Pages:
Author

Topic: Are 1070 TI's the next best GPU for mining? 4.7 sols per watt - page 3. (Read 9258 times)

jr. member
Activity: 50
Merit: 1
How do you adjust the power limit as a percentage, I'm using simple mining os and the only power limit options I have myself are the actual wattage draw.  Do I just run these at maximum and then backwardsly work out the percentage ?

QuintLeo broke the wattage down in another 1070ti thread,
https://bitcointalk.org/index.php?topic=2324651.20

Got a suprise this morning - my EVGA SC 1070 ti was on my doorstep when I wasn't expecting it 'till Monday.

 I had already installed the latest 388.13 drivers on my Win7 single-slot "test rig" machine while doing some testing on a new 1080 ti earlier this week, so the test rig was ready to go.
 I HAVE noticed some sluggishness on this machine, but it's a low-end Athlon 5050E dual core with 2 GB of ram so that's not unusual on it (machine was a LOT snappier on XP back in the day).
 Testing on this rig was done in an A/C environment at 80F give or take a degree.

 First thing I notice - the memory clock is WAY LOW, 7600 base clock for some reason. Due to that, this is the first EVGA card I can crank up to "+700" memory in Afterburner and actually have it work - as that's only 9000 actual EFFECTIVE memory clock.
 Boost is NOT limited to the "factory spec" numbers - it's like any other 1070/1080/1080ti it'll boost as high as the power settings and temp let it.
 Default TDP is 180 watts, same as a base 1080 or some higher-TDP 1070 models.
 Another odd thing - Afterburner will let me OVERVOLT the card, but it won't let me UNDERVOLT it at all.

 Did some testing on a copy of Genoil I had left over on that machine from ETH mining days - meh, pretty much the same as a 1070, obviously ETH is seriously memory limited on this card (not unexpected given other widespread reporting).

 Moved to ZEC testing with EBWF.
 Testing methodology - make a change in Afterburner, wait *at least* 2 rounds of EWBF stats posted, THEN average the next 3 stat rounds for the posted numbers.

 At stock settings (except for using my "standard" Afterburner fan profile):
 Hashrate settles in at about 467 sol/s with VERY little variation with the wattage figure in EWBF bouncing between 168 and 180 watts most of the time for an efficiency around 2.6-2.7 range.
 Core is boosting to 1848 Mhz pretty much as a straight line, 62-63C at 64-66% fan.
 Memory rock solid at "3798" (effective 7596).

 Crank the memory up to +700 in Afterburner no other changes:
 Hashrate settles in around 525 sol/s with noticeable variation, power however is pretty solid at 180 watts for efficiency at a little over 2.9
 Core is boosting to between 1810 and 1823 with a lot of bounce between those figures, temp 64C at 68% fan.
 Memory is rock solid at "4498" (effective 8996) and STAYS that way in the rest of the "+700 memory" testing.

 Dial in +200 core clock in Afterburner:
 Hashrate settles in to about 538 sol/s with some variation, power pretty solid at 180, efficiency a bit over 3.
 Core is bouncing some, mostly 1987 with lows down to 1967 and occasional bounces up to 2000. I forgot to check temps, think they bounced up to 65C at 70% fan or so.

 Drop TDP to 90% in Afterburner:
 Hashrate at 540 sol/s and pretty SOLID, power solid at 161 watts for efficiency at 3.38 - and yes, the average hashrate DID GO UP A HAIR vs 100% TDP.
 Core is bouncing 1911 and 1924 with an occasional 1898 or 1937, temp drops to 63C at 66% fan
 
 Drop TDP to 80% in Afterburner:
 Hashrate finally drops to 529 sol/s pretty solid, power pretty solid at 146, efficiency pretty solid at 3.64
 Core bouncing a lot, 1810 to 1873 range, Temp to 60C at 60% fan.

 Drop TDP to 70% in Afterburner:
 Hashrate to 509 sol/s but bouncing some, power pretty solid at 126 watts for efficiency of appx 3.96 with one reading above 4 (517 sol/s)
 Core bouncing in the 1709 to 1759 range, temp 56C at 52% fan.

 Drop TDP to 60% in Afterburner (WATCH THESE FIGURES):
 Hashrate 457 sol/s pretty stable, 103 solid watts, 4.46 efficiency.
 Core bouncing in the 1557 to 1607 range, temp 54C at 48% fan.

 This would be a VERY NICE setting to run a quiet multi-card rig in your BEDROOM, I have to put my head less than 2 feet from the GPU fans to hear them AT ALL.
 This is also THE highest efficiency figure I've seen to date out of ANY card I've owned, and I don't remember seeing a better figure reported (though I might have missed or forgotten a slightly better reported figure somewhere).


 Drop TDP to 50% in Afterburner:
 Core clock drops to under 1200 and the hashrate drops so far the efficiency gets WORSE. Don't go this low on this card.





As of today I am running 9 GTX 1070ti off of one EVGA 1600 T2 power supply, so my settings in simple miner are as follows,

Core
200,200,200,200,200,200,200,200,200
Memory   
700,700,700,700,700,700,700,700,700
Power Limit
136,136,136,136,136,136,136,136,136    with everything mining away, no monitor or keyboard / mouse attached, I am drawing 1,365 Watts at the wall,

getting around
========== Sol/s: 4548.0 Sol/W: 3.72 Avg: 4536.2 I/s: 2425.8 Sh: 49.35 1.00 209
member
Activity: 644
Merit: 24
So it looks like the price of the 1070s is the same price as 1070 Ti.  Is there any more that's better than the others or any that I should just avoid?

Read through this entire thread and there is another 1070ti thread that I've been watching too.  From what I have gathered, the fans are not up to par on the Gigabyte 1070ti's and I'm also staying away from the Zotac cards.  Vosk has some Zotac rigs running, but I think he said that they were not performing as well as others that he has tested.  That may just be the mini's ... Can't remember.  There is also some question about the fans in the Zotac cards.  That could also just be the mini's.  Not 100% certain.  But seriously read through these threads.  There is a wealth of information on 1070ti's.
member
Activity: 644
Merit: 24

When you're calculating 110w per card, is that with everything else factored in and measuring power consumption at the wall?

 110 watts as set as the power limit under Linux using Nvidia-settings, or reported by nvidia-smi in either Linux or Windows (or calculated from the %TDP setting in something like Afterburner).

 At the wall will vary more due to power supply efficiency curves.



Thank you! Thank you!  nvidia-smi is showing 107W-110W on each gpu.
newbie
Activity: 62
Merit: 0
So it looks like the price of the 1070s is the same price as 1070 Ti.  Is there any more that's better than the others or any that I should just avoid?
newbie
Activity: 34
Merit: 0
How do you adjust the power limit as a percentage, I'm using simple mining os and the only power limit options I have myself are the actual wattage draw.  Do I just run these at maximum and then backwardsly work out the percentage ?
member
Activity: 644
Merit: 24

When you're calculating 110w per card, is that with everything else factored in and measuring power consumption at the wall?

 110 watts as set as the power limit under Linux using Nvidia-settings, or reported by nvidia-smi in either Linux or Windows (or calculated from the %TDP setting in something like Afterburner).

 At the wall will vary more due to power supply efficiency curves.



Oh well in that case, I guess I'm way lower than I think.
newbie
Activity: 46
Merit: 0
Nice project! Good choices
legendary
Activity: 1498
Merit: 1030

When you're calculating 110w per card, is that with everything else factored in and measuring power consumption at the wall?

 110 watts as set as the power limit under Linux using Nvidia-settings, or reported by nvidia-smi in either Linux or Windows (or calculated from the %TDP setting in something like Afterburner).

 At the wall will vary more due to power supply efficiency curves.

legendary
Activity: 1498
Merit: 1030
Also possible - 490sols @ 95W / 4.95sol per W  Wink

http://i66.tinypic.com/2ewf8d1.jpg
'
 69% power limit set in Afterburner is 124 watts - something funky going on there.

 Probably the same sort of thing I've seen on "short efficiency spikes" in EBWF where it will report one card at way low power draw but normal hashrate - momentary down spike on power JUST as EBWF looked at it.

legendary
Activity: 1498
Merit: 1030
Am I missing something here? I'm checking profit calculations on whattomine.com and the ROI for the 1070ti is somewhere over 6 months even if I set the hash to 500 sol at 110 watts, while rx580 ROI in little over 4 months.

 You missed ZEC almost tripling in price over the 2-3 weeks after your post.

 8-P

member
Activity: 644
Merit: 24
It's looking like in order to keep myself right at 500 sol/s, I'm using 135W per card at the wall.  That's with +200 core clock and +750 memory clock.  I cut the fans back to 50%.  All my temps are still below 60c.

This thing is a hell of a lot quieter and cooler than my vega rig mining etn.
member
Activity: 644
Merit: 24
Good to know.  I have some EVGA FTW2 1070ti's coming on Friday ... (Hopefully) ... They are supposed to ship today.  I had some ZOTAC's on backorder and ended up cancelling those.  This will be my first NVIDIA rig build.
hopefully they rock! Our FTW2s have been our best 1070 TI, much better than the recent zotac 1070 TI mini's -- the ftw2 hashing 30-35 sols more w/ the same settings. The original zotac 1070 TI bought for testing performed well too though . . i sold that one xD

Hey Vosk,

I have my power down to 160W per card.  That's total power at the wall 960W divided by 6.  Power limit is set at 70%.  Core clock +200 and memory clock +700 on these FTW2's.  Averaging 525 sol/s per card.  When power was at 180W (80% limit in afterburner), I was getting a few more sol/s on each card.  I've never overclocked a gpu in my life, so this is making me slightly nervous ... Lol.  I'm not sure if I should keep going or not.  What do you have your FTW2's at?

Another question ... 3 of my FTW2's only have one fan running and I'm not sure why.  Any ideas?  I have my fan speed set to 80% in afterburner.  It's the back fan that's not running on all 3.

And there's more ...

I just set my fan speed to auto, and then back to 80% ... Now the only card with the back fan not running is the card on the end of my open air frame.  60 seconds later ... I have 3 back fans not running, but they aren't the same 3 cards as the first time.  Lol ... This is weird.  Do they not run constantly?  They'll just turn on when a certain temp is hit?

My FTW2 Hold about 500-505 Sol's each at +200 core and +750 mem. Watts are 110 per card, fan set minimum 50%, and target temp of 70C. They stay around 60-65C.

Also, I'm pretty sure only one fan will run if temp is not high enough to justify both running. Right now my three are siting at 59C 62C and 60C, only one has the second fan running. My guess is that it's the one at 62C. And as I type this my middle ones second fan kicked on. So don't worry about the fan not working if you have one working.


EDIT: I would try to get down to a lower watt setting. While an extra 25 Sol/s per card seems nice, it's only netting about .10 more a day after power consumption, .25 before.

When you're calculating 110w per card, is that with everything else factored in and measuring power consumption at the wall?
newbie
Activity: 25
Merit: 0
Good to know.  I have some EVGA FTW2 1070ti's coming on Friday ... (Hopefully) ... They are supposed to ship today.  I had some ZOTAC's on backorder and ended up cancelling those.  This will be my first NVIDIA rig build.
hopefully they rock! Our FTW2s have been our best 1070 TI, much better than the recent zotac 1070 TI mini's -- the ftw2 hashing 30-35 sols more w/ the same settings. The original zotac 1070 TI bought for testing performed well too though . . i sold that one xD

Hey Vosk,

I have my power down to 160W per card.  That's total power at the wall 960W divided by 6.  Power limit is set at 70%.  Core clock +200 and memory clock +700 on these FTW2's.  Averaging 525 sol/s per card.  When power was at 180W (80% limit in afterburner), I was getting a few more sol/s on each card.  I've never overclocked a gpu in my life, so this is making me slightly nervous ... Lol.  I'm not sure if I should keep going or not.  What do you have your FTW2's at?

Another question ... 3 of my FTW2's only have one fan running and I'm not sure why.  Any ideas?  I have my fan speed set to 80% in afterburner.  It's the back fan that's not running on all 3.

And there's more ...

I just set my fan speed to auto, and then back to 80% ... Now the only card with the back fan not running is the card on the end of my open air frame.  60 seconds later ... I have 3 back fans not running, but they aren't the same 3 cards as the first time.  Lol ... This is weird.  Do they not run constantly?  They'll just turn on when a certain temp is hit?

My FTW2 Hold about 500-505 Sol's each at +200 core and +750 mem. Watts are 110 per card, fan set minimum 50%, and target temp of 70C. They stay around 60-65C.

Also, I'm pretty sure only one fan will run if temp is not high enough to justify both running. Right now my three are siting at 59C 62C and 60C, only one has the second fan running. My guess is that it's the one at 62C. And as I type this my middle ones second fan kicked on. So don't worry about the fan not working if you have one working.


EDIT: I would try to get down to a lower watt setting. While an extra 25 Sol/s per card seems nice, it's only netting about .10 more a day after power consumption, .25 before.
member
Activity: 644
Merit: 24
Good to know.  I have some EVGA FTW2 1070ti's coming on Friday ... (Hopefully) ... They are supposed to ship today.  I had some ZOTAC's on backorder and ended up cancelling those.  This will be my first NVIDIA rig build.
hopefully they rock! Our FTW2s have been our best 1070 TI, much better than the recent zotac 1070 TI mini's -- the ftw2 hashing 30-35 sols more w/ the same settings. The original zotac 1070 TI bought for testing performed well too though . . i sold that one xD

Hey Vosk,

I have my power down to 160W per card.  That's total power at the wall 960W divided by 6.  Power limit is set at 70%.  Core clock +200 and memory clock +700 on these FTW2's.  Averaging 525 sol/s per card.  When power was at 180W (80% limit in afterburner), I was getting a few more sol/s on each card.  I've never overclocked a gpu in my life, so this is making me slightly nervous ... Lol.  I'm not sure if I should keep going or not.  What do you have your FTW2's at?

Another question ... 3 of my FTW2's only have one fan running and I'm not sure why.  Any ideas?  I have my fan speed set to 80% in afterburner.  It's the back fan that's not running on all 3.

And there's more ...

I just set my fan speed to auto, and then back to 80% ... Now the only card with the back fan not running is the card on the end of my open air frame.  60 seconds later ... I have 3 back fans not running, but they aren't the same 3 cards as the first time.  Lol ... This is weird.  Do they not run constantly?  They'll just turn on when a certain temp is hit?
newbie
Activity: 14
Merit: 0
hmmm my results are a little better on DSTM vs EWBF

https://youtu.be/oPX864SOUPE
https://i.imgur.com/Q5H0rFkm.png

I messed around with my latest zotac 1070 ti mini's for awhile and I never really got them above 4.3 sols/watt -- im no OC guru though / I'm at
175 core 700 memory and 110 watts on DSTM


Like your input and mindset here + by the time the new cards roll out I imagine everyone buying now could have made their money back lol so why wait Cheesy

I am getting about the same sol/watt as Vosk on all of my Zotac 1070ti AMP editions running DSTM (+200 core / +700 mem / 115w)(Micron memory). They have been running stable since I put this rig together. I have one more open slot on this rig I am debating on trying an EVGA card this time.
sr. member
Activity: 1414
Merit: 487
YouTube.com/VoskCoin
Anyone managed to take the settings further with even better results / esp on linux?

 Probably card dependent, and seems to at least in part be mining software dependent.

 It DOES look like the best efficiency point on EBWF is at 104 watts not 106, but it's so close it could be an artifact of measurement limits.

 I tried pushing core and memory harder on one of my EVGA SC cards, but stability went bad very quickly above the +200 core and +700 memory settings (which was an unreported part of my original testing).
 That however might vary with the specific card and probably with the model of GPU.


 I'm not waiting, as I'm not convinced that the Volta versions of the 10xx series cards will be a LOT more efficient.
 10% probably, 20% maybe, 30% I'd bet against - we're still on the same process node, just "process improvements" and perhaps some additional minor optimisations to the designs are all we're likely to see - think of "RX 4xx vs RX 5xx" level changes being the most likely.

 Hint - "12nm" isn't a new node, it's enhancements to a current 16nm node process with a "new name" as strictly an ADVERTISING point.


 Samsung does seem to have some VERY small quantity of production happening on 10nm - but strictly for low-power devices like their smartphones at this point, and they're eating ALL of the output so far.
 Intel is supposed to FINALLY be shipping their first 10nm products next year - but the date keeps slipping.....

 Manufacturing efficiency for both is probably going to be VERY poor on anything they manage to ship at 10nm next year - it takes TIME to dial a new process in, and it seems to take longer every time a legitimate "shrink" happens TO get the new process dialed in.


hmmm my results are a little better on DSTM vs EWBF

https://youtu.be/oPX864SOUPE


I messed around with my latest zotac 1070 ti mini's for awhile and I never really got them above 4.3 sols/watt -- im no OC guru though / I'm at
175 core 700 memory and 110 watts on DSTM


Like your input and mindset here + by the time the new cards roll out I imagine everyone buying now could have made their money back lol so why wait Cheesy
sr. member
Activity: 1414
Merit: 487
YouTube.com/VoskCoin
I have to pickup my MSI Gaming X 1070 Ti from the post office tomorrow. What a waste to have it there one night in the box haha
I was quite impressed by the results people posted here - will report back tomorrow!
how is it performing?
sr. member
Activity: 1414
Merit: 487
YouTube.com/VoskCoin
Good to know.  I have some EVGA FTW2 1070ti's coming on Friday ... (Hopefully) ... They are supposed to ship today.  I had some ZOTAC's on backorder and ended up cancelling those.  This will be my first NVIDIA rig build.
hopefully they rock! Our FTW2s have been our best 1070 TI, much better than the recent zotac 1070 TI mini's -- the ftw2 hashing 30-35 sols more w/ the same settings. The original zotac 1070 TI bought for testing performed well too though . . i sold that one xD
member
Activity: 644
Merit: 24
Good to know.  I have some EVGA FTW2 1070ti's coming on Friday ... (Hopefully) ... They are supposed to ship today.  I had some ZOTAC's on backorder and ended up cancelling those.  This will be my first NVIDIA rig build.
member
Activity: 336
Merit: 15
I have to pickup my MSI Gaming X 1070 Ti from the post office tomorrow. What a waste to have it there one night in the box haha
I was quite impressed by the results people posted here - will report back tomorrow!

Just plugged it in and had some issues at the start. The MSI Gaming X 1070 Ti is only recognised if you add the 8+6 PIN power connectors (opposed to the GB Gaming G1 Version).

Just ran some quick tests on Nist5 Algorithm and it hashed 2 MH/s more than the GB version with the same core clock and both cards pulling around 135W. So yes - not disappointed with it and keeping in mind that the fans on the MSI should last much longer I will only go for these cards in the future.
Pages:
Jump to: