Topic: Kano vs Bitsyncom - page 7. (Read 15357 times)

Avalon customer service now brought to you by Inaba.

... No FUD please.

Wow, phrased this way it sounds an awful lot like ckolivas too.

Coz you are making an absurd claim that is false.

I too have been really irritated by kano and his "I'm entitled to free stuff" attitude.

But it goes beyond mere irritation; when he doesn't get free stuff he actually makes up false accusations in order to get attention.  For example, he made the absurd claim that the TML host software was derived from cgminer.  Of course this is ridiculous, and all of the source code for the TML host software is publicly available, every single line of it.  And it's written in a completely different language for crying out loud.

I still haven't gotten an apology from him.

Kano, you forgot to cite the hole in his code that steals 1/100000 hashs and sends it to his account so he can steal the equivalence of bit penny.....  Open Source or None at all

Click & snap again.

I mean kjj is kinda lost-cause here, he isn't even aware that he's at a poker table and laying your cards for all to see is not a winning strategy. But I know that this board is read by many young people who are capable of learning.

Hardware development is like a poker: you keep your cards close to the chest, maybe drop some and add some, place your bets and wait for the showdown. If you show your cards to other players before showdown you are going to lose (or at least you aren't going to win anything). And if you lose all your stake you will not be allowed to sit at any table. If you wont be able to front the money for the cheapest table you'll be just a beggar waiting for handouts from the real players at the casino's entrance.

I'm not sure how much TSMC values the non-disclosure about the manufacturing node that Avalon used. But before they had their chips manufactured by TSMC they had to sign something about obeying reasonable care to avoid disclosing TSMC-proprietary and whoever-else-proprietary information. SHA-2 is an example of a self-testing structure, something akin to the test structures used in the manufacturing process testing and calibration.

When Avalon is going to disclose their voltage regulator and clock synthesizer programming information it will allow competent people to obtain very detailed information about TSMC process used. I don't think that theres much commercial value in that, but it is the intentions that count. Avalon signed not to disclose, but allowed disclosure through carelessness. TSMC aren't going to be thrilled about it and will drive harder bargain when Avalon tries to order the 2nd batch. I'm not expecting somebody from Chronicle Technology open the account to post "Thanks, suckers.", but maybe some of the Avalon competitors will do that.

So Avalon is now in between the hammer of GPLv3 and the anvil of NDA with TSMC.

This concludes this my short lecture. If you plan to ever in your life play in the high-stakes game of hardware development: learn the rules. Otherwise you'll be forever fighting over the table scraps and leftovers: like Raspberry Pi where Broadcom/Alphamosaic Videocore GPU boots and controls the ARM CPU sandbox to let the kids play with their open cards poker game.

Why would I edit or delete my posts?  You are the insane one here, not me.  Heh, ok, calling you insane might be a bit over the top, so I may edit that to a nicer term later.

They used software that requires disclosure of changes.  No one forced them to use it.  They made the decision voluntarily.  They had the right to write their own software and keep it secret.  Hell, if they really wanted to, they could have tracked down all of the authors and negotiated a different license for the same software.

I'm no fan of copyright as applied to non-commercial distribution, or even merely copying, for that matter, but it is the world we live in.  And in this world, I support the rights of authors to game the system to preserve freedom, and I care a hell a lot more about that than I do about "competition in hardware business".  You are the one with the short thinking horizon here, not me.  You seem to care more about next year's products than about the next generation's freedom.

I think I've clarified my position on this sufficiently now.  The products/freedom issue is what divides the open source people from the Free Software people, and I think it should be plenty obvious which side both of us are on.

Oh dear god, not the manufacturing yield!  Anything but the manufacturing yield!

Seriously dude, get a grip.  The mining programs already spew FPGA bitstreams as opaque blobs provided by the manufacturer.  If the Avalon is using an FPGA for management, that blob is either on a ROM chip on the board, or in the filesystem of the control board.  No one cares.  What we do care about is knowing what they did to cgminer (GPL!) so that it could talk to the management module.  Plenty of us want to throw out the crappy wifi router they are using and connect our units, when we get them, into our existing infrastructure.

If there are big secrets in the patches they made to cgminer, then they fail because now that they have distributed it, they are required to disclose those changes.  If they didn't like that, they should not have used that software.

Again, I'm just quoting for future reference because this board allows ulimited message editing/deletion.

This is an example of how defense of international software copyright treaties kills competition in hardware business preventing the startups from recouping the NRE costs.

The biggest enemy of Bitcoin aren't banksters or whatever else powers-that-be. The enemies are the hormone-laden cholerics that simply cannot think on the horizon longer than a month or (rarely) year.

They were in violation of international copyright treaties the moment they shipped.  The GPL doesn't say that you must intend in the future to release the source code, it says that the physical product must be accompanied by either the source code or a written offer to provide the source code.

By default, there is no right of distribution.  They only way to get that right is through a license.  The GPL provides an automatic license to people that comply with the terms described.  Failure to comply with those terms = copyright violation and termination of license.  Section 8 provides ways to restore the license, but it does not excuse a violation and does not give a grace period during which violations are acceptable.

At any rate, I don't really care much about violations.  I had two points, the first being that 2112's notion that the GPL is an evil thing, strangling poor projects in their crib is nonsense, and the second that there was not, is not, and never will be, a good reason for Avalon to have failed to provide the software alongside the physical product.

I'm just quoting for posterity. This is a perfect example of a supporter that is worse than an enemy. Not understanding hardware is not a problem. Not understanding hardware and pretending to understand it is the most serious problem: for example the overclockability curve can be used to estimate manuafacturing yields.

Involvement of the people like kjj is the reason why so many hardware-related open source projects fail: they inadvertently disclose all the competitive information because they simply don't understand the manufacturing technology and planning: front-loaded NRE costs rule. The competition can run circles around them: a knowledgeable competitor using proper analytics would know more about their manufacturing than the ostensible project managers do.

This isn't a software business with no barrier to entry and where the costs are back-end loaded: mostly in the maintenance. Even if you don't understand this now, just copy this and paste it somewhere for the future reference. There is also a lot of similar discussion from they days where various people from around Linux Torvalds discussed merits and demerits of various licenses. In Bitcoin you have all that distilled to just a handfull of projects.

That would only be true if Avalon said they will not release source code; but they have always maintained that they will release source code.  At least that's what I was told.

Based on that, the only question is the date of release, and by extension, the patience level of forum posters

Bullshit.  There are no secrets here.  SHA is totally known.  Bitcoin mining is totally known.  USB is totally known.  There is absolutely fucking nothing in an Avalon unit to be protected by secrecy.  The hard part here is actually making the damn chip, not the LOL-programming they use in the FPGA that manages them.

Moreover, your notion of GPL "turning virulent" and killing the host is bullshit too.  If they didn't like the conditions, they didn't have to use any of it.  Not releasing their changes is just shitting on the community, and as I mentioned before, pointless.

Well, I'm thinking that Avalon people didn't have time to write a proper firmware and device driver. They are probably using the generic drivers from the Linux kernel. Then all the initialization/control/stabilization code is in the modified cgminer, which is all GPLv3. Avalon probably heavily modified the AMD/ATI fan control code from the GPU miner and USB-interface code from the FPGA miner.
The easiest it would be if only fan-control is used after initialization and voltage and frequency settings are hard-coded in the factory. But if they have the fully dynamic clocking like eldentyrell does in his tricone FPGA miner then it would be nearly impossible to split all that functionality into a binary blob or whatever else.

ACPI folks had years of experience with their technology and build up on the past experience with PnP etc. And even they still freqently can't get it right.

The kernel stuff is GPLv2.  Only cgminer is GPLv3, AFAICT.  Definitely a difference there, and there is plenty of de facto precedent for binary-only gadgets in the kernel -- even if hardcore Debian-legal denizens disagree with that legal interpretation/result.


There are any number of existing bytecode solutions that could be employed, to provide a sort of "firmware" that is executed by the host CPU.  Modern ACPI functions like this, as does http://en.wikipedia.org/wiki/Open_Firmware

Of course, that tends not to be performance-heavy code, but instead critical bootstrapping and initialization functions.

This is minor leagues kind of fun.

The real fun would (or will) start when the people involved could afford a real legal counsel. Imagine what would happen if a successfull preliminary injunction would get filled against importing Avalon units into USA. That would be like real fireworks.

Obviously I'm very curious what is in the unreleased source code in the Avalon driver. Fan-speed control is probably nearly open-source. But the voltage regulator programming and clock synthesizer programming may hide real secret information. Even if the code gets released the people involved may regret the ultimate results of the disclosure.

I understand that Avalon simply had no time to develop a proper firmware layer to isolate themselves from the results of adverse disclosure. They may be just another project snuffed by the situation where GPLv3 turns virulent and kills the host.

Anyone here has any constructive suggestions how to make the hardware driver detail disclosure non-adverse? Some sort of obfuscation scheme with programming a multitude of non-existent regsters with multiple pages full of hex values? Is there a quick way to distinguish which registers are real and which are fake?

Ah, I see now it actually ends where the picture ends. No katana indeed. I DID think the handguard looked very unjapanese. We need a weapon porn thread.

Nobody fucking cares about getwork/cpu mining anymore. cpuminer is a toy for people who want to lol at their linksys router mining at 20khash. considering how much has been rewritten/replaced with something less shitty, cgminer might as well have been forked from windows 98.