Pages:
Author

Topic: [POOL][Scrypt][Scrypt-N][X11] Profit switching pool - wafflepool.com - page 72. (Read 465716 times)

member
Activity: 413
Merit: 10
OK, Vertcoin has been added (looks to work), but isn't actively mining, as we don't have enough hashpower to be reasonable yet (would take 1+ days per block, not really worth the variance), it will be auto-enabled when we get to the point that we'd find a block every hour or so (around 200MHs Scrypt-N pool, shouldn't take long).

We've also moved the Scrypt-N pool to all of our endpoints, so you can mine Scrypt-N at any of the normal wafflepool servers (useast, uswest, sea, eu).

We'll be updating the stats pages throughout the day, and when we get to the point that stats are clean and concise, I'll expand the beta to more people (post a note on the site) Smiley
Will there be a separate API Scrypt-N stats? Just wanted to know if I'll need to make any changes to my widget. Thanks
sr. member
Activity: 322
Merit: 254
OK, Vertcoin has been added (looks to work), but isn't actively mining, as we don't have enough hashpower to be reasonable yet (would take 1+ days per block, not really worth the variance), it will be auto-enabled when we get to the point that we'd find a block every hour or so (around 200MHs Scrypt-N pool, shouldn't take long).

We've also moved the Scrypt-N pool to all of our endpoints, so you can mine Scrypt-N at any of the normal wafflepool servers (useast, uswest, sea, eu).

We'll be updating the stats pages throughout the day, and when we get to the point that stats are clean and concise, I'll expand the beta to more people (post a note on the site) Smiley
sr. member
Activity: 322
Merit: 254
Payouts were just run.

For those of you testing Scrypt-N, and had met the normal payout threshold (looks to be 3 of you), you got 2 payouts (assuming you were owed one for Scrypt as well).

I'm open to thoughts on how to deal with payouts as we add more algorithms.  Combining them into one payout is slightly more difficult on the backend, but I understand people not wanting to have a bunch of smaller payouts, and is definitely doable.  For the immediate future (until we're at a stable setup with the new algorithms) they'll continue to be sent as a transaction per algorithm (because that's how its working now, and changing is always more work than leaving alone).


In other news, the beta is looking really good.  I'm going to start adding a few more coins today (you'll likely see some disconnects/reconnects if you're testing), and ideally expand it to the rest of the cluster.

One thing that became apparent from a few users who were testing, is that the interface is not clear (some things are mixed stats, some things are separated), so a major goal of the next few days is to clean up the stats pages to be more clear of what you're looking at. Smiley
member
Activity: 71
Merit: 11
suchmoon and waffle, thanks for the clarification.
sr. member
Activity: 322
Merit: 254
Figured I'd make this a separate post (its a bit more important, didn't want to get it skipped over in the replies post).

Payouts last night were delayed purposefully.  We had a few accounts that didn't have balances showing correctly, and I wasn't able to get them corrected before needing to sleep.  I was reasonably confident they were just display issues (everything in database correct, just caching/displaying incorrectly), but I didn't want to run payouts and add one more layer of things that could make them incorrect (or worse, if they were actually wrong in the database, screw up payouts).

I think these have been taken care of this morning, but want to take a bit longer to confirm.  Payouts will go out as soon as I'm confident in the numbers, sorry for the delay.
sr. member
Activity: 322
Merit: 254
<-- noob.  I've read that scypt-n produces lower hashes, but wanted to make sure I'm drinking the right koolaid.
What we call "scrypt" is actually scrypt(N=1024,R=1,P=1).  Nscrypt (more correctly, adaptive-n-scrypt) means that N gets larger based on a certain schedule (e.g. at block X, n increases).  Currently, all relevant nscrypt coins are at N = 2048 = 2^11.  EXECoin and Vertcoin (in example), will increase to scrypt(N=4096,R=1,P=1) at a certain block, and N will double again (or considering N=2^n, n will increase by 1) at another specific block.

This means that 'nscrypt' (scrypt(N=2048)) takes twice as much memory as 'scrypt', and therefore the hashing speed is reduced by ~50%.

This is exactly correct.  You should expect to get ~50% hashrate you got with Scrypt (and ideally 200%+ profit on those hashes - making things roughly even as a base).


Looking forward to X11

Soon as Scrypt-N is working fully, it should be reasonably trivial to add X11 (it's the next on the list), the only major change is in calculating hashrates (very different difficulty bases).


just started testing...all I get is Rejects:

(Trying not to sound like a dick), this is why I was aiming for people who have mined Scrypt-N before on other pools to test first.  It sounds like a misconfigured miner (probably using Scrypt instead of Scrypt-N?), but we don't have any docs on how to setup a new miner at this time, I'd prefer to get it working 100%, and then I can get some docs/FAQs/startup guides out for others to start jumping in Smiley


I'm VERY interested in trying the new algorithms, but have never mined anything other than "vanilla" scrypt so far.  

The main thing I'm unclear about Waffle's implementation (and here comes the newb question) is do I just have to enable this on the Beta page, and perhaps reconfigure my config.conf file, OR do I have to, in order to be able to mine scrypt-n (and any other additional future algorithms, X11, etc), switch to a different mining software?  I'm currently using BAMT 1.5, which uses sgminer 4.0.0.  I know it's a bit out of date, but I haven't had great experience trying to change/upgrade things in the past, so I left it, once I got it stable on current configuration.

One other thing, if I click enable on the Beta page (which I haven't done yet), and later want to go back (disable effectively), can I do this?

Thanks, very excited about this new development, awesome work as always Waffle! Smiley

The button on the page to "enable" Scrypt-N is just an attempt to make people read that blob of text and understand that things might break.  Otherwise I get a shitload of emails like "OMG EVERYTHING BROKED! FIXPLS".  The actual code for it is live across the board (just a tiny dropdown in the header to switch which algo you're looking at for stats), that "Enable" button just sets it visible for a while (session - a few hours normally).  You're free to enable it and never use it, or enable it, try it, and go back to normal Scrypt.


Correct if I'm wrong, but 1 GPU MH/s Scrypt-N = 0.5 MH/s Scrypt?
And currently BTC / 1MH Scrypt-N 0.00532591 vs. BTC / 1MH Scrypt 0.00267616 would be actually 0.00266296 vs. 0.00267616 on the same RIG, mining Scrypt is more profitable?

That is roughly correct (GPUs will have differences that makes it 55% for some GPUs, 45% for others, etc).  There's also no guarantee that Scrypt-N (or X11) is more profitable than Scrypt, its just another algorithm, with the same sorts of changes in market, difficulty, coins, etc.  I'd caution you that the current numbers for Scrypt-N on WP are not terribly accurate in terms of performance.  You're comparing a reasonably large Scrypt pool (lower daily variance on average) to an extremely small Scrypt-N pool (huge variance daily).  That should change as we get more miners, more coins, etc Smiley


Yep, someone else emailed with the same thing, and I see it on their account.  I'm looking into it and hope to have it fixed in the morning.  My guess is its just something a bit off with stats and caching.

OK. Thanks for looking into this.

In case it helps, I'll PM you my BTC address.

Thanks, I'll check it out.  I made some tweaks this morning (were some problems with caching values as you switched algorithms).  It fixed the stats page for the other guy who was having the problems, so ideally it should be fixed on yours as well, but I'll double-check Smiley
legendary
Activity: 3654
Merit: 8909
https://bpip.org
I'm VERY interested in trying the new algorithms, but have never mined anything other than "vanilla" scrypt so far.  

The main thing I'm unclear about Waffle's implementation (and here comes the newb question) is do I just have to enable this on the Beta page, and perhaps reconfigure my config.conf file, OR do I have to, in order to be able to mine scrypt-n (and any other additional future algorithms, X11, etc), switch to a different mining software?  I'm currently using BAMT 1.5, which uses sgminer 4.0.0.  I know it's a bit out of date, but I haven't had great experience trying to change/upgrade things in the past, so I left it, once I got it stable on current configuration.

One other thing, if I click enable on the Beta page (which I haven't done yet), and later want to go back (disable effectively), can I do this?

Thanks, very excited about this new development, awesome work as always Waffle! Smiley

"Enable" just shows you instructions and allows to access stats for Scrypt-N. You will need to use a miner that supports Scrypt-N (I don't think sgminer 4.0.0 does, so you might want to upgrade to a newer BAMT) and you will need to change the pool port number. To go back to regular Scrypt you would change your miner config.
legendary
Activity: 3654
Merit: 8909
https://bpip.org
just started testing...all I get is Rejects:

Are you sure you're using the correct miner (vertminer, sgminer with the proper kernel, cudaminer, ...) ?
legendary
Activity: 1150
Merit: 1004
Yep, someone else emailed with the same thing, and I see it on their account.  I'm looking into it and hope to have it fixed in the morning.  My guess is its just something a bit off with stats and caching.

OK. Thanks for looking into this.

In case it helps, I'll PM you my BTC address.
newbie
Activity: 30
Merit: 0
Correct if I'm wrong, but 1 GPU MH/s Scrypt-N = 0.5 MH/s Scrypt?
And currently BTC / 1MH Scrypt-N 0.00532591 vs. BTC / 1MH Scrypt 0.00267616 would be actually 0.00266296 vs. 0.00267616 on the same RIG, mining Scrypt is more profitable?
member
Activity: 71
Merit: 11
I'm VERY interested in trying the new algorithms, but have never mined anything other than "vanilla" scrypt so far.  

The main thing I'm unclear about Waffle's implementation (and here comes the newb question) is do I just have to enable this on the Beta page, and perhaps reconfigure my config.conf file, OR do I have to, in order to be able to mine scrypt-n (and any other additional future algorithms, X11, etc), switch to a different mining software?  I'm currently using BAMT 1.5, which uses sgminer 4.0.0.  I know it's a bit out of date, but I haven't had great experience trying to change/upgrade things in the past, so I left it, once I got it stable on current configuration.

One other thing, if I click enable on the Beta page (which I haven't done yet), and later want to go back (disable effectively), can I do this?

Thanks, very excited about this new development, awesome work as always Waffle! Smiley
newbie
Activity: 8
Merit: 0
just started testing...all I get is Rejects:
newbie
Activity: 33
Merit: 0
Looking forward to X11
hero member
Activity: 700
Merit: 500
<-- noob.  I've read that scypt-n produces lower hashes, but wanted to make sure I'm drinking the right koolaid.
What we call "scrypt" is actually scrypt(N=1024,R=1,P=1).  Nscrypt (more correctly, adaptive-n-scrypt) means that N gets larger based on a certain schedule (e.g. at block X, n increases).  Currently, all relevant nscrypt coins are at N = 2048 = 2^11.  EXECoin and Vertcoin (in example), will increase to scrypt(N=4096,R=1,P=1) at a certain block, and N will double again (or considering N=2^n, n will increase by 1) at another specific block.

This means that 'nscrypt' (scrypt(N=2048)) takes twice as much memory as 'scrypt', and therefore the hashing speed is reduced by ~50%.

---

228% vs LTC so far today, with only 2 nscrypt coins!  Variance is obviously going to play a factor with the low hashrate so far, but that's looking darn good.
sr. member
Activity: 322
Merit: 254
Poolwaffle, where should we provide beta feedback? Should it be here in the public forum, or would you prefer PMs?

Either here or via email ([email protected]), either is fine Smiley

It looks like unconverted coins are piling up. Is this just a stats display problem? Maybe I'm seeing it because I'm on the Scrypt-N beta?

Yep, someone else emailed with the same thing, and I see it on their account.  I'm looking into it and hope to have it fixed in the morning.  My guess is its just something a bit off with stats and caching.
legendary
Activity: 1150
Merit: 1004
It looks like unconverted coins are piling up. Is this just a stats display problem? Maybe I'm seeing it because I'm on the Scrypt-N beta?
newbie
Activity: 7
Merit: 0
<-- noob.  I've read that scypt-n produces lower hashes, but wanted to make sure I'm drinking the right koolaid.
hero member
Activity: 700
Merit: 500
Great work PW!  I'm definitely one of the few mining.  

What's everyone using for miner?  I'm currently using vertminer 0.5.3

As for diff, I coded 16 into the password field, but the cmd window is showing 4.
I use the latest sgminer compiled from git. I ran into annoying bugs with vertminer a few times (crashes, declining hashrate, etc). Plus, sgminer let's you hot swap between scrypt and nscrypt pools without restarting the miner.

https://github.com/sgminer-dev/sgminer

Looks like there is even some new auto-tuning functionality that was committed recently.
newbie
Activity: 36
Merit: 0
I'll be switching my GPU rigs over to the new pool as soon as information is available for those of us who haven't done scrypt-n before. Appreciate the hard work in getting this project going PW.
newbie
Activity: 7
Merit: 0
Great work PW!  I'm definitely one of the few mining.  

What's everyone using for miner?  I'm currently using vertminer 0.5.3

As for diff, I coded 16 into the password field, but the cmd window is showing 4.
Pages:
Jump to: