Overclock.net banner

261 - 280 of 288 Posts

·
Watercooler
Joined
·
3,367 Posts

·
Typo Elemental
Joined
·
3,222 Posts
Discussion Starter #262
@simmons, I never did get my premium account from last month, can you help me with that seeing as how I didn't qualify this month for a prize?
Okay so firstly, if you want to @ me, make sure you @ the right person :p

In regards to the OC'd account, send a message to ENTERPRISE. He is the one who handles that.

And yes, because you won last month, you were inelligble for prizes for this month and next month. In the past we have had individuals who won prizes several months in a row, so this is just to share the wealth, so to speak.

I never got the January cash prize that I won... the prizes must be a little slow this year
I wish there was something that I could do about that. For you and everyone else that hasn't received their prizes, please send a message to ENTERPRISE. He's the one who facilitates the prizes between OCN and VS.
 

·
Premium Member
Joined
·
113 Posts
If it helps you would be surprised with a wall meter exactly how much power your computer pushes whilst folding.
9900K off with 2080Ti its pushing up to 600W. Insanity. Software was nowhere near accurate.
So a couple of thoughts / questions.

1. I suspect the UPS is more accurate than software and similar to a wall meter as I am thinking it is measuring the draw on the plug at the external UPS and then networking that actual measurement to the OS / software. That said I am curious if anyone has compared UPS power measurements to dedicated wall meters?

2. (As I am planning a 4 cards rig) I am curious how much the efficiency of the power supply affects the wall current draw verses the cards draw. In theory the peak draw for a card would be something like TDP * Peak Load as % of TDP * (1/Power Supply Efficiency)

For example my card has a TDP of 160. HWInfo typically shows a max draw at 115% of rated TDP. My power supply is rated 90% efficient at 100% draw. For other brands that range can drop to 80%. So a couple of sample calcs could look like

Wall Draw with 90% PS = 160 * 1.15 * (1/0.9) = 204 Watts
Wall Draw with 80% PS = 160 * 1.15 * (1/0.8) = 230 Watts

If you take that same card on the 90% Power Supply and drop load to 50% then the efficiency goes to 94% or
Wall Draw with 90% PS = 160 * 1.15 * (1/0.94) = 195 Watts

So in theory the same card under the same load can have an ~18% difference in Wall Draw based of the efficiency of the power supply. I haven't done wall tests on this yet, but have been thinking about it with planning this new rig and will probably do some tests. I'd be interested to hear what everyone else has found in practice, but there is definitely a significant component of wall current / power draw that is related to the power supply and it's current total load.

Note: base efficiency numbers pulled from https://en.wikipedia.org/wiki/80_Plus
 

·
Rig Advisor
Joined
·
2,232 Posts
Okay so firstly, if you want to @ me, make sure you @ the right person :p

In regards to the OC'd account, send a message to ENTERPRISE. He is the one who handles that.

And yes, because you won last month, you were inelligble for prizes for this month and next month. In the past we have had individuals who won prizes several months in a row, so this is just to share the wealth, so to speak.



I wish there was something that I could do about that. For you and everyone else that hasn't received their prizes, please send a message to ENTERPRISE. He's the one who facilitates the prizes between OCN and VS.
Eeek, my bad on the tag, yea I knew I was ineligible for the next 2 FaT's.
 

·
Premium Member
Joined
·
2,416 Posts
3 WU's wouldn't send last night, 3/4 million points wasted. The 3.21.157.11 Collection server is fried.

I was wondering if that was me. Had a few WUs on different computers that were refusing to send. Ah well.
 

·
Typo Elemental
Joined
·
3,222 Posts
Discussion Starter #266
Eeek, my bad on the tag, yea I knew I was ineligible for the next 2 FaT's.
No worries mate, Just making sure everything was clear.


I was wondering if that was me. Had a few WUs on different computers that were refusing to send. Ah well.
Make sure you aren't stuck in fail-send timer hell. My one workstation had a 4 hour Time-To-Retry when I finally got to it last night.
 

·
Rig Advisor
Joined
·
2,232 Posts
No worries mate, Just making sure everything was clear.




Make sure you aren't stuck in fail-send timer hell. My one workstation had a 4 hour Time-To-Retry when I finally got to it last night.
I am acutely aware of the timer's. The problem is that unlike a WU work server, the collection server will not change when it has been assigned. If there is an issue with a specific collection server like the one I mentioned, there is nothing you can do about. I checked the Forums over at [email protected] and there are multiple complaints about that particular server.
 

·
OC...D
Joined
·
27,802 Posts
I just keep folding around the dead server. So far the downloads have not stalled due to the borked WUs.
 

Attachments

·
Rig Advisor
Joined
·
2,232 Posts
I just keep folding around the dead server. So far the downloads have not stalled due to the borked WUs.
I know Jpm, but no thanks. I'll give it a day or so.
 

·
Watercooler
Joined
·
3,367 Posts
So a couple of thoughts / questions.

1. I suspect the UPS is more accurate than software and similar to a wall meter as I am thinking it is measuring the draw on the plug at the external UPS and then networking that actual measurement to the OS / software. That said I am curious if anyone has compared UPS power measurements to dedicated wall meters?

2. (As I am planning a 4 cards rig) I am curious how much the efficiency of the power supply affects the wall current draw verses the cards draw. In theory the peak draw for a card would be something like TDP * Peak Load as % of TDP * (1/Power Supply Efficiency)

For example my card has a TDP of 160. HWInfo typically shows a max draw at 115% of rated TDP. My power supply is rated 90% efficient at 100% draw. For other brands that range can drop to 80%. So a couple of sample calcs could look like

Wall Draw with 90% PS = 160 * 1.15 * (1/0.9) = 204 Watts
Wall Draw with 80% PS = 160 * 1.15 * (1/0.8) = 230 Watts

If you take that same card on the 90% Power Supply and drop load to 50% then the efficiency goes to 94% or
Wall Draw with 90% PS = 160 * 1.15 * (1/0.94) = 195 Watts

So in theory the same card under the same load can have an ~18% difference in Wall Draw based of the efficiency of the power supply. I haven't done wall tests on this yet, but have been thinking about it with planning this new rig and will probably do some tests. I'd be interested to hear what everyone else has found in practice, but there is definitely a significant component of wall current / power draw that is related to the power supply and it's current total load.

Note: base efficiency numbers pulled from https://en.wikipedia.org/wiki/80_Plus
I might have to look into a new power supply i think, My PSU is pushing on about 8 years of service. Hasn't had a hiccup though... if it helps with efficiency i'll wait till corona and buy a new one.
In regards to this i assume that being just idling with 1 core usage for GPU pushes the CPU to over 100W on it's own, then the GPU gets close to about 350W on average. Plus HDD's, Ram etc... I guess it wouldn't be to ludicrous.
 

·
Premium Member
Joined
·
113 Posts
I might have to look into a new power supply i think, My PSU is pushing on about 8 years of service. Hasn't had a hiccup though... if it helps with efficiency i'll wait till corona and buy a new one.
In regards to this i assume that being just idling with 1 core usage for GPU pushes the CPU to over 100W on it's own, then the GPU gets close to about 350W on average. Plus HDD's, Ram etc... I guess it wouldn't be to ludicrous.
Before you invest in a new one try to find the specs on your current one and see if you can determine what it's current efficiency is compared to a new one. Then you can run you KWH savings over a year and figure in your power rate and determine your break even point. Local utility rates vary quite a bit in different markets which really impact the cost analysis if by efficiency you are looking at Point/$ verses just being green and looking at Points/KWH or Points/Ton of CO2.
 

·
Rig Advisor
Joined
·
2,232 Posts
Before you invest in a new one try to find the specs on your current one and see if you can determine what it's current efficiency is compared to a new one. Then you can run you KWH savings over a year and figure in your power rate and determine your break even point. Local utility rates vary quite a bit in different markets which really impact the cost analysis if by efficiency you are looking at Point/$ verses just being green and looking at Points/KWH or Points/Ton of CO2.
^This. It is only 0.11c per KWh in my area. It's not much to run a folding rig. You said your costs were 26$ for a week. What are your rates?
 

·
Premium Member
Joined
·
113 Posts
^This. It is only 0.11c per KWh in my area. It's not much to run a folding rig. You said your costs were 26$ for a week. What are your rates?
My rate I had calculate from my last bill is 10.4 c (USD) per KWh.

So with 2 cards at 115 Watts Average + 300 Watts CPU + 14 * 10 Watts per fan + 35 Watts for Pump = 705 Watts Load on PSU

So that puts my PSU in the ~92% efficient range so my Cost per day (CPD) would look similar to

CPD = ((705 / 1000) * 24 * 0.104) / 0.92
or
= KW * hours per day * Electric Rate in KWH / PSU efficiency
= $1.91 / day in USD
= $57 / month (30 days) in USD

Note 1: Fan & pump numbers are made up not measured
Note 2: CPU draw is based on 180% TDP which is max draw on my CPU at current OC and I don't have a good average so this is over estimating hourly draw as this should be based on average and not max watts

While I don't know @Socks keep you warm electric rate I did find some numbers that electricity in Victoria is ~0.28 I'm assuming in AUD.

No idea if this is accurate, but will illustrate the difference in where a rig is run none the less. So moving my rig to this Victoria rate

CPD = ((705 / 1000) * 24 * 0.28) / 0.92
= 5.15 / day in AUD = 3.34 in USD
= 154 / month (30 days) in AUD = 100.02 in USD

So a few thoughts that came to mind running through this

1. Currency Conversion from AUD to USD are significant as the current exchange rate is 1 to 0.65, which depending on the direction of conversion can appear as a ~54% difference
2. Local electric rates vary widely based on what I found were looking at up to a 250% difference
3. For cost calculations draw should be done in Averages so @Avacado's UPS numbers showed average which is the correct number. I'm not sure duration @socks-keep-you-warm kill a watt meter measured, but based on the variances between WUs I would think it would be more accurate to use it to measure total KWs over at least a few days and then divide this out to get KWh which would give a better average for different WUs and idle time. A short average over a period of heavy draw would inflate the numbers.

I'm curious and will see if I can get my UPS networked and get some actual average draw numbers from my rig to see how high my above guesstimate is. I'm pretty sure I didn't see a $57 dollar bump when I started folding.
 

·
Rig Advisor
Joined
·
2,232 Posts
My rate I had calculate from my last bill is 10.4 c (USD) per KWh.

So with 2 cards at 115 Watts Average + 300 Watts CPU + 14 * 10 Watts per fan + 35 Watts for Pump = 705 Watts Load on PSU

So that puts my PSU in the ~92% efficient range so my Cost per day (CPD) would look similar to

CPD = ((705 / 1000) * 24 * 0.104) / 0.92
or
= KW * hours per day * Electric Rate in KWH / PSU efficiency
= $1.91 / day in USD
= $57 / month (30 days) in USD

Note 1: Fan & pump numbers are made up not measured
Note 2: CPU draw is based on 180% TDP which is max draw on my CPU at current OC and I don't have a good average so this is over estimating hourly draw as this should be based on average and not max watts

While I don't know @Socks keep you warm electric rate I did find some numbers that electricity in Victoria is ~0.28 I'm assuming in AUD.

No idea if this is accurate, but will illustrate the difference in where a rig is run none the less. So moving my rig to this Victoria rate

CPD = ((705 / 1000) * 24 * 0.28) / 0.92
= 5.15 / day in AUD = 3.34 in USD
= 154 / month (30 days) in AUD = 100.02 in USD

So a few thoughts that came to mind running through this

1. Currency Conversion from AUD to USD are significant as the current exchange rate is 1 to 0.65, which depending on the direction of conversion can appear as a ~54% difference
2. Local electric rates vary widely based on what I found were looking at up to a 250% difference
3. For cost calculations draw should be done in Averages so @Avacado's UPS numbers showed average which is the correct number. I'm not sure duration @socks-keep-you-warm kill a watt meter measured, but based on the variances between WUs I would think it would be more accurate to use it to measure total KWs over at least a few days and then divide this out to get KWh which would give a better average for different WUs and idle time. A short average over a period of heavy draw would inflate the numbers.

I'm curious and will see if I can get my UPS networked and get some actual average draw numbers from my rig to see how high my above guesstimate is. I'm pretty sure I didn't see a $57 dollar bump when I started folding.
The 1.97$ a day seems about right on.
 

·
OC...D
Joined
·
27,802 Posts
you guys better not have daughters with hair dryers. :laughings
 

·
Rig Advisor
Joined
·
2,232 Posts
@Jpmboy This little overclocker won't need one for a few years!
 

Attachments

·
Watercooler
Joined
·
3,367 Posts
^This. It is only 0.11c per KWh in my area. It's not much to run a folding rig. You said your costs were 26$ for a week. What are your rates?
31c KW/H is what i set it too! Australia is a very expensive place to live. It's actually 51c KW/H but i get minus 20c because i get paid back from solar. I won't know the real cost until i see the next bill.
 

·
Premium Member
Joined
·
113 Posts
31c KW/H is what i set it too! Australia is a very expensive place to live. It's actually 51c KW/H but i get minus 20c because i get paid back from solar. I won't know the real cost until i see the next bill.
I have an old friend that moved to Brisbane probably 10 years ago and he says the same thing about cost. I know it was higher, but had no idea it was still around 3 times higher when accounting for currency.

you guys better not have daughters with hair dryers. :laughings
Or wives, I think hers is around 1800 watts and doubles as a space heater it's cold. :wheee:
 

·
Rig Advisor
Joined
·
2,232 Posts
This is the **** that makes me think they limit WU's based on IP address. see below!
 

Attachments

261 - 280 of 288 Posts
Top