Bitcoin Forum
January 30, 2023, 02:34:05 AM *
News: Latest Bitcoin Core release: 24.0.1 [Torrent]
 
   Home   Help Search Login Register More  
Pages: « 1 [2] 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 »
  Print  
Author Topic: GTX 1070ti mining  (Read 26794 times)
maxuscz
Member
**
Offline Offline

Activity: 118
Merit: 10


View Profile
November 03, 2017, 02:46:26 PM
Last edit: November 03, 2017, 04:03:30 PM by maxuscz
 #21

I got one today morning. Bad owrked on Win 7. Looked like PC overloaded. Slow scrypts. Slow mining soft reaciton . I mean it is bad current driver

On Win10 - worked but. Alexis10 -neoscrypt - hashrate half from my 1070 card. Don know where problem is.
Im just instal card to PC and start driver update from folder. Not install driver via setup cose I have another 6 1070 cards on this rig and dont wont change driver for they. But meibe it is problem

P.S. Only neoscrypt show problem. Better hashrate on klaust miner
but still less than 1070

All other algo - 1070ti show better result
1675046045
Hero Member
*
Offline Offline

Posts: 1675046045

View Profile Personal Message (Offline)

Ignore
1675046045
Reply with quote  #2

1675046045
Report to moderator
1675046045
Hero Member
*
Offline Offline

Posts: 1675046045

View Profile Personal Message (Offline)

Ignore
1675046045
Reply with quote  #2

1675046045
Report to moderator
You get merit points when someone likes your post enough to give you some. And for every 2 merit points you receive, you can send 1 merit point to someone else!
Advertised sites are not endorsed by the Bitcoin Forum. They may be unsafe, untrustworthy, or illegal in your jurisdiction.
1675046045
Hero Member
*
Offline Offline

Posts: 1675046045

View Profile Personal Message (Offline)

Ignore
1675046045
Reply with quote  #2

1675046045
Report to moderator
1675046045
Hero Member
*
Offline Offline

Posts: 1675046045

View Profile Personal Message (Offline)

Ignore
1675046045
Reply with quote  #2

1675046045
Report to moderator
QuintLeo
Legendary
*
Offline Offline

Activity: 1498
Merit: 1030


View Profile
November 03, 2017, 07:20:34 PM
 #22

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.




I'm no longer legendary just in my own mind!
Like something I said? Donations gratefully accepted. LYLnTKvLefz9izJFUvEGQEZzSkz34b3N6U (Litecoin)
1GYbjMTPdCuV7dci3iCUiaRrcNuaiQrVYY (Bitcoin)
heavyarms1912
Full Member
***
Offline Offline

Activity: 772
Merit: 114



View Profile
November 03, 2017, 08:04:31 PM
 #23

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.


what version of EBWF miner is that?
lunobird
Full Member
***
Offline Offline

Activity: 848
Merit: 115


View Profile
November 03, 2017, 10:58:17 PM
 #24

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.


Very Impressive.  Looks like if underclocked correclty it can be a very efficient miner.  I like and thanks for doing a deep dive on your settings and efficiency.

This 1080 gtx is $500 after rebate.  Would you pick your 1070 ti over this 1080 gtx?  Im on the fence on what to choose now. I would only be mining with equihash and other algos that's not ethereum.

I would like your pro miner opinion on this.

https://www.amazon.com/MSI-GTX-1080-ARMOR-8G/dp/B01GXOWUDQ/ref=sr_1_2?s=pc&ie=UTF8&qid=1509749825&sr=1-2&keywords=1080

philipma1957
Legendary
*
Offline Offline

Activity: 3654
Merit: 5222



View Profile
November 03, 2017, 11:48:42 PM
 #25

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.


Very Impressive.  Looks like if underclocked correclty it can be a very efficient miner.  I like and thanks for doing a deep dive on your settings and efficiency.

This 1080 gtx is $500 after rebate.  Would you pick your 1070 ti over this 1080 gtx?  Im on the fence on what to choose now. I would only be mining with equihash and other algos that's not ethereum.

I would like your pro miner opinion on this.

https://www.amazon.com/MSI-GTX-1080-ARMOR-8G/dp/B01GXOWUDQ/ref=sr_1_2?s=pc&ie=UTF8&qid=1509749825&sr=1-2&keywords=1080



I don't like any armor card.  but I never used that armor card.

I ordered 2 of these

https://www.evga.com/products/product.aspx?pn=08G-P4-5678-KR

3) Never invest more than you can afford to lose
2)Not your keys not your coins.
1) be like JJG just DCA
QuintLeo
Legendary
*
Offline Offline

Activity: 1498
Merit: 1030


View Profile
November 04, 2017, 12:06:02 AM
 #26

$469 for the EVGA SC I bought - there's a Zotac mini at $449.
Tough call vs the 1080 if you're pushing for high hashrate, but the 1070 ti is a better choice if you need to worry about efficiency.


I've got one of that Armor model - other than the 240 watt "default" TDP it's nothing special, cooling good but not great and otherwise it's right in the same pack with all of the other 1080 models I have (I don't have any Aorus or other HIGH END 1080 cards though).
I prefer the EVGA SC models overall - and I've sworn off that MSI "Aero" shortie 1070 card, the Gigabyte ITX blows it away across the board - clocks higher, cools better, but both has poor fans that will likely die too fast.



I'm no longer legendary just in my own mind!
Like something I said? Donations gratefully accepted. LYLnTKvLefz9izJFUvEGQEZzSkz34b3N6U (Litecoin)
1GYbjMTPdCuV7dci3iCUiaRrcNuaiQrVYY (Bitcoin)
TeamGreen
Newbie
*
Offline Offline

Activity: 52
Merit: 0


View Profile
November 04, 2017, 12:39:03 AM
 #27

Here are some early benchmarks for ZEC and ETH.

https://www.youtube.com/watch?v=2tQDcwZA7GQ

sevenmiles
Full Member
***
Offline Offline

Activity: 196
Merit: 100


View Profile
November 04, 2017, 03:58:55 AM
 #28

Here are some early benchmarks for ZEC and ETH.

https://www.youtube.com/watch?v=2tQDcwZA7GQ



Looks like 1070 ti could get 500 sol/s stable if overlock correctly and can also get the efficiency at about 5 sol/watts, pretty impressive on ZEC...
however, ETH is another story, i don;t see any improvement comparing to old 1070 cards, stock at 26 MH/s, and 32 MH/s overlocked...

while the prices is not that friendly for now:

EVGA GeForce GTX 1070 Ti SC GAMING ACX 3.0 Black Edition, 8GB GDDR5, EVGA OCX Scanner OC, White LED, DX12OSD Support (PXOC) Graphics Card 08G-P4-5671-KR
$ 469.99

just nope...

VoskCoin
Sr. Member
****
Offline Offline

Activity: 1414
Merit: 487


YouTube.com/VoskCoin


View Profile WWW
November 04, 2017, 04:07:32 AM
 #29

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.





Thanks for taking the time to do this , awesome write up

For 60% just to clarify what was your memory offset?

Check out my Crypto YouTube channel
https://www.youtube.com/VoskCoin
If you enjoy my content click Subscribe
QuintLeo
Legendary
*
Offline Offline

Activity: 1498
Merit: 1030


View Profile
November 04, 2017, 08:02:39 AM
 #30

60%, +200 core, +700 memory.

 Interesting that the card in that video had the SAME underclocked RAM mine did (same model card).
 Seems to be a batch issue not just my card.

 Their test bed was a lot higher end than than mine, not that it really mattered for a single card ZEC test rig.
 Afterburner version doesn't seem to matter much - I was using 4.4 on the test rig, they were using 4.3, and the final "production rig" the card is now in for me uses 4.2

 I AM starting to wonder if they looked at my review before they did their video - the numbers in the early part are SO....mine!

 9-)


I'm no longer legendary just in my own mind!
Like something I said? Donations gratefully accepted. LYLnTKvLefz9izJFUvEGQEZzSkz34b3N6U (Litecoin)
1GYbjMTPdCuV7dci3iCUiaRrcNuaiQrVYY (Bitcoin)
QuintLeo
Legendary
*
Offline Offline

Activity: 1498
Merit: 1030


View Profile
November 04, 2017, 08:33:59 AM
 #31

Here are some early benchmarks for ZEC and ETH.

https://www.youtube.com/watch?v=2tQDcwZA7GQ



Looks like 1070 ti could get 500 sol/s stable if overlock correctly and can also get the efficiency at about 5 sol/watts, pretty impressive on ZEC...
however, ETH is another story, i don;t see any improvement comparing to old 1070 cards, stock at 26 MH/s, and 32 MH/s overlocked...

while the prices is not that friendly for now:

EVGA GeForce GTX 1070 Ti SC GAMING ACX 3.0 Black Edition, 8GB GDDR5, EVGA OCX Scanner OC, White LED, DX12OSD Support (PXOC) Graphics Card 08G-P4-5671-KR
$ 469.99

just nope...

 Try 540 - go back and read my early test points.

 On a per-sol cost basis at a more efficient mining point, these cards are a hair BETTER than the 1070, and noticeably better than the 1080 or 1080 ti - and nothing AMD is even in the ballpark at CURRENT pricing.
 They're comparable when pushing for max hashrate as well to the other high-end NVidia 10 series cards on a hash/$ basis.

 For ETH, the only NVidia card that is even worth LOOKING AT is the 1060 at recent "gouge is receding" pricing - and even there the RX 470/480/570/580 has come down enough and availabilty has improved enough that the AMD option is better for ETH.


 They're efficient enough they have me rethinking my "future farm" planning.

I'm no longer legendary just in my own mind!
Like something I said? Donations gratefully accepted. LYLnTKvLefz9izJFUvEGQEZzSkz34b3N6U (Litecoin)
1GYbjMTPdCuV7dci3iCUiaRrcNuaiQrVYY (Bitcoin)
RealSwissMiner
Member
**
Offline Offline

Activity: 336
Merit: 15


View Profile WWW
November 04, 2017, 12:47:36 PM
 #32



These are my results. For all cards memory +500 and core +50. For the regular 1070 PL to 68%, for the 1070 TI PL to 70% (GPU0).

Lowering the PL further on the Ti has a big impact on the GPU frequency. The efficiency is better than for the regular 1070 cards and it hashes roughly 10% more.

Mining and investment blog - https://swisscryptominer.blogspot.ch
RealSwissMiner
Member
**
Offline Offline

Activity: 336
Merit: 15


View Profile WWW
November 04, 2017, 04:17:26 PM
 #33



These are my results. For all cards memory +500 and core +50. For the regular 1070 PL to 68%, for the 1070 TI PL to 70% (GPU0).

Lowering the PL further on the Ti has a big impact on the GPU frequency. The efficiency is better than for the regular 1070 cards and it hashes roughly 10% more.

Results on ETH are disappointing - with identical settings the regular 1070 hits around 31 MH/s while the Ti only does 27 MH/s although Afterburner shows the exact same memory speed across the board.

Mining and investment blog - https://swisscryptominer.blogspot.ch
QuintLeo
Legendary
*
Offline Offline

Activity: 1498
Merit: 1030


View Profile
November 04, 2017, 06:16:41 PM
 #34

I found that anything much lower than about 60% (106 watts) on my GTX 1070 ti started killing core clock frequency and performance in a hurry.

 Might vary a little depending on the specific card, but I doubt that much under 100 watts is going to be worth messing with on these cards.


I'm no longer legendary just in my own mind!
Like something I said? Donations gratefully accepted. LYLnTKvLefz9izJFUvEGQEZzSkz34b3N6U (Litecoin)
1GYbjMTPdCuV7dci3iCUiaRrcNuaiQrVYY (Bitcoin)
mettalmag
Legendary
*
Offline Offline

Activity: 1085
Merit: 1003


≡v≡


View Profile
November 04, 2017, 10:19:33 PM
 #35



 It won't outperform a GTX 1080 on Zcash - I'd guess more like 480-500 sol/s ballpark at optimal "full hash" settings.
 A hair fewer cores than the 1080, and the memory system isn't a big factor to Equihash algorithm.


im optimistic, i see it doing 550-600 sol/s


 Probability zero, 1080 cards don't manage that reliable if at all.

 FEWER CORES = LOWER HASHRATE on Equihash, it will hash SLOWER than a 1080 though probably not by a lot since it's only a FEW less cores.

 It also seems to be the same GPU, with one "compute unit" disabled, so it's very unlikely to overclock more than the 1080 to a measurable degree if at all.
 Rated TDP seems to be the same as the 1080 cards, which makes sense since the GPU is so close. I could see it turning out to be 170-175 watt range instead though for "reference" level cards.


 Equihash algo used for ZCash/ZEN/ZCL and suck is not nearly as memory limited as the algo used for ETH, or the 1080 would NOT outperform the 1070 by a fair bit and the 1080ti wouldn't blow both completely out of the water.


 I talk about ETH and ZEC and their clones because that's where the money is - in terms of "total GPUs mining" ETH has more GPUs working then everything else combined, with ZEC in a solid second at a little less than half that (perhaps a third if enough ZEC miners are using NVidia cards) then a LONG drop to whatever is third at the time.



My MSI Gamixg X 1080 easily gets 575-585 H/s at 75% power limit. I can get it over 600 H/s if I increase the power limit.



I think high end overclocked 1070 Ti's will get 32-33 MH/s on ETH and ~520 H/s on Equihash, depending on the model. Not worth it considering the 1080 is about the same price and AMD RX series cards are much more cost effective for ETH.
please tell me more about your 1080 and ab settings
splat1
Newbie
*
Offline Offline

Activity: 26
Merit: 0


View Profile
November 05, 2017, 12:04:01 AM
 #36

1070TI founders

80% power
200+ core
600+ mem

Ran through nicehash benchmarks (Not precise). without any real tweeking.  Claymore dual is almost exactly the same as the regular 1070's I have in this machine.  Lyra - Skunk - equihash seem to have some gains though.

Keccak .871GH/s
NeoScript 1.127 MH/s
Lyra2REv2 44.3 MH/s
Dagger (ethminer) 30.5 MH/s
Dagger (Claymore) 31.1 MH/s
DaggerDecred 30.2/0.302GH/s
DaggerSia 30.2/0302GH/s
Decred (ccminer) 3.086GH/s
CryptoNight 675H/s
Lbry 0.399GH/s
Equihash (excavator) 450.5 Sol/s 
Equihash (EWBF) 485 Sol/s  << 1070fe does 412 on same machine
Pascal (excavator) 1.195Gh/s
X11Ghost (ccminer) 6.889MH/s
Sia (ccminer) 2.002GH/s
Blake2s(ccminer) 4.455GH/s
Skunk (ccminer) 35.610MH/s
bucketofsocks
Member
**
Offline Offline

Activity: 85
Merit: 11


View Profile
November 05, 2017, 12:58:01 AM
 #37

I bought a 1070ti and a standard 1070 today both gigabyte G1s. I set them up in afterburner so they were running the same clocks and power limits and put them through paces and I have to say there were absolutely no surprises. The ti is a little better in everything (eth when dual mining, max single is about the same) uses negligibly more power, efficiency varies but is usually within 0.0x of one another. Groestl 1070 38 ti 41, dual mining dcr 1070 27/800 ti 29/888  zec 1070 452 ti 501 etc.
goo38
Newbie
*
Offline Offline

Activity: 49
Merit: 0


View Profile
November 05, 2017, 06:23:40 AM
 #38

I bought a 1070ti and a standard 1070 today both gigabyte G1s. I set them up in afterburner so they were running the same clocks and power limits and put them through paces and I have to say there were absolutely no surprises. The ti is a little better in everything (eth when dual mining, max single is about the same) uses negligibly more power, efficiency varies but is usually within 0.0x of one another. Groestl 1070 38 ti 41, dual mining dcr 1070 27/800 ti 29/888  zec 1070 452 ti 501 etc.

But can the ti play crysis while dual mining  Shocked  Grin
rednoW
Legendary
*
Offline Offline

Activity: 1514
Merit: 1003


View Profile
November 05, 2017, 07:11:48 AM
 #39

Please, can somebody test gtx1070ti with neoscrypt?
I was getting 1.4mhs with gtx1070 at max PL and alexis miner.
Wonder if TI will give 1.6 with moderate power consumption. If yes - I'll buy.
RealSwissMiner
Member
**
Offline Offline

Activity: 336
Merit: 15


View Profile WWW
November 05, 2017, 07:37:59 AM
 #40

Please, can somebody test gtx1070ti with neoscrypt?
I was getting 1.4mhs with gtx1070 at max PL and alexis miner.
Wonder if TI will give 1.6 with moderate power consumption. If yes - I'll buy.
With what minee should this this be done?

Mining and investment blog - https://swisscryptominer.blogspot.ch
Pages: « 1 [2] 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 »
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.19 | SMF © 2006-2009, Simple Machines Valid XHTML 1.0! Valid CSS!