giagge
Legendary
Offline
Activity: 1134
Merit: 1001
|
|
November 30, 2016, 02:32:17 PM |
|
980 watt x 900hs at 3300$ ? .
|
|
|
|
scryptr
Legendary
Offline
Activity: 1797
Merit: 1028
|
|
November 30, 2016, 02:53:13 PM |
|
980 watt x 900hs at 3300$ ? . EIGHT 280X CARDS WILL BEAT THAT-- I have two 4-card 280X rigs that have already paid for themselves at least twice. They hash at over 700Sols/s per rig. The wattage for both rigs together is likely more than 980W, I heven't measured. If they were purchased used, with two 4-PCIe slot motherboards, and PSUs, the total cost would be less than $1600. I don't see the advantage. --scryptr
|
|
|
|
abctoz
Newbie
Offline
Activity: 11
Merit: 0
|
|
November 30, 2016, 06:34:07 PM |
|
I think 1070 loses efficiency powering the 8GB of memory, the 1060 3GB should be best, except for in ethereum.
Probably, but it's hard to gauge just by how much. I think the number is miniscule and it's nice not to get locked with 3-4GB in the future as I'm not planning on selling 1xxx series cards in less than well over a year. I could only find estimates on memory power consumption which vary wildly: http://www.anandtech.com/show/10193/micron-begins-to-sample-gddr5x-memory (1/3 of the page) Based on other sources older cards consumed 4.35 W per gigabyte of GDDR5 (17.4W for 4GB and 34.8W for 8GB) but the same source says it's closer to 20W or slightly more for 8GB. There are also much higher numbers, like 50W for an R9 290X. isn't it watt per die instead of eg. 1 GB I don't know. Maybe that's one of the reason why the numbers are so different. It would be per die... The amount of power memory consumes is pretty small compared to the entirety of the card. Some people also forget that Ethereum is not the only coin in existence and building your ecosystem around one coin that will be getting hit hard in the coming months is not such a great idea. Nvidia is also not responsible for the kind of memory manufacturers use. That would be up to manufacturers such as MSI/Gigabyte/Asus. Ethereum continues to give the largest % of mining revenue by far, without it we can all throw our cards away. Also 30W/card would not be negligible. Only NVIDIA pulls this memory switching shit every generation, your argument fails because AMD and its AIB partners don't do this.
|
|
|
|
|
bensam1231
Legendary
Offline
Activity: 1764
Merit: 1024
|
|
December 01, 2016, 01:57:49 AM |
|
I think 1070 loses efficiency powering the 8GB of memory, the 1060 3GB should be best, except for in ethereum.
Probably, but it's hard to gauge just by how much. I think the number is miniscule and it's nice not to get locked with 3-4GB in the future as I'm not planning on selling 1xxx series cards in less than well over a year. I could only find estimates on memory power consumption which vary wildly: http://www.anandtech.com/show/10193/micron-begins-to-sample-gddr5x-memory (1/3 of the page) Based on other sources older cards consumed 4.35 W per gigabyte of GDDR5 (17.4W for 4GB and 34.8W for 8GB) but the same source says it's closer to 20W or slightly more for 8GB. There are also much higher numbers, like 50W for an R9 290X. isn't it watt per die instead of eg. 1 GB I don't know. Maybe that's one of the reason why the numbers are so different. It would be per die... The amount of power memory consumes is pretty small compared to the entirety of the card. Some people also forget that Ethereum is not the only coin in existence and building your ecosystem around one coin that will be getting hit hard in the coming months is not such a great idea. Nvidia is also not responsible for the kind of memory manufacturers use. That would be up to manufacturers such as MSI/Gigabyte/Asus. Ethereum continues to give the largest % of mining revenue by far, without it we can all throw our cards away. Also 30W/card would not be negligible. Only NVIDIA pulls this memory switching shit every generation, your argument fails because AMD and its AIB partners don't do this. Ethereum is going away this Januaryish as it shifts to PoS. Only Nvidia switches memory? lol... When I mined with AMD I had three different manufacturers. Hynix, Elpedia, and Samsung from four different card manufacturers. Some changing between the same cards and models. They do it because they find a better partner that can source memory for cheaper or better to them. Mining is a small drop in the bucket when it comes to the entirety of GPUs. As I mentioned AMD/Nvidia have nothing to do with who the memory is sourced from, only the type that is used (GDDR5/X, HBM). That's all the card manufacturers.
|
I buy private Nvidia miners. Send information and/or inquiries to my PM box.
|
|
|
klondike_bar
Legendary
Offline
Activity: 2128
Merit: 1005
ASIC Wannabe
|
|
December 01, 2016, 02:33:54 AM |
|
I think 1070 loses efficiency powering the 8GB of memory, the 1060 3GB should be best, except for in ethereum.
Probably, but it's hard to gauge just by how much. I think the number is miniscule and it's nice not to get locked with 3-4GB in the future as I'm not planning on selling 1xxx series cards in less than well over a year. I could only find estimates on memory power consumption which vary wildly: http://www.anandtech.com/show/10193/micron-begins-to-sample-gddr5x-memory (1/3 of the page) Based on other sources older cards consumed 4.35 W per gigabyte of GDDR5 (17.4W for 4GB and 34.8W for 8GB) but the same source says it's closer to 20W or slightly more for 8GB. There are also much higher numbers, like 50W for an R9 290X. isn't it watt per die instead of eg. 1 GB I don't know. Maybe that's one of the reason why the numbers are so different. It would be per die... The amount of power memory consumes is pretty small compared to the entirety of the card. Some people also forget that Ethereum is not the only coin in existence and building your ecosystem around one coin that will be getting hit hard in the coming months is not such a great idea. Nvidia is also not responsible for the kind of memory manufacturers use. That would be up to manufacturers such as MSI/Gigabyte/Asus. Ethereum continues to give the largest % of mining revenue by far, without it we can all throw our cards away. Also 30W/card would not be negligible. Only NVIDIA pulls this memory switching shit every generation, your argument fails because AMD and its AIB partners don't do this. Ethereum is going away this Januaryish as it shifts to PoS. Only Nvidia switches memory? lol... When I mined with AMD I had three different manufacturers. Hynix, Elpedia, and Samsung from four different card manufacturers. Some changing between the same cards and models. They do it because they find a better partner that can source memory for cheaper or better to them. Mining is a small drop in the bucket when it comes to the entirety of GPUs. As I mentioned AMD/Nvidia have nothing to do with who the memory is sourced from, only the type that is used (GDDR5/X, HBM). That's all the card manufacturers. ethereum originally planned to go PoS by summer 2016. then fall. now early 2017.dont hold your breath - IMO the devs and those close to the project are mining it, and advertising a PoS switch "soon" is an excellent deterrent to others from building farms
|
|
|
|
bensam1231
Legendary
Offline
Activity: 1764
Merit: 1024
|
|
December 01, 2016, 03:51:58 AM |
|
I think 1070 loses efficiency powering the 8GB of memory, the 1060 3GB should be best, except for in ethereum.
Probably, but it's hard to gauge just by how much. I think the number is miniscule and it's nice not to get locked with 3-4GB in the future as I'm not planning on selling 1xxx series cards in less than well over a year. I could only find estimates on memory power consumption which vary wildly: http://www.anandtech.com/show/10193/micron-begins-to-sample-gddr5x-memory (1/3 of the page) Based on other sources older cards consumed 4.35 W per gigabyte of GDDR5 (17.4W for 4GB and 34.8W for 8GB) but the same source says it's closer to 20W or slightly more for 8GB. There are also much higher numbers, like 50W for an R9 290X. isn't it watt per die instead of eg. 1 GB I don't know. Maybe that's one of the reason why the numbers are so different. It would be per die... The amount of power memory consumes is pretty small compared to the entirety of the card. Some people also forget that Ethereum is not the only coin in existence and building your ecosystem around one coin that will be getting hit hard in the coming months is not such a great idea. Nvidia is also not responsible for the kind of memory manufacturers use. That would be up to manufacturers such as MSI/Gigabyte/Asus. Ethereum continues to give the largest % of mining revenue by far, without it we can all throw our cards away. Also 30W/card would not be negligible. Only NVIDIA pulls this memory switching shit every generation, your argument fails because AMD and its AIB partners don't do this. Ethereum is going away this Januaryish as it shifts to PoS. Only Nvidia switches memory? lol... When I mined with AMD I had three different manufacturers. Hynix, Elpedia, and Samsung from four different card manufacturers. Some changing between the same cards and models. They do it because they find a better partner that can source memory for cheaper or better to them. Mining is a small drop in the bucket when it comes to the entirety of GPUs. As I mentioned AMD/Nvidia have nothing to do with who the memory is sourced from, only the type that is used (GDDR5/X, HBM). That's all the card manufacturers. ethereum originally planned to go PoS by summer 2016. then fall. now early 2017.dont hold your breath - IMO the devs and those close to the project are mining it, and advertising a PoS switch "soon" is an excellent deterrent to others from building farms It's still in the code, it was just moved back a year apparently. So next fall. I was operating off old information. -_- This just makes Nvidia even more worthless to mine with now. Everything else is still relevant.
|
I buy private Nvidia miners. Send information and/or inquiries to my PM box.
|
|
|
bathrobehero
Legendary
Offline
Activity: 2002
Merit: 1051
ICO? Not even once.
|
|
December 01, 2016, 04:08:12 AM |
|
Ethereum continues to give the largest % of mining revenue by far
Only if you're a plug and play miner who will all have a very bad time in the next few months and will shake you out of mining - even with very cheap electricity - until the next big thing comes out for a few weeks, maybe months. Personally, I don't even remember the last time I mined Ethereum.
|
Not your keys, not your coins!
|
|
|
bensam1231
Legendary
Offline
Activity: 1764
Merit: 1024
|
|
December 01, 2016, 07:46:47 AM |
|
Ethereum continues to give the largest % of mining revenue by far
Only if you're a plug and play miner who will all have a very bad time in the next few months and will shake you out of mining - even with very cheap electricity - until the next big thing comes out for a few weeks, maybe months. Personally, I don't even remember the last time I mined Ethereum. Ethereum is earning on par with Equihash right now, especially considering Nvidia miners are locked to Nicehash's miner. Apparently Ethereum armageddon was moved back to next year. It would be nice if there was a 'next big thing'. I have heard about anything coming out since Zcash.
|
I buy private Nvidia miners. Send information and/or inquiries to my PM box.
|
|
|
mirny
Legendary
Offline
Activity: 1108
Merit: 1005
|
|
December 01, 2016, 07:48:07 AM |
|
ethereum originally planned to go PoS by summer 2016. then fall. now early 2017.dont hold your breath - IMO the devs and those close to the project are mining it, and advertising a PoS switch "soon" is an excellent deterrent to others from building farms
Is It surely early 2017? because I found this http://cryptomine.weebly.com/cryptomine-blog/ethereum-to-switch-to-pos-before-november-1-2017
|
This is my signature...
|
|
|
sp_ (OP)
Legendary
Offline
Activity: 2954
Merit: 1087
Team Black developer
|
|
December 01, 2016, 08:16:04 AM |
|
Ethereum is earning on par with Equihash right now, especially considering Nvidia miners are locked to Nicehash's miner. Apparently Ethereum armageddon was moved back to next year.
It would be nice if there was a 'next big thing'. I have heard about anything coming out since Zcash.
LBRY is profitable again with the sp-mod private kernel.. $2.25 per day if you mine with 300MHASH
|
|
|
|
Nikolaj
|
|
December 01, 2016, 10:39:09 AM |
|
Ethereum is earning on par with Equihash right now, especially considering Nvidia miners are locked to Nicehash's miner. Apparently Ethereum armageddon was moved back to next year.
It would be nice if there was a 'next big thing'. I have heard about anything coming out since Zcash.
LBRY is profitable again with the sp-mod private kernel.. $2.25 per day if you mine with 300MHASH you're overstimating the amounts, and it's a pretty constant with your statements sp 1.6 on avg. It's better zcash ps: with the opensource 1.8.3 in any case, a bit better than private sp1 with 1070's (given the same performance but the absence of fees and reconnections)
|
|
|
|
sp_ (OP)
Legendary
Offline
Activity: 2954
Merit: 1087
Team Black developer
|
|
December 01, 2016, 10:58:43 AM |
|
yiimp is paying 0.00935BTC per day for 1000 MHASH (based on the payouts in the last 24 hours) 0.00935 * 0.3 * $753 = $2.11 LBRY sp-mod works best on the maxwell cards. 300 mhash on the 980ti. The opensource does around 250mhash. Lbry sp-mod also works good on the Gtx 1060 3gb and 6gb.
|
|
|
|
Nikolaj
|
|
December 01, 2016, 11:05:31 AM |
|
yiimp is paying 0.00935BTC per day for 1000 MHASH (based on the payouts in the last 24 hours) 0.00935 * 0.3 * $753 = $2.11 LBRY sp-mod works best on the maxwell cards. 300 mhash on the 980ti. The opensource does around 250mhash 0.007 on avg * 0.29 (0.29 for you without fees, 0.29*0.965 for the others = 0.279) * 745 = 1.51. If you count 0.279 it's 1.45$/day gross (1070) If you add very high power consumption of gtx 980Ti, that works at more than 220W, in EU with 0.15 EUR/KWh it's 0.84$, that minus 1.55/1.65 gross it's 0.66/0.9$ net/day. Simple math If you count the immediate 1 minute, and irrelevant, profitability it's another case, but you could probably find something better than lbc. It matters the 24h AVG amount, not the spikes. In this case we've had values well below 0.007 BTC/MH/Day with lbc
|
|
|
|
sp_ (OP)
Legendary
Offline
Activity: 2954
Merit: 1087
Team Black developer
|
|
December 01, 2016, 11:13:42 AM |
|
The 980ti doesn't use 220W. Reduce the tdp to 50-75% lower the memclock -500. and overclock the core a bit 300mhash@ around 130-150w
|
|
|
|
Amph
Legendary
Offline
Activity: 3248
Merit: 1070
|
|
December 01, 2016, 11:19:00 AM Last edit: December 01, 2016, 11:47:00 AM by Amph |
|
it's not $2.5 in any case(not a single coin is giving that much with one gpu), it's better to mine something else and hold lbry until it get pumped if ever
|
|
|
|
sp_ (OP)
Legendary
Offline
Activity: 2954
Merit: 1087
Team Black developer
|
|
December 01, 2016, 11:19:57 AM |
|
EU with 0.15 EUR/KWh i
In Norway I pay 0.04€ / KWh
|
|
|
|
Nikolaj
|
|
December 01, 2016, 11:20:34 AM |
|
Have you got a proof about this statement? 130W with 300MH/s on pbc with a gtx 980ti? I haven't found a single one. Please post some proofs. A 970? Thank you EU with 0.15 EUR/KWh i
In Norway I pay 0.04€ / KWh Another major element that it's incompatible with the general situation. In EU it could even be 0.26. Sp, making a general announcement of a gross profitability it's totally irrelevant, more so if your electrical infrastructure costs are so low. The AVG in EU it's 0.15, as mentioned by the latest report of the European Union in matter.
|
|
|
|
sp_ (OP)
Legendary
Offline
Activity: 2954
Merit: 1087
Team Black developer
|
|
December 01, 2016, 11:23:59 AM |
|
it's not $2.5 in any case(not a single coin is giving that much with oen gpu), it's better to mine something else and hold lbry until it get pumped if ever
Skein on zpool.ca 24hours payouts: 0.00709 / 1000MHASH. A single gtx 1070 (sp-mod private) 500MHASH. $2,65809 per card.
|
|
|
|
sp_ (OP)
Legendary
Offline
Activity: 2954
Merit: 1087
Team Black developer
|
|
December 01, 2016, 11:30:00 AM |
|
Have you got a proof about this statement? 130W with 300MH/s on pbc with a gtx 980ti? I haven't found a single one. Please post some proofs.
|
|
|
|
|