[ANN][BURST] Burst | Efficient HDD Mining | New 1.2.3 Fork block 92000

[twig123]

Here is the new version. I reduced the amount of memory used from 40KB to about 1KB per unit. The only drawback is that it requires twice the global memory as before. I will search a mean to reduce this overhead later.
In CPU mode, it all goes pretty well (when no graphic card is detected).
The GPU mode is still kind of buggy on my graphic card (an old GeForce 9300M GS), don't know the exact reason yet. Sometimes it works, sometimes not. I will try to fix this issue tomorrow.

Here are the files :
gpuPlotGenerator-src-1.1.0.7z : https://mega.co.nz/#!iYFWAL5B!BvtmRQ5qGq4gGwjDglFNtDtNIX4LDaUvATBtClBdTlQ
gpuPlotGenerator-bin-win-x86-1.1.0.7z : https://mega.co.nz/#!aBVGBBQD!tBsRtb8VrHR12_anrFTrl41U0fPQu_OqFnxyi5nCyBY

For the linux users, the Makefile has a new target named "dist" that builds and copy/paste all the necessary files to the "bin" directory.

The executable usage is : ./gpuPlotGenerator <path> <address> <start nonce> <nonces> <stagger size> <threads>
<path> : the path to the plots directory
<threads> : number of parrallel threads for each work group

Tried the Windows version on my Lappy (has a nvidia geforce 335M, 1GB) and I get the following:

Code:

gpuPlotGenerator.exe C:\plots 11111222223333344444 0 5000 500 1

Code:

GPU plot generator v1.1.0
Author: Cryo
Bitcoin: 138gMBhCrNkbaiTCmUhP9HLU9xwn5QKZgD
Burst: BURST-YA29-QCEW-QXC3-BKXDL
--------------
Path: C:\plots
Nonces: 0 to 5000 (1 GB)
Process memory: 125MB
Threads number: 1
--------------
Retrieving OpenCL platform
Retrieving OpenCL GPU device
Creating OpenCL context
Creating OpenCL command queue
Creating CPU buffer
Creating OpenCL GPU generation buffer
Creating OpenCL GPU scoops buffer
Creating OpenCL program
Building OpenCL program
Creating OpenCL kernel
Setting OpenCL kernel arguments
Generating from nonce #0
An OpenCL error occured in the generation process, aborting...
>>> [-5] Error in synchronous read

Each time I run it, the screen goes black for a moment and the comes back. Then I get a popup stating that the Nvidia display driver has stopped responding and has recovered.

I have my desktop with a ATI 7800 already in progress of plotting with dcct's plot generator on Linux. But I will see if I can figure out how to test this on linux on an ATI card.

Edit: Tested it on my Ubuntu install and I get the following when issuing the 'make' command:

Code:

Compiling [gpuPlotGenerator.cpp]
make: /c/_data/cryo/_apps/mingw/bin/g++: Command not found
make: *** [gpuPlotGenerator.o] Error 127

Seems this may be a hard coded location?
I've already installed 'build-essential'... how do I change this so I can actually test it?

[Conduit]

A show of hands... how many of you have said "Aww... crap..."  Delete... start plotting over again... 

*raises hand*

I think I did that at least three or four times before I got everything sorted out and working.

[wmikrut]

Here is the new version. I reduced the amount of memory used from 40KB to about 1KB per unit. The only drawback is that it requires twice the global memory as before. I will search a mean to reduce this overhead later.
In CPU mode, it all goes pretty well (when no graphic card is detected).
The GPU mode is still kind of buggy on my graphic card (an old GeForce 9300M GS), don't know the exact reason yet. Sometimes it works, sometimes not. I will try to fix this issue tomorrow.

Here are the files :
gpuPlotGenerator-src-1.1.0.7z : https://mega.co.nz/#!iYFWAL5B!BvtmRQ5qGq4gGwjDglFNtDtNIX4LDaUvATBtClBdTlQ
gpuPlotGenerator-bin-win-x86-1.1.0.7z : https://mega.co.nz/#!aBVGBBQD!tBsRtb8VrHR12_anrFTrl41U0fPQu_OqFnxyi5nCyBY

For the linux users, the Makefile has a new target named "dist" that builds and copy/paste all the necessary files to the "bin" directory.

The executable usage is : ./gpuPlotGenerator <path> <address> <start nonce> <nonces> <stagger size> <threads>
<path> : the path to the plots directory
<threads> : number of parrallel threads for each work group

I am giving it a try now!

[Prelude]

Come on miners! Get your SHA rigs pointed at stratum+tcp://pool.burstmultipool.com:5555

Our profitability stats for the current shift update on the main page every 10 minutes! We need more scrypt miners to start solving some blocks, but our SHA miners are currently nearly 1/3 more profitable than BTC directly.

Why settle for a generic NOMP rip?

Just pointed 1TH at the SHA256 port, and will later switch 150MH at the scrypt port. Any eta on the stats page? Hate not seeing what my workers are doing.

Hey man,

The 'My Stats' page is a bit borked right now in terms of showing you the separate coins.  It is still however displaying your historical hashrates correctly, as well as your estimated number of BURST at next payout.
The entire server is under a constant DDOS attack in excess of 5 GB, thankfully our mitigation service is working well.

A shortcut that can be used to access your miners stats is: http://burstmultipool.com/miner/<WALLET>
so for example:
http://burstmultipool.com/miner/BURST-79PK-DGC2-M4XP-HUAVB


I will get the coins all displaying on it properly later tonight.

OK cool, that's better than nothing!   I just switched over 150MH of ASICs on scrypt.

It seems your pool is only handing out diff 512 shares as max? If so could you increase it to at least 2048? (ideally up to 4096 for A2s)

[n4ru]

How relevant is read/write speed once the plot is generated?

[THeZoiD]

is there a way to change account id from plot file?

[jamoes]

How relevant is read/write speed once the plot is generated?

Write speed is irrelevant after plotting. Read speed must be fast enough to read all plots fast enough to find the next block. If your read speed is too slow, you may miss out on blocks you otherwise would have found. Increasing the stagger size makes read performance better because you have to do less seeks.

is there a way to change account id from plot file?

Nope. Plots are all hashes of your account id. You have to replot if you want to change the account id. You can pool mine (v2) using your solo plots though.

[THeZoiD]

ive put the wrong account Id and ive been mining since Friday I just got back home and notice it


been mining on http://burst-pool.cryptoport.io/ since Friday and im popping shares like crazy

[n4ru]

How relevant is read/write speed once the plot is generated?

Write speed is irrelevant after plotting. Read speed must be fast enough to read all plots fast enough to find the next block. If your read speed is too slow, you may miss out on blocks you otherwise would have found. Increasing the stagger size makes read performance better because you have to do less seeks.

is there a way to change account id from plot file?

Nope. Plots are all hashes of your account id. You have to replot if you want to change the account id. You can pool mine (v2) using your solo plots though.

Would having a bunch of USB docks with HDDs in them be sufficient? Do the drives have to be in RAID and displayed as one drive? Or can I have, say, a USB docking station with 3-4 different drives?

[HoldTheLine]

Hi everyone,

After many hours of setup I finally made it. I have a 1Tb generation in progress and 3x100Gb already finished.
I would like to test the V2 pool but I haven't any BURST for now. Could someone send me 1 BURST to test it please ? Here is my address : BURST-YA29-QCEW-QXC3-BKXDL.

Regarding the plot generation, I found an OpenCL implementation of Shabal (https://github.com/aznboy84/X15GPU/blob/master/kernel/shabal.cl) that could be used to make a GPU version of the generator. I will try to work on it when I have some free time.

Regards

Hi everyone,

As promised I have been working on a GPU plot generator on the last few days. I made a little program built on top of OpenCL, and it seems to work pretty well in CPU mode. Unfortunately, I can't test the GPU mode as it requires a very powerfull graphic card (with at least 46kB private memory per compute unit, because the algorithm needs at least 4096*64 static bytes to store an entire plot).

Here is a preview you can test for now :
gpuPlotGenerator-src-1.0.0.7z : https://mega.co.nz/#!bcF2yKKL!3Ud86GaibgvwBehoxkbO4UNdiBgsaixRx7ksHrgNbDI
gpuPlotGenerator-bin-win-x86-1.0.0.7z : https://mega.co.nz/#!HJsziTCK!UmAMoEHQ3z34R4RsXoIkYo9rYd4LnFtO_pw-R4KObJs

I will build another release in the end of the day with some minor improvements (threads per compute unit selection, output of OpenCL error codes, improvement of the Makefile to generate the distribution directly).
I will also try to figure out another mean to dispatch the work between the GPU threads to reduce the amount of private memory needed by the program.

For the windows people, you can use the binary version directly.
For the linux people, just download the source archive, make sure to modify the OpenCL library and lib path in the makefile (and maybe the executable name), and build the project via "make". To run the program, you need the "kernel" and the "plots" directories beside the executable.

The executable usage is : ./gpuPlotGenerator <address> <start nonce> <nonces> <stagger size>
The parameters are the same as the original plot generator, without the threads number.

If you find bugs or if you want some new features, let me now.

If you want to support me, here are my Bitcoin and Burst addresses :
Bitcoin: 138gMBhCrNkbaiTCmUhP9HLU9xwn5QKZgD
Burst: BURST-YA29-QCEW-QXC3-BKXDL

Regards

Unfortunately, I can't test the GPU mode as it requires a very powerfull graphic card (with at least 46kB private memory per compute unit, because the algorithm needs at least 4096*64 static bytes to store an entire plot).

It's nice to see someone else working on this, since I seem to have failed in it.

Private memory is actually part of global on AMD cards, so storing it in private isn't any better than just using global for everything; it's local that needs to aimed for for the massive speedup. No AMD cards have more than 64KB local per workgroup, which makes storing it all in local impossible however.

I haven't tried your implementation yet, but on my own first attempt, I also used global on everything also, and the result was faster than the java plotter, but slower than dcct's c plotter. My 2nd attempt used a 32KB local buffer I rotated through for storing the currently being hashed stuff, however I couldn't figure out how to get it copied also to global fast enough, and the local -> global copy killed the performance.

You might be interested in those kernels here: https://bitcointalk.org/index.php?topic=731923.msg8695829#msg8695829

Thanks, I will look at your kernels to see if I can find a better solution.

Here is the new version. I reduced the amount of memory used from 40KB to about 1KB per unit. The only drawback is that it requires twice the global memory as before. I will search a mean to reduce this overhead later.
In CPU mode, it all goes pretty well (when no graphic card is detected).
The GPU mode is still kind of buggy on my graphic card (an old GeForce 9300M GS), don't know the exact reason yet. Sometimes it works, sometimes not. I will try to fix this issue tomorrow.

Here are the files :
gpuPlotGenerator-src-1.1.0.7z : https://mega.co.nz/#!iYFWAL5B!BvtmRQ5qGq4gGwjDglFNtDtNIX4LDaUvATBtClBdTlQ
gpuPlotGenerator-bin-win-x86-1.1.0.7z : https://mega.co.nz/#!aBVGBBQD!tBsRtb8VrHR12_anrFTrl41U0fPQu_OqFnxyi5nCyBY

For the linux users, the Makefile has a new target named "dist" that builds and copy/paste all the necessary files to the "bin" directory.

The executable usage is : ./gpuPlotGenerator <path> <address> <start nonce> <nonces> <stagger size> <threads>
<path> : the path to the plots directory
<threads> : number of parrallel threads for each work group

Hey man

It doesn't seem to work. Gets to "building openCL program" and then stops with "gpuplotgenerator has stopped working".

[jamoes]

How relevant is read/write speed once the plot is generated?

Write speed is irrelevant after plotting. Read speed must be fast enough to read all plots fast enough to find the next block. If your read speed is too slow, you may miss out on blocks you otherwise would have found. Increasing the stagger size makes read performance better because you have to do less seeks.

is there a way to change account id from plot file?

Nope. Plots are all hashes of your account id. You have to replot if you want to change the account id. You can pool mine (v2) using your solo plots though.

Would having a bunch of USB docks with HDDs in them be sufficient? Do the drives have to be in RAID and displayed as one drive? Or can I have, say, a USB docking station with 3-4 different drives?

USB 2.0 is probably sufficient based on what people have been saying here. I'm not sure how to test this exactly though...

There is no need to use RAID or any other disc-combining method. In fact, this will make your mining setup more fragile, because if one drive fails, you lose all your plots. Just use separate partitions, and run a separate instance of the miner for each of your partitions. Just make sure none of your plots overlap.

[n4ru]

How relevant is read/write speed once the plot is generated?

Write speed is irrelevant after plotting. Read speed must be fast enough to read all plots fast enough to find the next block. If your read speed is too slow, you may miss out on blocks you otherwise would have found. Increasing the stagger size makes read performance better because you have to do less seeks.

is there a way to change account id from plot file?

Nope. Plots are all hashes of your account id. You have to replot if you want to change the account id. You can pool mine (v2) using your solo plots though.

Would having a bunch of USB docks with HDDs in them be sufficient? Do the drives have to be in RAID and displayed as one drive? Or can I have, say, a USB docking station with 3-4 different drives?

USB 2.0 is probably sufficient based on what people have been saying here. I'm not sure how to test this exactly though...

There is no need to use RAID or any other disc-combining method. In fact, this will make your mining setup more fragile, because if one drive fails, you lose all your plots. Just use separate partitions, and run a separate instance of the miner for each of your partitions. Just make sure none of your plots overlap.

What exactly is plot overlapping? How do I avoid it?

[dcct]

What exactly is plot overlapping? How do I avoid it?

If two or more plotfiles contain the same nonces, they will result in the same deadlines. That won't increase your chances of finding a block and should be avoided.

You can check your files with this tool:

https://bchain.info/BURST/tools/overlap

[coinits]

So what happens if I run a miner for each BURST pool using the same plots / nonces at the same time?

[HoldTheLine]

So what happens if I run a miner for each BURST pool using the same plots / nonces at the same time?

Nothing. You can only win once. It's like buying a lotto ticket with the same numbers from 4 different stores.

[BurstBurst]

Can't get it work and What is reco <threads> for R9 Series Card ?

Hi everyone,

After many hours of setup I finally made it. I have a 1Tb generation in progress and 3x100Gb already finished.
I would like to test the V2 pool but I haven't any BURST for now. Could someone send me 1 BURST to test it please ? Here is my address : BURST-YA29-QCEW-QXC3-BKXDL.

Regarding the plot generation, I found an OpenCL implementation of Shabal (https://github.com/aznboy84/X15GPU/blob/master/kernel/shabal.cl) that could be used to make a GPU version of the generator. I will try to work on it when I have some free time.

Regards

Hi everyone,

As promised I have been working on a GPU plot generator on the last few days. I made a little program built on top of OpenCL, and it seems to work pretty well in CPU mode. Unfortunately, I can't test the GPU mode as it requires a very powerfull graphic card (with at least 46kB private memory per compute unit, because the algorithm needs at least 4096*64 static bytes to store an entire plot).

Here is a preview you can test for now :
gpuPlotGenerator-src-1.0.0.7z : https://mega.co.nz/#!bcF2yKKL!3Ud86GaibgvwBehoxkbO4UNdiBgsaixRx7ksHrgNbDI
gpuPlotGenerator-bin-win-x86-1.0.0.7z : https://mega.co.nz/#!HJsziTCK!UmAMoEHQ3z34R4RsXoIkYo9rYd4LnFtO_pw-R4KObJs

I will build another release in the end of the day with some minor improvements (threads per compute unit selection, output of OpenCL error codes, improvement of the Makefile to generate the distribution directly).
I will also try to figure out another mean to dispatch the work between the GPU threads to reduce the amount of private memory needed by the program.

For the windows people, you can use the binary version directly.
For the linux people, just download the source archive, make sure to modify the OpenCL library and lib path in the makefile (and maybe the executable name), and build the project via "make". To run the program, you need the "kernel" and the "plots" directories beside the executable.

The executable usage is : ./gpuPlotGenerator <address> <start nonce> <nonces> <stagger size>
The parameters are the same as the original plot generator, without the threads number.

If you find bugs or if you want some new features, let me now.

If you want to support me, here are my Bitcoin and Burst addresses :
Bitcoin: 138gMBhCrNkbaiTCmUhP9HLU9xwn5QKZgD
Burst: BURST-YA29-QCEW-QXC3-BKXDL

Regards

Unfortunately, I can't test the GPU mode as it requires a very powerfull graphic card (with at least 46kB private memory per compute unit, because the algorithm needs at least 4096*64 static bytes to store an entire plot).

It's nice to see someone else working on this, since I seem to have failed in it.

Private memory is actually part of global on AMD cards, so storing it in private isn't any better than just using global for everything; it's local that needs to aimed for for the massive speedup. No AMD cards have more than 64KB local per workgroup, which makes storing it all in local impossible however.

I haven't tried your implementation yet, but on my own first attempt, I also used global on everything also, and the result was faster than the java plotter, but slower than dcct's c plotter. My 2nd attempt used a 32KB local buffer I rotated through for storing the currently being hashed stuff, however I couldn't figure out how to get it copied also to global fast enough, and the local -> global copy killed the performance.

You might be interested in those kernels here: https://bitcointalk.org/index.php?topic=731923.msg8695829#msg8695829

Thanks, I will look at your kernels to see if I can find a better solution.

Here is the new version. I reduced the amount of memory used from 40KB to about 1KB per unit. The only drawback is that it requires twice the global memory as before. I will search a mean to reduce this overhead later.
In CPU mode, it all goes pretty well (when no graphic card is detected).
The GPU mode is still kind of buggy on my graphic card (an old GeForce 9300M GS), don't know the exact reason yet. Sometimes it works, sometimes not. I will try to fix this issue tomorrow.

Here are the files :
gpuPlotGenerator-src-1.1.0.7z : https://mega.co.nz/#!iYFWAL5B!BvtmRQ5qGq4gGwjDglFNtDtNIX4LDaUvATBtClBdTlQ
gpuPlotGenerator-bin-win-x86-1.1.0.7z : https://mega.co.nz/#!aBVGBBQD!tBsRtb8VrHR12_anrFTrl41U0fPQu_OqFnxyi5nCyBY

For the linux users, the Makefile has a new target named "dist" that builds and copy/paste all the necessary files to the "bin" directory.

The executable usage is : ./gpuPlotGenerator <path> <address> <start nonce> <nonces> <stagger size> <threads>
<path> : the path to the plots directory
<threads> : number of parrallel threads for each work group

[coinits]

So what happens if I run a miner for each BURST pool using the same plots / nonces at the same time?

Nothing. You can only win once. It's like buying a lotto ticket with the same numbers from 4 different stores.

Just wondered if I would get more shares.

[HoldTheLine]

So what happens if I run a miner for each BURST pool using the same plots / nonces at the same time?

Nothing. You can only win once. It's like buying a lotto ticket with the same numbers from 4 different stores.

Just wondered if I would get more shares.

No you would not.

You don't get paid twice for doing the same work for the same employer just because you send them your bill to two different locations.

[paradigmflux]

Come on miners! Get your SHA rigs pointed at stratum+tcp://pool.burstmultipool.com:5555

Our profitability stats for the current shift update on the main page every 10 minutes! We need more scrypt miners to start solving some blocks, but our SHA miners are currently nearly 1/3 more profitable than BTC directly.

Why settle for a generic NOMP rip?

Just pointed 1TH at the SHA256 port, and will later switch 150MH at the scrypt port. Any eta on the stats page? Hate not seeing what my workers are doing.

Hey man,

The 'My Stats' page is a bit borked right now in terms of showing you the separate coins.  It is still however displaying your historical hashrates correctly, as well as your estimated number of BURST at next payout.
The entire server is under a constant DDOS attack in excess of 5 GB, thankfully our mitigation service is working well.

A shortcut that can be used to access your miners stats is: http://burstmultipool.com/miner/<WALLET>
so for example:
http://burstmultipool.com/miner/BURST-79PK-DGC2-M4XP-HUAVB


I will get the coins all displaying on it properly later tonight.

OK cool, that's better than nothing!   I just switched over 150MH of ASICs on scrypt.

It seems your pool is only handing out diff 512 shares as max? If so could you increase it to at least 2048? (ideally up to 4096 for A2s)
[/quote
Will make that adjustment right now, will require a short stratum reboot.
Let's all hop on #BURSTCOIN

[fivebells]

The GPU mode is still kind of buggy on my graphic card (an old GeForce 9300M GS), don't know the exact reason yet. Sometimes it works, sometimes not. I will try to fix this issue tomorrow.

I would never say never, but I am quite skeptical about the nonce computation ever being modified for major speed gains on current GPUs.  It requires each thread to crawl over every part of a 256k buffer while computing shabal hashes for all of it, and the cores on top-end GPUs only contain 64kb of memory.  I can't see any way to avoid a huge amount memory bandwidth.  This is a different problem than for scrypt (where you can trade the memory for extra computation), in that it actually requires access to the memory, or perhaps some fast way to find shabal preimages.