They had better ship all of batch 1 before the whole company goes on a 1 week vacation. Otherwise there will be some very angry early adopters.

Bitmain has updated their site. The AntRouter test is gone, and they've added batch 2 and 3 of the S7.

Interesting that batch 2 is 4.66TH/s for $1585 and batch 2 is 4.86TH/s for $1658. Both ship in the same date range.

So I guess you pay a little more for better performance? Have they ever done that before?

In the past separate batches would have separate ship date ranges, not slightly different specs.

I just wish they'd ship the rest of batch 1. Still waiting for my second order from Sept 5.

Playing with your stick huh?

Are you sure you're on the right forum?

Sorry, maybe I shouldn't have had that second glass of wine tonight

Beat me to it

For completeness sake though, when doing a new install on a Pi it's a good idea to go through the full upgrade process, like this:


On a resource limited machine like a Pi this can take a really long time.

You're welcome. Post some picts when you've got your rig set up.

The only downside of egg crate is that when you cut it, you can end up with a jagged edge. You can scrape yourself on it if you're not careful.

Currently both of them are running at 306. That results in about 16.9 GH/s with 0.08% HW errors.

I would have gone higher, but one of the units is a little weaker. I had to turn up the voltage proportionally more on the weak unit to get to 300-ish.

I probably could mess around a bit more with the voltage and frequency to go higher. Or I could explore per-unit frequency options since one stick outperforms the other. Assuming that cgminer-gekko can in fact do that (I don't know).

In the end it seemed good enough to get double the stock hash rate for the amount of fiddling I did.

Egg crate is made of stiff polystyrene plastic, so it's easy to cut. I typically use diagonal cutters. The plastic easily splits along the cut line. Very easy to work with. Got the idea to use this stuff from a very different hobby (salt water aquariums).

I bought a big panel a couple of years ago and have built out three small scale rigs out of the stuff. Usually I just tack down the power supplies, hubs, etc. with zip ties. Once the components are in place, the rig is effectively portable.

It's typically used as paneling for lighting fixtures, especially comercial fluorescents, in the US at least. Any building supply or lighting company should have it for pretty cheap. If you search for "egg crate plastic" you'll find a ton of options.

I'm a big fan of egg crate for smaller rigs. It's not pretty, but it's super versatile.

Here's what my stick rig with Y cables looks like (repurposed from running a bunch of DualMiners recently):

This SP20E is a super reliable workhorse. Its entire career has been spent running at stock speeds in a dust free, air conditioned data center.

This beast has been pampered by a cold environment. Input temps have been around 20 decrees C or less. Currently the unit is indicating 18 degrees C on the input sensor.

Looking for $320.

You supply the shipping label from the US. I'll ship in the original box for safe arrival.

If you want escrow, we'll use OgNasty and you pay his fee.

Here's a picture (excuse the bad phone camera quality):

OK. We'll keep things simple and as-is for now.

We can figure something out that works for everyone once we're on the other side of the 30 days.

In the meantime, I'll give this some more thought. But I'm back to thinking that a shared dropbox link of an unformatted text file is the simplest approach. It's easy to share the file for editing. And following redirects in Python is no problem, so the notification script can easily read a dropbox file. Under the hood, I might use a list like this myself just to make it easier to update my instance of the notification script. I can share the link with TheRealSteve if he's interested.

Anyway, we can return to this when the time is right...

In the meantime, I finally got my Y connectors and time to use them. Yesterday I doubled the output of my 2 sticks to a whopping 33.8 GH/s!

Originally I thought the text file would be unformatted with one worker address per line. But if you need to tell donation addresses from regular ones, then maybe we need a simple JSON format that separates the categories.

That means Phil would have to be careful to add addresses to the right places. But I'm sure he's edited tons of config files in the past, so this would be no different.

No problem. I code for living so I'm sure we can make this work

Phil, what I'm thinking is you'd have the master text file that would list the club members' worker addresses. When a change needs to be made, you'd edit the file and then post it.

I can probably provide a simple bash script to do the posting to make that easier.

Periodically the notification script would run and get a fresh copy of the posted list via http. I assume Steve's IRC bot can do the same. The effect is that changes would be automatically propagated.

Anyone else who wants to see the list can just click on the link in their browser, or use the IRC bot, which would show the latest stats.

Sound good?

The question is where to post the list. I played around with Dropbox, but in order to support raw files they employ a redirect, which is kind of a pain.

Unless someone has another idea, it back to looking like the simplest way to do this is going to be for me to set up a VM Web server that supports receiving the file from Phil.

Unfortunately I don't know much about google docs. I could learn though.

But if you're looking for a current display of the club members and their stats, doesn't TheRealSteve's IRC bot provide that?

One thing that I'm thinking about is solving a somewhat different problem, and that's management the list itself. If we've got several different ways at getting at the data, then there are several different people managing the list of workers and keeping it current.

Maybe what we need is a shared text file (something easily parseable, like one worker address per line) that one person (probably Phil) can post somewhere and the other consumers can access. If that existed, I could extend the script to dynamically get the worker list from this posted text file. TheRealSteve could probably do the same with his IRC bot.

If this sounds good, I'm fine with setting up a simple web server that would host the text file. We have an ESXi host at the office that has spare resources. I could give Phil access to it so that he could post new copies of the text file (probably via scp). The URL would be public so that anyone can grab the current list via a browser.

Does this make sense, or does anyone else have alternative suggestions?

It worked for me on 10.10 Yosemite. But I wouldn't update just yet (if you're thinking about it). Better to wait for El Cap, which is coming soon.

I logged on and checked out the bot output yesterday. Very cool. Unfortunately I couldn't stick around since it was a workday...

Note that notification script also stores the best shares value locally in a stats file. So it won't be reset either when the pools stats reset.

I agree with you that best share not a meaningful statistic. Although it's fun to cheer your miners on to do better

The problem is that I don't think there's a direct interface to get information about found blocks from the pool. Currently the script uses an indirect method.

Namely, the script queries blockchain.info for the pool's fee address (1PKN98VN2z5gwSGZvGKS2bj8aADZBkyhkZ). The assumption is that if there's a new transaction from this address, then the recipient must have found a block.

While this heuristic approach works (I just got a notification email this morning that someone else found a block), it means that we really don't have any visibility into the worker that found the block.

Given the low likelihood that we'll find a block, I doubt that CK would be too thrilled about extending the JSON API to provide some sort of block history. Maybe instead we could just rely on CK to tell us which worker found a block, assuming that that information is logged somewhere on his server.

+1! I've had good experiences with FedEx too.

My first S7 order was exactly like this. Took 3 days and was delivered before 10:30 AM, just as FedEx said it would. No delays and no drama.

Of course carriers are like hard drives manufacturers. Everyone has a different favorite because of personal experience.

Thanks for the favorable mini-review SlimePuppy!

Notlist3d, the current cgminer-gekko in the image does not support U3's yet. I put that on hold because there are compile problems when you enable both gekko and icarus.

Novak, have you managed to get to this yet? Maybe you fixed the sources and I missed it?

I've updated the CK solo pool notification script with the following changes:

• When notified of changes (like a new block found or best share improvements) the hash rate stats of the affected monitored addresses is also reported. So if the club finds a block, the hash rates of all the monitored workers will be reported.
• If the --verbose option is on, the email text will be logged locally. This could be useful if the email can't be sent but an important event like a found block happened. In this situation, the email text could be pulled out of the log.
• Fixed some problems sorting mining addresses.
• Cleaned up the formatting of the email body text.
• Added an testing/developer option that allows the caller to pretend like a specified address found a block. Previously I had to hard code script globals to force block found testing.

Here's the github URL for anyone who's interested in playing with the script:

http://edonkeycoin.github.io/ckpoolnotify/

I've tested the latest version with the current list of club workers, and it seems to work fine. I was a bit concerned that a large number of workers might make the command line too big, but it seems fine for now (on the Mac at least).

If you've followed the read me instructions and gotten your email credentials to work with a test email, then a command line like the following will monitor the list of known workers at this time (substitute your own email credentials, of course):


At some point I'll support a config file to make it easier to manage a large list of workers.

If you're using the script and you want to update to a new version, all you have to do is stop the current script, replace it with the latest, then start it up again.

If you're on the Mac and you opted to set up the script as a daemon, then you'll have to unload and reload the daemon. Here's the basic command lines to use (replace the path to the new script before using):


One odd thing I noticed while testing the latest is that none of the pool workers have a 7 day hash rate average. I don't know why this is. Maybe if there are any club members that have less than 7 days in, then none of the 7 day averages for the others get reported?

It might not matter if we rely on the 1 day average, or if we add a mechanism to remember a few days worth of hash rate (which Phil suggested).

The S7 is here! Thank you FedEx!

Luckily it's destined for a data center, so we won't hear it.