greatwolf
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 17, 2014, 03:00:40 AM |
|
The reject rate on this is horrible. Getting almost a 20% reject using cudaminer 750 ti.
|
|
|
|
phzi
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 17, 2014, 03:17:39 AM Last edit: April 17, 2014, 03:30:57 AM by phzi |
|
The reject rate on this is horrible. Getting almost a 20% reject using cudaminer 750 ti.
I have been averaging around 5% rejects from western Canada, which is perfectly acceptable for a pool located in the EU (no geo-located stratum endpoints here yet, but apparently coming soon). please help me understand if i have these calculations right.
if i were to buy hashpower of 0.1 (that's 100mh/s correct?) and spend 5btc at a price of 5btc/day
i would get 100mh/s for 24 hours, is that correct?
.1GH would be 100MH, correct. If you ordered 5BTC of hashing at .1GH, at a rate of 5BTC/GH/Day, you would get 10 days of 100MH. also, i would essentially be bidding against other people that are purchasing hash power right? so the higher the price i list, the most likely my offer will get picked up? it looks like there is a wide range of bids, would a bid of say 2.5 btc/day get picked up? it's hard to tell it looks like every bid is being picked up right now.
Yes, you are in a bidding competition for hashing power. The higher you bid, the higher priority your job has. Keep in mind, however, that miners can set minimum payout rates. For example, I have p=5 set, so if there is no work paying >5BTC/GH/day, my miners failover to another pool and won't work on lower paying jobs. Furthermore, there have been rare few cases where the active jobs have been < 5BTC/GH/day since the pool started, so the chances of a 2.5BTC/GH/Day job getting completely quickly are very slim. kenshirothefist, can you explain how the hashing power is distributed from the seller's perspective, especially now that there are limits on the GH per job?
I certainly hope it's an averaging scheme, where all workers are getting paid for their portion of work completed irrelevant of what task it was on... although that might complicate matters regarding minimum profitability settings. If this isn't how it works, then I am going to setup a string of NiceHash failover pools with gradually decreasing minimum profitability, so that I'm always mining the highest paying job =p. --- Can you please add date/time added (or maybe "Time Waiting") to the current orders table?
|
|
|
|
dspair
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 17, 2014, 03:44:43 AM |
|
I certainly hope it's an averaging scheme, where all workers are getting paid for their portion of work completed irrelevant of what task it was on... although that might complicate matters regarding minimum profitability settings. I'm pretty sure it's not. 15 minutes ago I was mining @ 5.5BTC/GH, now I've been switched to 5.9BTC/GH, even though there is two 6.5BTC/GH orders.
|
|
|
|
phzi
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 17, 2014, 04:04:47 AM |
|
I certainly hope it's an averaging scheme, where all workers are getting paid for their portion of work completed irrelevant of what task it was on... although that might complicate matters regarding minimum profitability settings. I'm pretty sure it's not. 15 minutes ago I was mining @ 5.5BTC/GH, now I've been switched to 5.9BTC/GH, even though there is two 6.5BTC/GH orders. I am actually pretty sure it is an average. The chart shows my profitability has been fractional amounts like 5.54, 5.88, 6.12, etc since the change, and fluctuating much more rapidly. Previously it was 5.4 for a while, and then 5.6, back to 5.4, etc. Last 5 minutes have been at a profitability of 5.75BTC/GH/day. A very rough average based on GH limit rather then real GH rate would be (6.5*.1)+(5.9*.3)+(5.8*.2)+(5.6*.4)+(5.4*.01) = 5.874. Which is close enough to make me think the payout rate is definitely an average.
|
|
|
|
TheMathGuy
Newbie
Offline
Activity: 26
Merit: 0
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 17, 2014, 04:16:43 AM |
|
I certainly hope it's an averaging scheme, where all workers are getting paid for their portion of work completed irrelevant of what task it was on... although that might complicate matters regarding minimum profitability settings. I'm pretty sure it's not. 15 minutes ago I was mining @ 5.5BTC/GH, now I've been switched to 5.9BTC/GH, even though there is two 6.5BTC/GH orders. I am actually pretty sure it is an average. The chart shows my profitability has been fractional amounts like 5.54, 5.88, 6.12, etc since the change, and fluctuating much more rapidly. Previously it was 5.4 for a while, and then 5.6, back to 5.4, etc. Last 5 minutes have been at a profitability of 5.75BTC/GH/day. A very rough average based on GH limit rather then real GH rate would be (6.5*.1)+(5.9*.3)+(5.8*.2)+(5.6*.4)+(5.4*.01) = 5.874. Which is close enough to make me think the payout rate is definitely an average. I am pretty sure it is not, since my graph says 5.90 at the moment, while countless others have 5.50 On a different note, is anyone else noticing the random crashes/resets? After a while almost all the hashpower seems to fallback to the unlimited orders, whether they offer more or not.
|
|
|
|
ca1ne
Newbie
Offline
Activity: 11
Merit: 0
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 17, 2014, 04:42:31 AM |
|
Highest, unlimited, but only 7 miners. Majority of miners are on one of the lowest paid tiers.
Miners Speed GH/s #492 Alive 6.10 0.13888305 0.01% 1561.21 0.10 0.00 No 7 0.0004 #490 Alive 6.00 0.01839692 5.67% 1.39 15.91 0.06 0.10 80 0.0530 #472 Alive 5.90 0.12783510 29.62% 1.69 787.77 13.61 0.35 161 0.3084 #484 Alive 5.90 0.26161980 9.09% 2.49 383.25 8.34 No 249 0.4266 #486 Alive 5.90 0.09758923 0.32% 31.69 4.58 0.05 0.10 4 0.0125 #485 Alive 5.70 0.48990200 0.00% ∞ 0.00 0.00 0.20 0 0.0000 #475 Alive 5.60 0.09712124 0.80% ∞ 12.05 0.18 0.10 0 0.0000 #481 Alive 5.60 0.19590200 0.00% ∞ 0.00 0.00 0.10 0 0.0000 #483 Alive 5.60 0.97990200 0.00% ∞ 0.00 0.00 0.40 0 0.0000 #489 Alive 5.60 0.09790200 0.00% ∞ 0.00 0.00 0.10 0 0.0000 #436 Alive 5.50 0.71617964 26.91% 4.89 4142.87 113.39 No 355 0.6388 #474 Alive 5.50 0.09790200 0.00% ∞ 0.00 0.00 0.10 0 0.0000 #406 Alive 5.40 0.16885492 17.49% 81.82 572.55 12.39 No 5 0.0092
|
|
|
|
TheMathGuy
Newbie
Offline
Activity: 26
Merit: 0
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 17, 2014, 04:54:49 AM |
|
Highest, unlimited, but only 7 miners. Majority of miners are on one of the lowest paid tiers.
Miners Speed GH/s #492 Alive 6.10 0.13888305 0.01% 1561.21 0.10 0.00 No 7 0.0004 #490 Alive 6.00 0.01839692 5.67% 1.39 15.91 0.06 0.10 80 0.0530 #472 Alive 5.90 0.12783510 29.62% 1.69 787.77 13.61 0.35 161 0.3084 #484 Alive 5.90 0.26161980 9.09% 2.49 383.25 8.34 No 249 0.4266 #486 Alive 5.90 0.09758923 0.32% 31.69 4.58 0.05 0.10 4 0.0125 #485 Alive 5.70 0.48990200 0.00% ∞ 0.00 0.00 0.20 0 0.0000 #475 Alive 5.60 0.09712124 0.80% ∞ 12.05 0.18 0.10 0 0.0000 #481 Alive 5.60 0.19590200 0.00% ∞ 0.00 0.00 0.10 0 0.0000 #483 Alive 5.60 0.97990200 0.00% ∞ 0.00 0.00 0.40 0 0.0000 #489 Alive 5.60 0.09790200 0.00% ∞ 0.00 0.00 0.10 0 0.0000 #436 Alive 5.50 0.71617964 26.91% 4.89 4142.87 113.39 No 355 0.6388 #474 Alive 5.50 0.09790200 0.00% ∞ 0.00 0.00 0.10 0 0.0000 #406 Alive 5.40 0.16885492 17.49% 81.82 572.55 12.39 No 5 0.0092
Something is clearly wrong with the hash distribution, the FAQ says miners will slowly move from one order to another but this is ridiculous.
|
|
|
|
evolvia31
Member
![*](https://bitcointalk.org/Themes/custom1/images/star.gif)
Offline
Activity: 74
Merit: 10
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 17, 2014, 05:00:46 AM |
|
HI,
it's possible to add on the stats page the job which are working on ?
Now with the new features to limit the hash power by job, miners are split onto few jobs, so it ccould be interesting to know on which job each miner works no ?
|
|
|
|
phzi
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 17, 2014, 06:30:07 AM |
|
I am pretty sure it is not [an average], since my graph says 5.90 at the moment, while countless others have 5.50
Hm, maybe it's just because I have multiple rigs mining here, and they are on different jobs. -- I think the Speed GH/s measurement is laggy or over a fairly long window.
|
|
|
|
bruiser
Member
![*](https://bitcointalk.org/Themes/custom1/images/star.gif)
Offline
Activity: 68
Merit: 13
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 17, 2014, 06:30:36 AM |
|
Great to see the updates put in so quick for the renters! I would also ask for the ability to switch pools without having to cancel/recreate orders (I think thats probably causing the biggest amount of order spam) but I also know with the way your system is setup it might be hard to switch that info without an actual "reset" of the order.
Anyways, great stuff. I like your approach to renting.
|
|
|
|
itos
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 17, 2014, 06:52:29 AM Last edit: April 17, 2014, 07:02:44 AM by itos |
|
This needs A LOT more hashing to cover more orders. If not it would just a private service for rich investors and would probably fail later. Hopefully we can see more miners joining and covering more orders so the price doesn't go to the moon and no one will want to use the service.
EDIT: This site will become the pump and dump of mining! Like exchanges awesome.
|
|
|
|
dspair
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 17, 2014, 07:20:24 AM Last edit: April 17, 2014, 07:37:55 AM by dspair |
|
Highest, unlimited, but only 7 miners. Majority of miners are on one of the lowest paid tiers.
Yep, current orders: ![](https://ip.bitcointalk.org/?u=https%3A%2F%2Fi.imgur.com%2FVHcL4gv.png&t=665&c=74Nqpja-bf7FNQ) But I mine on a 6.40 order. Looks like restarting sgminer makes you switch.
|
|
|
|
kenshirothefist (OP)
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 17, 2014, 08:08:36 AM |
|
OK, so we were very busy in the past hours due to several service upgrades. Notably the most important is the option to limit the hash power. Behind the scenes some other features have been added to lower rejects, improve efficiency and allow miners to set lower diff. Providers/Sellers:I see the majority of the questions are regarding how providers are assigned to orders. For now this is simplified and we are using the option that is most suitable for now (keep in mind that our service is still young!) and this is round-robin. Providers are randomly assigned to orders in a round-robin fashion to provide fairness between providers. This means that you'll be mining a certain amount of time on lower paid orders and a certain amount of time on higher paid orders. Since we are doing round-robin providers will be more or less equally distributed among different orders which does guarantee fairness between providers. Although there still might be some slight difference at the end of the day. Differences between prices are relatively low - a couple of percents between highest and lowest paying active order (10% in max for a very short time between order switching) and if we take into account round-robin "fairness" you should see <1% difference between providers payments (in a daily average). However, since we strive to 100% fair game, we're already working on a even better solution. What we're trying to do is to implemente a scheme where all workers are getting paid for their portion of work completed irrelevant of what task it was on. However, this is not at all a trivial task. There are several issues we are facing; orders are dynamic, can be canceled at any time, providers are dynamic, and there are tons of shares to be processed at any time which bring us to the issue of the "precision" of IEEE double precision floating point if we're to calculate the pricing distribution for each an every single share that is being processed through our system. It is doable, but first we have to mathematically prove that our algorithm is correct and then we have to do test on real examples to see what kind of error are we getting due to rounding. I would just like to stress out that the current round-robin scheme is doing very good in average. It's also worth noting that any kind of multi-pool is having similar issues (immature/unexchaged/rounding precision) ... it is very very hard to do an absolute 100% fair game for all providers. But hey - what's important is that you're getting paid a good price for your hash power. It shouldn't really matter if you're paid 1% more than your friend in the morning while in the afternoon it will be vice-versa, and at the and of the day you'll both be happy Buyers:Calculations: BTC/GH/Day = BTC/1000*MH/Day BTC/TH/Day = BTC/1000*GH/Day How much time will I get for a specific amount of BTC at a specific rate? Time in days = (amount BTC / speed GH) / rate BTC/GH/day Example: I wanna pay 0.05 BTC at 5 BTC/GH/day for 0.1 GH of speed (0.05 BTC / 0.1 GH) / 5 BTC/GH/day = 0.5 BTC/Gh / 5 BTC/GH/day = (0.5 / 5) day = 0.1 day = 2.4 hours You'll get 2.4 hours of 0.1 GH speed for 0.05 BTC at 5 BTC/GH/day. Will add this calculation on the order submit page today. Order submission: What we're seeing is an excessive fight for the "unlimited" hash power. We're happy to see this, however very small/short orders are killing the efficiency of the system. Therefore we'll probably have to introduce minimum time frame for the duration of your order at submit time (you'll still be able to cancel it at any time, though). It's pointless to run a 1.5Gh/s job for only 15 minutes, this only messes-up the orders and probably isn't efficient for you as well (we're thinking about 0.5h minimum time at submit time). We'll also consider the upper limit for order speed in dependence of the total available speed (for example, max speed 1GH if 1.5GH is available, 2GH if 3 GH is availabe ... 7 GH if 10GH is available, etc.). Providers/sellers and buyers! Please keep in mind that NiceHash is still a novel system and we'll try to adapt to both of you to make a service as valuable as possible for both of you! Thank you for using NiceHash! btw: more providers are needed, currently there are good priced orders!
|
|
|
|
phzi
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 17, 2014, 08:11:03 AM |
|
This needs A LOT more hashing to cover more orders. If not it would just a private service for rich investors and would probably fail later. Hopefully we can see more miners joining and covering more orders so the price doesn't go to the moon and no one will want to use the service.
That makes almost no sense... If the price climbs high, then that means there are orders that high already, so we don't need more people to buy hashrate... Growth is good, but crazy growth would just clear the order books (which would be very bad). Awesome to see BTC/GH/Day rates climbing right now! >7 is pure awesome.
|
|
|
|
suchmoon
Legendary
Offline
Activity: 3766
Merit: 9005
https://bpip.org
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 17, 2014, 08:26:28 AM |
|
OK, so we were very busy in the past hours due to several service upgrades. Notably the most important is the option to limit the hash power. Behind the scenes some other features have been added to lower rejects, improve efficiency and allow miners to set lower diff.
I just wanted to say - as I think I haven't mentioned this among all the petty complaints ![Wink](https://bitcointalk.org/Smileys/default/wink.gif) - you guys are doing an awesome job. Keep it up!
|
|
|
|
CoinBuzz
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 17, 2014, 08:37:59 AM |
|
I see the majority of the questions are regarding how providers are assigned to orders. For now this is simplified and we are using the option that is most suitable for now (keep in mind that our service is still young!) and this is round-robin. Providers are randomly assigned to orders in a round-robin fashion to provide fairness between providers. This means that you'll be mining a certain amount of time on lower paid orders and a certain amount of time on higher paid orders. Since we are doing round-robin providers will be more or less equally distributed among different orders which does guarantee fairness between providers. Although there still might be some slight difference at the end of the day. Differences between prices are relatively low - a couple of percents between highest and lowest paying active order (10% in max for a very short time between order switching) and if we take into account round-robin "fairness" you should see <1% difference between providers payments (in a daily average).
So, for me as a buyer, Since miners are distributed randomly equal, if i order some GHash with the lowest possible price, i get equal mining power as if i was ordering with highest price?
|
|
|
|
itos
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 17, 2014, 08:45:56 AM |
|
I think miners are thinking that is great that price increases, but as a buyer you put an order an immediately another one goes up. Price will 10x and then it will be a ghost town where no one wants to use this service, since the prices is just too expensive. Only 1 person a whale will use it. I am sure the owners can think in a better system for distribution and fairness for both buyers and providers. Or maybe we really need more miner providers that can put more hash rate available and fill more orders quickly.
|
|
|
|
kenshirothefist (OP)
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 17, 2014, 08:53:59 AM |
|
So, for me as a buyer, Since miners are distributed randomly equal, if i order some GHash with the lowest possible price, i get equal mining power as if i was ordering with highest price?
The distribution of miners of course doesn't affect buyers. All providers (miners/sellers) are of the "same quality" (NiceHash won't charge you for invalid shares, produced by providers, if there are some). So you get mining power of "equal quality", but of course, if there are too many better paying orders and there are no miners left for lower paying orders, your order will be on pause until better paying orders are processed. I think miners are thinking that is great that price increases, but as a buyer you put an order an immediately another one goes up. Price will 10x and then it will be a ghost town where no one wants to use this service, since the prices is just too expensive. Only 1 person a whale will use it. I am sure the owners can think in a better system for distribution and fairness for both buyers and providers. Or maybe we really need more miner providers that can put more hash rate available and fill more orders quickly.
Agree, we don't want for prices to go to the Venera ![Wink](https://bitcointalk.org/Smileys/default/wink.gif) , because this wouldn't be too good neither for buyers nor for seller on the long run. As I mentioned in my previous post we'll introduce some limitations for the service to be attractive to various kinds of buyers. We still wan't to be able to provide massive hash rate to some buyers - that's one of our important features ... but we'll make sure even smaller orders will get its hash rate. Then again, I can't promise you a completely whales-proof system ... it's like in the coin market ... if a whale comes in you just can't stop it, it'll just buy it all ... unfortunately (it's a free market, what can we do...).
|
|
|
|
itos
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 17, 2014, 09:04:11 AM |
|
So, for me as a buyer, Since miners are distributed randomly equal, if i order some GHash with the lowest possible price, i get equal mining power as if i was ordering with highest price?
The distribution of miners of course doesn't affect buyers. All providers (miners/sellers) are of the "same quality" (NiceHash won't charge you for invalid shares, produced by providers, if there are some). So you get mining power of "equal quality", but of course, if there are too many better paying orders and there are no miners left for lower paying orders, your order will be on pause until better paying orders are processed. I think miners are thinking that is great that price increases, but as a buyer you put an order an immediately another one goes up. Price will 10x and then it will be a ghost town where no one wants to use this service, since the prices is just too expensive. Only 1 person a whale will use it. I am sure the owners can think in a better system for distribution and fairness for both buyers and providers. Or maybe we really need more miner providers that can put more hash rate available and fill more orders quickly.
Agree, we don't want for prices to go to the Venera ![Wink](https://bitcointalk.org/Smileys/default/wink.gif) , because this wouldn't be too good neither for buyers nor for seller on the long run. As I mentioned in my previous post we'll introduce some limitations for the service to be attractive to various kinds of buyers. We still wan't to be able to provide massive hash rate to some buyers - that's one of our important features ... but we'll make sure even smaller orders will get its hash rate. Then again, I can't promise you a completely whales-proof system ... it's like in the coin market ... if a whale comes in you just can't stop it, it'll just buy it all ... unfortunately (it's a free market, what can we do...). Agree whales will always exist. But right now as a buyers doesn't make much sense to mine for 15 seconds and then immediately stop mining because there is another high order. Could be wrong but at this pace we will be 20 Bitcoin per Gh/s per day and of course at that price no one will buy for long term, just a bunch of small volume orders of people that want to show off high hashrate for low time. Higher hash rate available will solve most of this problems since orders could be filled more quickly and cover more buy orders, not just the top one. Overall great job with the site, I know it's on it's beginning so a lot of improvements will come.
|
|
|
|
kenshirothefist (OP)
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 17, 2014, 01:01:30 PM |
|
OK, new update has been applied to the service, both buyers and sellers should benefit from this update.
Editing orders: A) We've finally enabled the ability to edit orders. Now you're able to edit running orders on the fly. You can change the price as well as (optional) max hash limit. This will allow you to bid higher within your's existing orders. As a consequence, miners (sellers) won't be disconnected (basically miners will "travel" with you). Edit: Price is only allowed to be increased (otherwise you'd be able to "pull" miners with you to lower paying order).
Order submition: B) When adding new order you'll see the approximate duration of the order (time), presuming your order will run at selected speed full time.
With these updates we would like to encourage you to properly limit and update you orders and thus keep the system in high-efficient mode. We've also slightly raised the "new order fee" to prevent spamming with orders and to encourage you to rather edit existing orders over canceling/submitting new orders (of course there is no fee for editing orders).
btw: when you edit your order you can't change the amount of BTC (this can't be done and will probably never be enabled), so we suggest you to load your order with a reasonable amount of BTC for you to be able to make proper bids. Of course, you can still cancel you order at any time with no penalty.
We all wish you nice hashing and thanks for bringing us close to 2 GHs!
|
|
|
|
|