Plotting and Mining on Amazon Cloud Drive
-
@RichBC
Almost right Rich.
It was actually from the exact same PC rather than seperate PCs.
Everything else was as you described.
So the safest bet seems to get all the uploading done, then once the uploading is complete start mining.
It seems to be if you have uploaded and then downloaded in close succession.
-
@machasm OK understood. At the end of my current Month on the VPS I will try just Mining and see what happens. However I have this nasty feeling that it is the Mining that get's us locked?
Rich
-
@RichBC
Sheesh, I hope not after all the palava to get to that point.
Also found a better VPS provider called xhosts.uk.
They have a better control panel to build whatever OS you want, The machines are running true 3.5Ghz Xeon (not the 1Ghz that shows using CPU-ID on theworldservers ones) and the network connection can be as fast as 10Gbps.
-
-
@RichBC
Pricing isn't too bad.
I decided to go with an SSD solution since the disk read/writes take a while and changed the script to just create two files per upload rather than three due to ssd space limitations. Upload speeds seam to be similar,however, perhaps there is a bottleneck (how you can call 350Mbps a bottleneck!) at ACD preventing anything faster?
-
@machasm OK interesting would be good to see your final spec and also some output from the Log file to see how things compare?
Rich
-
@RichBC
This was what it looked like when I was creating 16Gb files with xhosts.uk18/02/2017 Start nonce 431000000 Stagger-65536 Size-16384
11:28:06.73 Start Plot 10
0:07:13.78
11:35:20.64 Start Plot 11
0:07:31.50
11:42:52.19 Start Optimise
0:09:21.78
11:52:14.00 Start Upload
0:31:35.7812:23:50.23 Start Plot 12
0:06:30.82
12:30:21.61 Start Plot 13
0:07:14.70
12:37:36.35 Start Optimise
0:08:36.46
12:46:12.84 Start Upload
0:18:43.2613:04:56.94 Start Plot 14
0:08:18.01
13:13:15.01 Start Plot 15
0:07:31.82
13:20:46.88 Start Optimise
0:08:18.64
13:29:05.56 Start Upload
0:16:46.7613:45:52.80 Start Plot 16
0:06:11.12
13:52:05.36 Start Plot 17
0:06:41.53
13:58:46.78 Start Optimise
0:08:31.75
14:07:18.60 Start Upload
0:24:32.40I have since increased the file sizes to 24GB and will upload a log as soon as the first run is finished.
This is what 24GB files look like from theworldservers
19/02/2017 Start nonce 101000000 Stagger-98304 Size-32768
12:44:39.13 Start Plot 10
0:52:51.90
13:37:31.00 Start Plot 11
0:40:50.00
14:18:21.02 Start Plot 12
0:41:48.25
15:00:09.28 Start Optimise
0:30:40.71
15:30:50.01 Start Upload
0:48:56.4315:19:44.50 Start Plot 13
0:39:10.31
15:58:54.82 Start Plot 14
0:37:37.68
16:36:32.50 Start Plot 15
0:40:09.51
17:16:42.01 Start Optimise
0:28:07.65
17:44:49.67 Start Upload
As you can see the plotting and optimising takes some time but I only have two cores at theworldservers whereas I have eight at xhosts.
Upload speed seems slighty faster at worldservers though?BTW how do you get the code to paste in a separate window like your post above?
-
@machasm Interesting numbers, the plot & optimise look much faster, upload similar. To get the code in a window select </> in the line above you are typing in and paste.
Rich
-
@RichBC
Here is a copy from the 24GB log19/02/2017 Start nonce 442769472 Stagger-98304 Size-16384 17:48:50.80 Start Plot 10 0:10:03.60 17:58:54.87 Start Plot 11 0:09:35.04 18:08:29.97 Start Plot 12 0:12:10.14 18:20:40.22 Start UploadThe upload took around 40 mins but had to re-upload since the files didn't appear at ACD end.
Also found a couple of tweaks to windows 10 that seemed to speed things up a littleEdit: Bogus info deleted.
-
Has anyone done the maths regarding how long this is going to take to read once the plots are in place?
Lets say we end up with about 30TB of plots all around 24GB in size and fully optimised.
How long are they going to take to read by the miner on a relatively good connection, say 100Mb/s?
It might take so long that a new block is found before the plots have finished being read!
-
@machasm
I read somewhere that in order to read the plots the amount of data required to download is 0.00025 of the total value?
If so that would mean on a 30TB plot you would need to download 7GB?
Have I got that right??
-
@machasm Yes I have read .024% of the data being read. Not certain of the maths, may have got the noughts wrong, but doesn't that work out to 720MB for 30TB Plot?
Rich
-
@machasm its 1/4096 th of the file size = 0.000244140625
30TB = 30000GB
30000GB * 0.000244140625 = 7.32421875 GBto see how long it would take to download will depend on your max download speed.
example 100Mb (Megabit) internet connection
100 Mb / 8 = 12.5 MB ( Megabytes per second )
12.5 MB /1000 = 0.0125 GB (GigaBytes per second)
7.324 GB / 0.0125 GB = 585.92 seconds = 9.72 Min
-
@iKnow0 Thought I had got the noughts wrong.... Bottom line is that with 30TB of plots you are really going to need a 1000Mb/s Connection and even then it will take about a minute to read.
Rich
-
@RichBC I also assumed there would be no network congestion, peak internet hours would also slow it down for sure. A VM in the same location as the storage would work also, but I doubt if it would be cost efficient, at least not yet anyway.
-
@iKnow0
Thanks for clearing that up.
Looks like I might try things when we get near the 10TB mark. See how it fares (fairs?)
-
@iKnow0 So does that means I would be using 7.5 GB of bandwidth for every block? :0 (for a 30TB plot)
-
@Glen yep
-
@iKnow0
Thanks for the math. Looks like it is spot on!
I have a 12MB/s connection and a 2.9TB plot and it takes about 62 secs to read so not far off the figure you quoted if extrapolating out to 30TB.
I think that unless you have a Gbps connection (Who has one of those?) then the plots are probably going to have to be no more than around 10TB max as you probably wont get to finish reading them when the next block appears.
-
Any new experiences with the mining on ACD?


