Burst Mining Issue and Questions (Open Bounties)

[DebitMe]

Current Bounties
Bounty 1 - Teamviewer into my machine and get Jminer setup.  This should use my graphics card which is currently a 280x, which should be able to check over 200 Mb/s with this working properly.

Pays - .02 btc

_______________________________________________________________________________ ________________________________________


I have recently gotten into Burstcoin mining, for various reasons that won't be discussed in this thread.  But am having some issues with a couple of aspects of it.  Here is my current setup and questions...

I bought three Seagate 8 TB hard drives from amazon, link below...
https://www.amazon.com/gp/product/B01HAPGEIE/ref=oh_aui_detailpage_o00_s00?ie=UTF8&psc=1
which I am working on getting set up to mine.

First, I have a rig that I am using for plotting, so far it has gotten faster plot times than my mining rig, so I have been using it.
I was able to get GPUplotter up and running on it, but am only able to get about 5200 nonces/minute.  I have 4 cards in this machine, a 1080 TI, 1080, 1080, and 1070 and am currently using the 1080Ti to plot.  Here is my device file...
0 0 1792 256 8192
and here is my start file
gpuPlotGenerator generate buffer d:\Burst\plots\6122256529182459080_50000000_30498624_2496
The machine only has 4 GB of RAM, but if I understand it correctly that shouldn't matter since it uses the 1080Ti only.

So question one is, shouldn't I be able to get faster plot time than about 5200?  I have been reading that some people are in the 20k range and figured a 1080Ti should be up there pretty high?

Second, I have a rig that I have the hard drives plugged into once they are done plotting.  I have gotten through 1 of the 8 TB HD's, and have a few smaller external HD's that I had laying around that I have been testing with.
This rig has 16 GB of RAM and an i5-4670K CPU at 3.40GHz (4 CPU's) and 2x 280x 4 GB GPU's.
The hard drives are plugged into a powered USB hub, which is then plugged into the computer.
When I run the CPU miner that comes with the Burst wallet, I get about 35 Mb/s when scanning the extra external hard drives, but it drops to only about 20 Mb/s when scanning the 8 TB ones.

Question two is, should I be getting faster scan times with these computer specs?  It just seems painfully slow to go through almost 12 TB in nearly a minute and a half to two minutes and I don't think I would be able to expand anymore, which is against my plan.

Question three is, I have been trying to get the Jminer to run, which uses the GPU, but all that has happened is a black screen shows up and then disappears after a quarter of second and that is it.  I tried it on both the mining and plotting rig and both the same issue.  I downloaded Java 8 from here...
http://www.oracle.com/technetwork/java/javase/overview/java8-2100321.html
and I am pretty sure my OpenCL is fine since both machines are able to mine with no issues.  So I am not sure what the problem is.

The first problem I need to solve is how to scan through the hard drives faster.  It takes about 3 days to plot a 8 TB drive currently, but I can at least grow the operation with that, not so much with the current mining speeds.  Any help would be greatly appreciated and I would be willing to pay a bounty to anyone who can get these things working.

Thanks,

DebitMe

[Elder III]

I am in the same approximate situation with starting to mine Burst.  I don't have an answer for #2 or #3, but with #1 I believe you need to increase the amount of memory allocated to the gpuplotgenerator. I'm using a rig full of old parts and it only has 4GB of DDR2 in it.  I have it set to use 2GB of RAM since Windows uses 1.-5GB or more by itself. I am getting about 14,500 nonces/minute with an RX 480 4GB plotting an 8TB 5900rpm HDD. I ordered 8GB  DDR2 off eBay to allow me to plot at faster rates. Once it's here next week I will test to see how much higher the nonces/minute is.

[DebitMe]

I am in the same approximate situation with starting to mine Burst.  I don't have an answer for #2 or #3, but with #1 I believe you need to increase the amount of memory allocated to the gpuplotgenerator. I'm using a rig full of old parts and it only has 4GB of DDR2 in it.  I have it set to use 2GB of RAM since Windows uses 1.-5GB or more by itself. I am getting about 14,500 nonces/minute with an RX 480 4GB plotting an 8TB 5900rpm HDD. I ordered 8GB  DDR2 off eBay to allow me to plot at faster rates. Once it's here next week I will test to see how much higher the nonces/minute is.

But doesn't the GPUplotgenerator only use the RAM from the GPU?  I didn't think it had anything to do with the RAM in your computer when using the GPUplotgenerator?

If I am mistaken, which setting is it that would change to fix that?  Can you share your device file line?

[felixbrucker]

i can answer number 2: your plots are not optimized (and afaik no gpu plotter can do it), so you are either limited to using the cpu plotter to generate optimized plots (slow on smr, and slow because of cpu instead of gpu) or write unoptimized plots with gpu plotter and then optimize them. The optimization process isnt inplace so you will need a second 8tb disk for that.


my 8tb disk i tested runs with a constant 45mb/s, i created optimized plots with cpu

you can easily identify optimized plots by the last two numbers of the plot file, they have to be the same, eg: 1339169714501365211_500000001_4100096_4100096

[DebitMe]

i can answer number 2: your plots are not optimized (and afaik no gpu plotter can do it), so you are either limited to using the cpu plotter to generate optimized plots (slow on smr, and slow because of cpu instead of gpu) or write unoptimized plots with gpu plotter and then optimize them. The optimization process isnt inplace so you will need a second 8tb disk for that.


my 8tb disk i tested runs with a constant 45mb/s, i created optimized plots with cpu

you can easily identify optimized plots by the last two numbers of the plot file, they have to be the same, eg: 1339169714501365211_500000001_4100096_4100096

Gotcha,

How long did it take you to write 8 TB with your CPU?

[Elder III]

I am in the same approximate situation with starting to mine Burst.  I don't have an answer for #2 or #3, but with #1 I believe you need to increase the amount of memory allocated to the gpuplotgenerator. I'm using a rig full of old parts and it only has 4GB of DDR2 in it.  I have it set to use 2GB of RAM since Windows uses 1.-5GB or more by itself. I am getting about 14,500 nonces/minute with an RX 480 4GB plotting an 8TB 5900rpm HDD. I ordered 8GB  DDR2 off eBay to allow me to plot at faster rates. Once it's here next week I will test to see how much higher the nonces/minute is.

But doesn't the GPUplotgenerator only use the RAM from the GPU?  I didn't think it had anything to do with the RAM in your computer when using the GPUplotgenerator?

If I am mistaken, which setting is it that would change to fix that?  Can you share your device file line?

Try changing the end of your line:

gpuPlotGenerator generate buffer d:\Burst\plots\6122256529182459080_50000000_30498624_2496  --- to double that and see if the speed increases. Mine is set at 10,000 stagger size. I originally had it set at 20,000 but I was running out of RAM and freezing the system. I read that the end nonce number needs to be dividable by your stagger size, so round numbers work best for that.

[DebitMe]

I am in the same approximate situation with starting to mine Burst.  I don't have an answer for #2 or #3, but with #1 I believe you need to increase the amount of memory allocated to the gpuplotgenerator. I'm using a rig full of old parts and it only has 4GB of DDR2 in it.  I have it set to use 2GB of RAM since Windows uses 1.-5GB or more by itself. I am getting about 14,500 nonces/minute with an RX 480 4GB plotting an 8TB 5900rpm HDD. I ordered 8GB  DDR2 off eBay to allow me to plot at faster rates. Once it's here next week I will test to see how much higher the nonces/minute is.

But doesn't the GPUplotgenerator only use the RAM from the GPU?  I didn't think it had anything to do with the RAM in your computer when using the GPUplotgenerator?

If I am mistaken, which setting is it that would change to fix that?  Can you share your device file line?

Try changing the end of your line:

gpuPlotGenerator generate buffer d:\Burst\plots\6122256529182459080_50000000_30498624_2496  --- to double that and see if the speed increases. Mine is set at 10,000 stagger size. I originally had it set at 20,000 but I was running out of RAM and freezing the system. I read that the end nonce number needs to be dividable by your stagger size, so round numbers work best for that.

Yea, I believe the end number needs to be divisible by 8 and the second to last needs to be divisible by the last.

So to run the optimizer, it doesn't just write over the current file?  I need a new space to save it to?

[felixbrucker]

So to run the optimizer, it doesn't just write over the current file?  I need a new space to save it to?

The optimization process isnt inplace so you will need a second 8tb disk for that.

Gotcha,

How long did it take you to write 8 TB with your CPU?

i was generating nonces faster than the disks write speed so i plotted two in parallel (probably could have done more), took about 10-12 days

[DebitMe]

I am tired of messing with this jminer to try to get my GPU's to run so I will be opening bounties to get the parts done that I need done, starting with the ones that are holding me back the most.

Bounty 1 - Teamviewer into my machine and get Jminer setup.  This should use my graphics card which is currently a 280x, which should be able to check around 200 Mb/s with this working.

Pays - .02 btc

[exus]

I am tired of messing with this jminer to try to get my GPU's to run so I will be opening bounties to get the parts done that I need done, starting with the ones that are holding me back the most.

Bounty 1 - Teamviewer into my machine and get Jminer setup.  This should use my graphics card which is currently a 280x, which should be able to check around 200 Mb/s with this working.

Pays - .02 btc

check your bat file
if it contains -d64 you need 64 bit java version

(bat example)
java -jar -d64 -XX:+UseG1GC burstcoin-jminer-0.4.10-SNAPSHOT.jar
pause      (<- add this pause to see errors)

you can find 64bit java here
https://www.java.com/it/download/manual.jsp

in jminer.properties file set:
plotPaths=Z:/   (comma separated)
poolMining=true
numericAccountId=yourNumericAccount
poolServer=http://poolofd32th.club:8124    (I suggest you this one)
walletServer=    (leave blank)
platformId=   your gpu id 

just tested on a 280x and it works

[lunas]

I just plotted a WD 4TB My Book Desktop External Hard Drive - USB 3.0 last night on my 280x was getting around 26k nounces per min completed the process in about 10 hours. i might replot the other 4tb i have to try making it go faster... latest drive run in about 46 seconds vs the other 4tb which goes for 74 seconds...

my settings were

with a 4096 stagger size

devices file for gpu plotter

0 0 2048 128 4096

here is the numbers for a 1tb plot
gpuPlotGenerator generate buffer F:\burst\plots\xxxxxxxxxxxxxxxxxxxxxxxx_1_3735936_4096

same settings on my laptop went at 16k/min with a 4gb r9 m375

[DebitMe]

I am tired of messing with this jminer to try to get my GPU's to run so I will be opening bounties to get the parts done that I need done, starting with the ones that are holding me back the most.

Bounty 1 - Teamviewer into my machine and get Jminer setup.  This should use my graphics card which is currently a 280x, which should be able to check around 200 Mb/s with this working.

Pays - .02 btc

check your bat file
if it contains -d64 you need 64 bit java version

(bat example)
java -jar -d64 -XX:+UseG1GC burstcoin-jminer-0.4.10-SNAPSHOT.jar
pause      (<- add this pause to see errors)

you can find 64bit java here
https://www.java.com/it/download/manual.jsp

in jminer.properties file set:
plotPaths=Z:/   (comma separated)
poolMining=true
numericAccountId=yourNumericAccount
poolServer=http://poolofd32th.club:8124    (I suggest you this one)
walletServer=    (leave blank)
platformId=   your gpu id  

just tested on a 280x and it works

It looks like the Java version is the issue.  I put in the pause statement, which was very helpful, and it is showing this...
"Error: This Java instance does not support a 64-bit JVM.
Please install the desired version"

I went to the website you put there and downloaded and installed the Java version and it still shows the same message.  I have not rebooted the rig since it is currently plotting, could that be the issue?

EDIT:  I restarted the machine and no change from that message.

EDIT2: Freaking A, I was clicking the wrong download link, got 64 bit installed and it seems to be starting up.  Let me test on main rig and Ill let you know.

[DebitMe]

Ok, so the GPU miner is running, but I am not getting any faster read times from my hard drives.  All but one was plotted using CPU, so they should all be optimized besides that 8 TB drive.  I took it out and am getting slightly faster read times, but when I put it back in it is just as slow as the CPU reader.

I would think that whether it is optimized or not, the GPU should get faster read times than the CPU right?

[QuintLeo]

You might be getting limited on drive read speed by the use of USB - if your drives or hub or connection is USB2 at ANY point, that is a significant limitation on drive read speed.

 I'm currently CPU plotting one of those Seagate 8TB now - into the second day and a bit more than halfway done with a FX 8570, but I'm also doing other stuff on the machine which slows the plotter down.
 I'm not plotting it in one chunk though - half TB at a time is close enough - and it seems to READ about as fast as any other non-SSD drive I've worked with, easily hitting 80 MB/sec on jminer running the CPU to mine with (when it's not plotting at the time).

 I've never been able to get GPUplotter to run at all, but the CPU plotting had been "fast enough" up to now.

[DebitMe]

Ok, I had been using a Rosewill 10 port hub that is only 2.0 read speed.  I took the drive off of that and plugged it straight into a 3.0 USB socket attached to the MOBO and am still getting the same speeds.

I downloaded the optimizer and that is currently running to optimize the 8 TB plot.  That is about my last hope at this point.  Not sure what else to do.  I kind of figured i would go ahead and build another PC for this, but frankly without hard evidence that I can get the speeds I think I should be able to, I don't really want to shell out 1.5k without being totally sure.

[DebitMe]

What does, "Unable to get mining info from wallet" mean in Jminer?

[QuintLeo]

What does, "Unable to get mining info from wallet" mean in Jminer?

 I haven't figured that one out myself, but I just ignore it as it doesn't seem to affect anything.

[Antigo]

What does, "Unable to get mining info from wallet" mean in Jminer?

As far as I understand it, the Jminer tries to get information about the winner of a block from a wallet to display it in the log. It's an optional feature and can either use a local or an online wallet. You should be able to deactive this feature by setting

Code:

walletServer=

to an empty string in the jminer.properties file

[QuintLeo]

What does, "Unable to get mining info from wallet" mean in Jminer?

As far as I understand it, the Jminer tries to get information about the winner of a block from a wallet to display it in the log. It's an optional feature and can either use a local or an online wallet. You should be able to deactive this feature by setting

Code:

walletServer=

to an empty string in the jminer.properties file

 Which does NOT stop the "Unable to get mining info from wallet" error messages.
 I've DONE that.

 I'm not sure if that's a bug or if the message comes from something else.

[DebitMe]

What does, "Unable to get mining info from wallet" mean in Jminer?

As far as I understand it, the Jminer tries to get information about the winner of a block from a wallet to display it in the log. It's an optional feature and can either use a local or an online wallet. You should be able to deactive this feature by setting

Code:

walletServer=

to an empty string in the jminer.properties file

 Which does NOT stop the "Unable to get mining info from wallet" error messages.
 I've DONE that.

 I'm not sure if that's a bug or if the message comes from something else.

Huh, yea its really weird.  I will get anywhere from 6 to 20 after a round finishes for no reason.  Everything gets submitted fine and I am getting paid no problem, so it doesn't really bother me.

I am hoping my issues will be resolved once this drive finishes optimizing.  It is about 25% of the way done and I started the miner on the new plot that is being created. and it seems to be reading in the 800 to 2000 mh/s range.  If that is the real read speed, then I am ready to expand this thing big time.  I am picking up a NAS that someone was selling on craigslist that has 8 slots and comes with 3 TB hard drives in all slots for a total of 24 TB.  Will be cool if I can get this thing going.