jr. Member
Offline
Activity: 54
Merit: 10
|
|
October 31, 2017, 09:04:12 AM |
|
I already have one here. Eager to test it now.
Sharing is caring, what about the Results of your Test?
|
https://bitcointalk.org/index.php?topic=2306383.0 Buy AWS 100$ EDU Coupon
|
|
|
|
ZenFr
Legendary
Offline
Activity: 1260
Merit: 1046
|
|
October 31, 2017, 02:37:54 PM |
|
Thank you for sharing. These resultsare bad, very bad (at stock...). Approximatively 6% more with Equihash... only with +25% in cuda cores and much power consumption. The same result or even less with Dagger-Hashimoto and ever muche more watts burned.
|
|
|
|
P00P135
|
|
October 31, 2017, 02:40:45 PM |
|
Thanks for taking the time to test them. Hopefully they hash faster than this though
|
|
|
|
flip4flop
|
|
October 31, 2017, 02:48:37 PM |
|
Yikes based on this I am happy I didn’t preorder. Hopefully that’s not the average results. I will have to jump on some of the price dropped 1070s if this seems to be the overall results. The question I guess now becomes to jump up to the 1080 ti or not. Hopefully some more results for 1070 ti start rolling in with higher hash rates. I was really expecting more overall gains from the 1070s.
|
|
|
|
PranDoW
|
|
October 31, 2017, 02:58:55 PM |
|
In my country GTX 1070ti gonna cost more expensive as usual. Not sure that I gonna buy them right now I am alright with my 1070
|
|
|
|
laineux
|
|
October 31, 2017, 03:15:33 PM |
|
Very disappointed That is the exact model I preordered... What clock timings have you used ?
|
|
|
|
pfft
|
|
October 31, 2017, 03:24:31 PM |
|
in my country right now 1070ti and 1080 have the same price....what should i buy ?
|
|
|
|
laineux
|
|
October 31, 2017, 03:43:31 PM |
|
in my country right now 1070ti and 1080 have the same price....what should i buy ?
1080 are bad to ETH mining
|
|
|
|
|