Buy nice or buy twice. --scryptr
Words to live by.
Topic: Claymore's ZCash/BTG AMD GPU Miner v12.6 (Windows/Linux) - page 38. (Read 3839201 times)
50 TO 70 PERCENT LOAD IS OPTIMAL--
The PSU will degrade faster at higher load. That is the fact. A good PSU will work at 90% load, but it will wear out at a faster rate as the components degrade more rapidly. --scryptr
Optimal for what? It's possible to argue that if you won't unpack your PSU it will be even better for its longevity. However personal attachment to the quality of the PSU capacitors and/or whatever internal components may degrade with time should be outweighed by the practical needs and financial reasoning...
A PSU WILL EXPLODE--
Sitting next to a capacity loaded "quality" PSU when it explodes and pops smoke is frightening. I think I have more "bronze" paperweights than I need. I keep putting off stripping the good fans and scrap steel off of them for re-use and resale.
The fact is the fact. Your money is yours. Buy nice or buy twice. --scryptr
50 TO 70 PERCENT LOAD IS OPTIMAL--
The PSU will degrade faster at higher load. That is the fact. A good PSU will work at 90% load, but it will wear out at a faster rate as the components degrade more rapidly. --scryptr
Optimal for what? It's possible to argue that if you won't unpack your PSU it will be even better for its longevity. However personal attachment to the quality of the PSU capacitors and/or whatever internal components may degrade with time should be outweighed by the practical needs and financial reasoning...
EVGA MAKES GOOD EQUIPMENT--
And they stand by their warranties. There are other known, popular brands with "80+" labels that will simply die if placed at full load. A PSU will last longer if used at 50-70% of full load. The capacitors degrade more rapidly at full load.
After having several failures, I always purchase a PSU with a higher rating than my expected load. --scryptr
I see your points...yet not sure if I follow your logic.
There're lots of makers doing good quality PSUs, so buying one should not be that difficult. And investing in a top tier PSU will most likely save you money in long term.
Few days of cards being idle may cover the price difference and trying to keep poor quality PSU in pristine conditions by paying premium for larger output and then running it at 50-70% doesn't make sense either.
So I'd say the optimal PSU should be a quality one (good models from EVGA, Seasonic, Corsair, etc) and run at up to 90%(from the wall) of it's nominal capacity, which makes it about 80% actual load...
50 TO 70 PERCENT LOAD IS OPTIMAL--
The PSU will degrade faster at higher load. That is the fact. A good PSU will work at 90% load, but it will wear out at a faster rate as the components degrade more rapidly. --scryptr
EVGA MAKES GOOD EQUIPMENT--
And they stand by their warranties. There are other known, popular brands with "80+" labels that will simply die if placed at full load. A PSU will last longer if used at 50-70% of full load. The capacitors degrade more rapidly at full load.
After having several failures, I always purchase a PSU with a higher rating than my expected load. --scryptr
I see your points...yet not sure if I follow your logic.
There're lots of makers doing good quality PSUs, so buying one should not be that difficult. And investing in a top tier PSU will most likely save you money in long term.
Few days of cards being idle may cover the price difference and trying to keep poor quality PSU in pristine conditions by paying premium for larger output and then running it at 50-70% doesn't make sense either.
So I'd say the optimal PSU should be a quality one (good models from EVGA, Seasonic, Corsair, etc) and run at up to 90%(from the wall) of it's nominal capacity, which makes it about 80% actual load...
actually you are incorrect in the matters of that we don't pay for it, which we do pay for in the form of a dev fee. that said not much advancement can be made with amd and equihash, my question is why has he not reduced the dev fee considering amd gpus for the moment have hit there max, better yet why hasn't he removed the dev fee atleast with the zcash miner since he cannot make anymore optimizations in the code. which quite franky makes sense to do
you can turn off the dev fee
anyway till now i only had 1 dead psu with mining, the cx430
over wise i have all corsair series and 2 zalman they all load at 90% or some more
over wise i have all corsair series and 2 zalman they all load at 90% or some more
The optimal range of most PSU is 50-70 percent of the rated load.
simply not so. good quality psu usually performs at loads of 90% better than at 50%.
go and see some graphs at www.jonnyguru.com
fyi also psu output is the number on the box, not the consumption at the wall.
Not sure where you got that information, certainly not at jonnyguru. A PSU peak efficiency is at around 50% and it decreases as you get either lower or higher. No PSU will perform better at 90% than it does at 50%. The reason you can see a lot of tests at higher loads in jonnyguru is because that’s exactly where most PSUs have difficulty hitting their 80+ ratings.
As for your last statement I’m not sure what you mean, but the number on the box is the effective power the PSU can supposedly deliver to your PC components. As no PSU has 100% efficiency, it will draw more than that from the wall to effectively deliver what it says on the box. How much more will depend on its efficiency
PS: Just do a simple google image search for PSU efficiency curve
I agree with that. Highest efficiency is around 50% of the load.
Every psu with 80plus certificate is tested on this site, you can browse efficiency curves as well:
https://plugloadsolutions.com
https://plugloadsolutions.com/psu_reports/EVGA_SUPERNOVA%201000%20G3_1000W_ECOS%204750_Report.pdf
https://plugloadsolutions.com
https://plugloadsolutions.com/psu_reports/EVGA_SUPERNOVA%201000%20G3_1000W_ECOS%204750_Report.pdf
This is a typical picture. You are correct about the efficiency. Its often best at ~90% (for many reasons)
However talking about actually using 50% of your PSU's nominal capacity sounds like a completely absurd thing to do.
also to your second point, that's exactly what I meant. Just was typing from mobile via VNC, so it was kinda hard:)
(1KW DC pulls 1.1KW AC with efficiency of 0.9)
EVGA MAKES GOOD EQUIPMENT--
And they stand by their warranties. There are other known, popular brands with "80+" labels that will simply die if placed at full load. A PSU will last longer if used at 50-70% of full load. The capacitors degrade more rapidly at full load.
After having several failures, I always purchase a PSU with a higher rating than my expected load. --scryptr
https://s12.postimg.org/cxhj6qdn1/evga-g3-1000.png
This is a typical picture. You are correct about the efficiency. Its often best at ~90% (for many reasons)
However talking about actually using 50% of your PSU's nominal capacity sounds like a completely absurd thing to do.
also to your second point, that's exactly what I meant. Just was typing from mobile via VNC, so it was kinda hard:)
(1KW DC pulls 1.1KW AC with efficiency of 0.9)
This is a typical picture. You are correct about the efficiency. Its often best at ~90% (for many reasons)
However talking about actually using 50% of your PSU's nominal capacity sounds like a completely absurd thing to do.
also to your second point, that's exactly what I meant. Just was typing from mobile via VNC, so it was kinda hard:)
(1KW DC pulls 1.1KW AC with efficiency of 0.9)
Yes, you’re correct. I was not suggesting that you should buy a PSU to run it at 50%. However I would advice not to buy a PSU that you’ll run at 90% or more, not only efficiency starts to drop more significantly, but you also have to take into account degradation with aging
This is a typical picture. You are correct about the efficiency. Its often best at ~90% (for many reasons)
However talking about actually using 50% of your PSU's nominal capacity sounds like a completely absurd thing to do.
also to your second point, that's exactly what I meant. Just was typing from mobile via VNC, so it was kinda hard:)
(1KW DC pulls 1.1KW AC with efficiency of 0.9)
The optimal range of most PSU is 50-70 percent of the rated load.
simply not so. good quality psu usually performs at loads of 90% better than at 50%.
go and see some graphs at www.jonnyguru.com
fyi also psu output is the number on the box, not the consumption at the wall.
Not sure where you got that information, certainly not at jonnyguru. A PSU peak efficiency is at around 50% and it decreases as you get either lower or higher. No PSU will perform better at 90% than it does at 50%. The reason you can see a lot of tests at higher loads in jonnyguru is because that’s exactly where most PSUs have difficulty hitting their 80+ ratings.
As for your last statement I’m not sure what you mean, but the number on the box is the effective power the PSU can supposedly deliver to your PC components. As no PSU has 100% efficiency, it will draw more than that from the wall to effectively deliver what it says on the box. How much more will depend on its efficiency
PS: Just do a simple google image search for PSU efficiency curve
The optimal range of most PSU is 50-70 percent of the rated load.
simply not so. good quality psu usually performs at loads of 90% better than at 50%.
go and see some graphs at www.jonnyguru.com
fyi also psu output is the number on the box, not the consumption at the wall.
if it s a good 400w it should run a 390 with a low cpu
i run my 390 with an athlon 620 what use 90w on a corsair cx430 (died after 5 month) when i realize the 12v rail only pull out 336w when it was actually using close to 400w
i run 4x 270 on corsair 650 (600w on 12V) for a year before modify the ring (now i have watt meter, i test and it use 605w at the wall)
5 x 380 on a 1000w using 1100 at the wall ( now decrease power limit to keep under 1000)
now my 390 runs on a zalman 700w with an fx8300 oc and an 470 under clocked (using 690w) max 700w on 12V, i limited the power limit to keep under 700w at the wall
i run my 390 with an athlon 620 what use 90w on a corsair cx430 (died after 5 month) when i realize the 12v rail only pull out 336w when it was actually using close to 400w
i run 4x 270 on corsair 650 (600w on 12V) for a year before modify the ring (now i have watt meter, i test and it use 605w at the wall)
5 x 380 on a 1000w using 1100 at the wall ( now decrease power limit to keep under 1000)
now my 390 runs on a zalman 700w with an fx8300 oc and an 470 under clocked (using 690w) max 700w on 12V, i limited the power limit to keep under 700w at the wall
YOU MUST LIKE THE SMELL OF BURNING WIRES--
The optimal range of most PSU is 50-70 percent of the rated load. You are abusing your power supplies. --scryptr
AGREED! Unless you had one of those new fancy Seasonic PSUs that are prepared to go 100w overspec and still achieve an 80+ gold rating, and that would still not be optimal, for your current setup you should get at minimum an 850w PSU, and go with a more reputable brand than Zalman.
With an 850w you would still not be in the optimal range of a PSU, but at least it wouldn’t be a fire hazard
if it s a good 400w it should run a 390 with a low cpu
i run my 390 with an athlon 620 what use 90w on a corsair cx430 (died after 5 month) when i realize the 12v rail only pull out 336w when it was actually using close to 400w
i run 4x 270 on corsair 650 (600w on 12V) for a year before modify the ring (now i have watt meter, i test and it use 605w at the wall)
5 x 380 on a 1000w using 1100 at the wall ( now decrease power limit to keep under 1000)
now my 390 runs on a zalman 700w with an fx8300 oc and an 470 under clocked (using 690w) max 700w on 12V, i limited the power limit to keep under 700w at the wall
i run my 390 with an athlon 620 what use 90w on a corsair cx430 (died after 5 month) when i realize the 12v rail only pull out 336w when it was actually using close to 400w
i run 4x 270 on corsair 650 (600w on 12V) for a year before modify the ring (now i have watt meter, i test and it use 605w at the wall)
5 x 380 on a 1000w using 1100 at the wall ( now decrease power limit to keep under 1000)
now my 390 runs on a zalman 700w with an fx8300 oc and an 470 under clocked (using 690w) max 700w on 12V, i limited the power limit to keep under 700w at the wall
YOU MUST LIKE THE SMELL OF BURNING WIRES--
The optimal range of most PSU is 50-70 percent of the rated load. You are abusing your power supplies. --scryptr
if it s a good 400w it should run a 390 with a low cpu
i run my 390 with an athlon 620 what use 90w on a corsair cx430 (died after 5 month) when i realize the 12v rail only pull out 336w when it was actually using close to 400w
i run 4x 270 on corsair 650 (600w on 12V) for a year before modify the ring (now i have watt meter, i test and it use 605w at the wall)
5 x 380 on a 1000w using 1100 at the wall ( now decrease power limit to keep under 1000)
now my 390 runs on a zalman 700w with an fx8300 oc and an 470 under clocked (using 690w) max 700w on 12V, i limited the power limit to keep under 700w at the wall
i run my 390 with an athlon 620 what use 90w on a corsair cx430 (died after 5 month) when i realize the 12v rail only pull out 336w when it was actually using close to 400w
i run 4x 270 on corsair 650 (600w on 12V) for a year before modify the ring (now i have watt meter, i test and it use 605w at the wall)
5 x 380 on a 1000w using 1100 at the wall ( now decrease power limit to keep under 1000)
now my 390 runs on a zalman 700w with an fx8300 oc and an 470 under clocked (using 690w) max 700w on 12V, i limited the power limit to keep under 700w at the wall
Hey guys,
I have a problem which I'm troubleshooting for days, and still not solved.
I got R9 280X MSI 3GB, 12GB of RAM, i3 4160, ASUS H81M-E, PSU ThermalTake 400W
On this setup my last gpu 7870 2GB worked perfectly on NiceHash mining ZCash on Claymore's miner, when I set up 280X, there are problems income..
First problem was GPU #0 returned incorrect data! on Windows 10 WITHOUT OC, everything stock. I also tryed many drivers, 15.12, then lastest, everything the same..
Now I got another HDD and installed Windows 7 64bit(Basic theme, not aero), got NiceHash setted up, Claymore's ZCash miner, 15.12 driver for win7, and still the same. Then with lastest driver, same situation. Here I saw another problem, monitor goes off for 1 second and it says "Display driver stopped responding and has recovered".
Can anyone help me, what to do ? I tryed to underclock core and memory, It worked for about 6 hours without error and then it came again, I can't let it run for 24/7 without assistance. Also lower clocks get me about -20.000 sols than stock.
Can anyone at least give me recommendation for batch file for my GPU in this case ?
I have a problem which I'm troubleshooting for days, and still not solved.
I got R9 280X MSI 3GB, 12GB of RAM, i3 4160, ASUS H81M-E, PSU ThermalTake 400W
On this setup my last gpu 7870 2GB worked perfectly on NiceHash mining ZCash on Claymore's miner, when I set up 280X, there are problems income..
First problem was GPU #0 returned incorrect data! on Windows 10 WITHOUT OC, everything stock. I also tryed many drivers, 15.12, then lastest, everything the same..
Now I got another HDD and installed Windows 7 64bit(Basic theme, not aero), got NiceHash setted up, Claymore's ZCash miner, 15.12 driver for win7, and still the same. Then with lastest driver, same situation. Here I saw another problem, monitor goes off for 1 second and it says "Display driver stopped responding and has recovered".
Can anyone help me, what to do ? I tryed to underclock core and memory, It worked for about 6 hours without error and then it came again, I can't let it run for 24/7 without assistance. Also lower clocks get me about -20.000 sols than stock.
Can anyone at least give me recommendation for batch file for my GPU in this case ?
You're too optimistic in regard of PSU. Check what wattage it is capable of for 12V line and if there are more 12V lines You should balance power for R9280X. It is power hungry card in comparison to undervolted 7870.
Yes, hadn’t noticed that, check your PSU. A 400w PSU is probably not enough, except if it is single rail, and even then it’s pretty close.
I have a 375w PSU and my 7970 won’t even try to turn on with that PSU. I also have a very power hungry cpu..
I'm having some weird issues with latest versions - the application occasionally crashes in Win10. While running 12.2 it hashes at the same rate and doesn't crash...
The rig is a mixed bunch of 4*470 and a single 370(in slot)
The rig is a mixed bunch of 4*470 and a single 370(in slot)
Hey guys,
I have a problem which I'm troubleshooting for days, and still not solved.
I got R9 280X MSI 3GB, 12GB of RAM, i3 4160, ASUS H81M-E, PSU ThermalTake 400W
On this setup my last gpu 7870 2GB worked perfectly on NiceHash mining ZCash on Claymore's miner, when I set up 280X, there are problems income..
First problem was GPU #0 returned incorrect data! on Windows 10 WITHOUT OC, everything stock. I also tryed many drivers, 15.12, then lastest, everything the same..
Now I got another HDD and installed Windows 7 64bit(Basic theme, not aero), got NiceHash setted up, Claymore's ZCash miner, 15.12 driver for win7, and still the same. Then with lastest driver, same situation. Here I saw another problem, monitor goes off for 1 second and it says "Display driver stopped responding and has recovered".
Can anyone help me, what to do ? I tryed to underclock core and memory, It worked for about 6 hours without error and then it came again, I can't let it run for 24/7 without assistance. Also lower clocks get me about -20.000 sols than stock.
Can anyone at least give me recommendation for batch file for my GPU in this case ?
I have a problem which I'm troubleshooting for days, and still not solved.
I got R9 280X MSI 3GB, 12GB of RAM, i3 4160, ASUS H81M-E, PSU ThermalTake 400W
On this setup my last gpu 7870 2GB worked perfectly on NiceHash mining ZCash on Claymore's miner, when I set up 280X, there are problems income..
First problem was GPU #0 returned incorrect data! on Windows 10 WITHOUT OC, everything stock. I also tryed many drivers, 15.12, then lastest, everything the same..
Now I got another HDD and installed Windows 7 64bit(Basic theme, not aero), got NiceHash setted up, Claymore's ZCash miner, 15.12 driver for win7, and still the same. Then with lastest driver, same situation. Here I saw another problem, monitor goes off for 1 second and it says "Display driver stopped responding and has recovered".
Can anyone help me, what to do ? I tryed to underclock core and memory, It worked for about 6 hours without error and then it came again, I can't let it run for 24/7 without assistance. Also lower clocks get me about -20.000 sols than stock.
Can anyone at least give me recommendation for batch file for my GPU in this case ?
You're too optimistic in regard of PSU. Check what wattage it is capable of for 12V line and if there are more 12V lines You should balance power for R9280X. It is power hungry card in comparison to undervolted 7870.
Jump to: