bitlane (OP)
Internet detective
Sr. Member
Offline
Activity: 462
Merit: 250
I heart thebaron
|
|
June 02, 2012, 10:13:24 AM |
|
Now that there are quite a few 7970 owners out there, have any of you taken the time to check the ASIC Quality percentages from each of your cards ?
I currently have 15x 7970s running 5-card per rig x 3. Doing a few 'spot checks' I have noticed that there is a bit of a variance of ASIC Quality readings from some of my cards.
My lowest ASIC Quality % seems to be about 65+%, with the highest being just shy of 90%.....the rest in the high 70% / mid-to-low 80% range.
I am wondering if it would be worth my time testing and separating the 15 cards into my 3 rigs, based on ASIC Quality, to further 'fine-tune' under-voltaging and overclocking ? For instance, make a 60+% rig, a 70+% rig and finally an 80+% rig all based on what GPU-Z tells me.
Has anyone gone this deep into their setups yet ? ...or taken advantage of the ASIC Quality Readings now available ?
|
|
|
|
ssateneth
Legendary
Offline
Activity: 1344
Merit: 1004
|
|
June 02, 2012, 11:06:31 AM |
|
You should already be finding the max clock at the voltage you are comfortable with on a percard basis. Not doing so is just robbing you of better MH/j and more shares/minute
|
|
|
|
Buckwheet
Full Member
Offline
Activity: 169
Merit: 100
Green Mining 85% Cheaper
|
|
June 02, 2012, 10:08:25 PM |
|
Now that there are quite a few 7970 owners out there, have any of you taken the time to check the ASIC Quality percentages from each of your cards ?
I currently have 15x 7970s running 5-card per rig x 3. Doing a few 'spot checks' I have noticed that there is a bit of a variance of ASIC Quality readings from some of my cards.
My lowest ASIC Quality % seems to be about 65+%, with the highest being just shy of 90%.....the rest in the high 70% / mid-to-low 80% range.
I am wondering if it would be worth my time testing and separating the 15 cards into my 3 rigs, based on ASIC Quality, to further 'fine-tune' under-voltaging and overclocking ? For instance, make a 60+% rig, a 70+% rig and finally an 80+% rig all based on what GPU-Z tells me.
Has anyone gone this deep into their setups yet ? ...or taken advantage of the ASIC Quality Readings now available ?
I have returned or exchanged every one of the cards with very low ASIC quality. I have had nothing but problems with high temps and crashes on ones that are in the 60-70% range. I of course first contact the mfg and give them the temps and numbers I am seeing and they give me the okay to return them to the store. Also I found buying the overclocked editions usually resulted in higher ASIC quality and lower thermals. I am not sure if they are just binned that way or what, but my last batch of basic HIS cards were all returned because they were complete crap for thermals and ASICs. When I say crap, I mean 65C idle 85C playing Diablo 3.
|
|
|
|
aqrulesms
|
|
June 02, 2012, 10:15:05 PM |
|
ASIC quality?
I am a 7970 owner with 1x 7970 and 1x 5850 for mining.
What exactly is ASIC and ASIC quality?
|
|
|
|
|
BCMan
|
|
June 03, 2012, 02:18:25 AM |
|
"ASIC quality reading not supported on this card". That was for 6770, 6870 and 5770. So only 7xxx cards are supported?
|
|
|
|
aqrulesms
|
|
June 03, 2012, 02:46:58 AM |
|
"ASIC quality reading not supported on this card". That was for 6770, 6870 and 5770. So only 7xxx cards are supported? Yes from the description of the software I think only 7xxx and above are supported.
|
|
|
|
cmg5461
|
|
June 03, 2012, 03:10:12 AM |
|
I have 76.7% from an XFX reference black edition. Watercooled that baby ALSO: 28nm has huge leakage. Once those temps are up, power draw is huge. Hashing with the cash at 30C has given me a huge increase in MH/J
|
If I've helped: 1CmguJhwW4sbtSMFsyaafikJ8jhYS61quz
Sold: 5850 to lepenguin. Quick, easy and trustworthy.
|
|
|
goxed
Legendary
Offline
Activity: 1946
Merit: 1006
Bitcoin / Crypto mining Hardware.
|
|
June 03, 2012, 08:46:52 AM |
|
I have a Radeon 7970 with 56% ASIC quality It requires higher voltage 1.1V to run at the default clocks.
|
Revewing Bitcoin / Crypto mining Hardware.
|
|
|
goxed
Legendary
Offline
Activity: 1946
Merit: 1006
Bitcoin / Crypto mining Hardware.
|
|
June 03, 2012, 08:47:52 AM |
|
Hashing with the cash at 30C has given me a huge increase in MH/J
How much do you think is the gain in MH/J with water cooling?
|
Revewing Bitcoin / Crypto mining Hardware.
|
|
|
AzN1337c0d3r
Full Member
Offline
Activity: 238
Merit: 100
★YoBit.Net★ 350+ Coins Exchange & Dice
|
|
June 03, 2012, 10:47:19 AM |
|
Hashing with the cash at 30C has given me a huge increase in MH/J
How much do you think is the gain in MH/J with water cooling? https://bitcointalk.org/index.php?topic=57410.0He was using 232W/card at stock clocks on air, while using 191W/card at stock clocks on water. So about a saving of 40W, or if you rathar deal with MH/J numbers it's 2.37 MH/J (air) to 2.87 MH/J (water). If you pay 10c/kwh, the waterblocks save you about $35/year.
|
|
|
|
crazyates
Legendary
Offline
Activity: 952
Merit: 1000
|
|
June 03, 2012, 12:41:02 PM |
|
Hashing with the cash at 30C has given me a huge increase in MH/J
How much do you think is the gain in MH/J with water cooling? https://bitcointalk.org/index.php?topic=57410.0He was using 232W/card at stock clocks on air, while using 191W/card at stock clocks on water. So about a saving of 40W, or if you rathar deal with MH/J numbers it's 2.37 MH/J (air) to 2.87 MH/J (water). If you pay 10c/kwh, the waterblocks save you about $35/year. I still don't see how swapping out 1 fan for a watercooler saves 40W...
|
|
|
|
cmg5461
|
|
June 03, 2012, 02:23:43 PM |
|
Hashing with the cash at 30C has given me a huge increase in MH/J
How much do you think is the gain in MH/J with water cooling? https://bitcointalk.org/index.php?topic=57410.0He was using 232W/card at stock clocks on air, while using 191W/card at stock clocks on water. So about a saving of 40W, or if you rathar deal with MH/J numbers it's 2.37 MH/J (air) to 2.87 MH/J (water). If you pay 10c/kwh, the waterblocks save you about $35/year. I still don't see how swapping out 1 fan for a watercooler saves 40W... The fan has nothing to do with it. My fan used 10W while my pump uses ~30w (not sure) As transistors get hot, they consume more power. We're talking going from 70C to ~ 30C
|
If I've helped: 1CmguJhwW4sbtSMFsyaafikJ8jhYS61quz
Sold: 5850 to lepenguin. Quick, easy and trustworthy.
|
|
|
crazyates
Legendary
Offline
Activity: 952
Merit: 1000
|
|
June 03, 2012, 06:01:34 PM |
|
Hashing with the cash at 30C has given me a huge increase in MH/J
How much do you think is the gain in MH/J with water cooling? https://bitcointalk.org/index.php?topic=57410.0He was using 232W/card at stock clocks on air, while using 191W/card at stock clocks on water. So about a saving of 40W, or if you rathar deal with MH/J numbers it's 2.37 MH/J (air) to 2.87 MH/J (water). If you pay 10c/kwh, the waterblocks save you about $35/year. I still don't see how swapping out 1 fan for a watercooler saves 40W... The fan has nothing to do with it. My fan used 10W while my pump uses ~30w (not sure) As transistors get hot, they consume more power. We're talking going from 70C to ~ 30C Gotcha. That makes sense. Thanks
|
|
|
|
aqrulesms
|
|
June 04, 2012, 01:32:41 AM |
|
Hashing with the cash at 30C has given me a huge increase in MH/J
How much do you think is the gain in MH/J with water cooling? https://bitcointalk.org/index.php?topic=57410.0He was using 232W/card at stock clocks on air, while using 191W/card at stock clocks on water. So about a saving of 40W, or if you rathar deal with MH/J numbers it's 2.37 MH/J (air) to 2.87 MH/J (water). If you pay 10c/kwh, the waterblocks save you about $35/year. I still don't see how swapping out 1 fan for a watercooler saves 40W... Watercooling allows the GPU to run much cooler. If temperatures are high, resistance is also high. It requires more power to overcome that resistance. This therefore saves a good amount of energy.
|
|
|
|
|