Kano vs Bitsyncom

[Vicus]

which would also normally be done if any of the devs actually had the hardware

You're like a broken record... Change it.

[1715267700]

1715267700

[1715267700]

1715267700

[mobodick]

My position is that rational supporters of Bitcoin would attempt to come up with some middle-of-the-road solution to safeguard the existence of viable competition of multiple vendors in the Bitcoin ASIC business. What I see is almost exact opposite: they are asking Avalon to nearly commit suicide for the sake of an ilusory freedom. Ilusory, because for the gain of few pages of source the whole Bitcoin ecosystem is paying the price of severely disadvantaging one of the ASIC vendors, to the point that in the next iteration we could have a monopoly.

Why would you assume that releasing the source would disadvantage avalon?
What about the modifications is so special yet generic enough it can be applied to a competing product?
You talk about rational people but i have seen little ratio at times from the asic companies themselfs.
How can you rationally manage such irrational behaviour?
And how the hell can you call your pathetic "they are asking Avalon to nearly commit suicide for the sake of an ilusory freedom boo hoo leave britney avalon alone" bullshit anything resembling rational? lol

[nathanrees19]

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.

Yeah, how dare he request that they release the mask sets for their ASICs. At least ask for something more reasonable like the GPL'ed mining software.

[Monster Tent]

IMHO the best proverb to describe this thread is this:

http://en.wikipedia.org/wiki/Cutting_off_the_nose_to_spite_the_face

No, this thread is about the GPL license requirements of cgminer.
It's not about some random person (you) idea about how that fits in with their thoughts on GPL and hardware.

GitSyncom has stated (for the 3rd time ... this time with a date that either just expired or is yet another week away) that they will release the source.

The issue why I brought this argument up was actually mainly due to the bogus excuse as to why they hadn't yet released it.
They supposedly completed 2 Avalon's on the 20th of Jan that contain the results of all of their different source code.
They have supposedly since then been making 12 Avalon a day (each with binaries made from that source code) and yet the bogus excuse was that they needed to remove something from it before they could release it.
Of course there are 2 issues: 1) They can't do that with cgminer and keep to the required licence 2) If they are referring to the non-cgminer code, then well that just means they've now made and shipped over 200 Avalons that have this so called problematic debugging code that is required to be removed - unless the "12 Avalons a day" was bogus.

I can tell you up front:
The cgminer code will be using the FPGA serial-USB library.
The cgminer code will have modified the scanhash is some manner to acquire ~24 work items at a time (instead of 1) coz the Avalon requires a group of items sent to it at a time ... no big deal ... and it's not ground breaking coz it's not even a queue as per discussion about improving performance.
Neither of those are any sort of advancement on cgminer, the first is something I've been removing from cgminer, the 2nd is something that is being developed properly now for the BFL SC's if they appear soon and use it in an optimal way ... as a queue.
The other changes they may have made in cgminer will not be ground breaking in any manner or form.
i.e. there really is no excuse to not have already released the cgminer code by a company who promotes themselves as Open Source to the Bitcoin community.
The results of that are that if anyone else does get an Avalon (which is still questionable more than 3 weeks after they shipped the first one) they are unable to deal with any cgminer problems themselves - and also an already known memory problem cannot be implemented by anyone who gets an Avalon, until the source code is released, without having to reinvent the code (which would also normally be done if any of the devs actually had the hardware ... though I'll be the one to help Xiangfu to do that ... if he does get an Avalon)

I dont think the GPL license says anything about getting free hardware because you commit some code.

[kano]

IMHO the best proverb to describe this thread is this:

http://en.wikipedia.org/wiki/Cutting_off_the_nose_to_spite_the_face

No, this thread is about the GPL license requirements of cgminer.
It's not about some random person (you) idea about how that fits in with their thoughts on GPL and hardware.

GitSyncom has stated (for the 3rd time ... this time with a date that either just expired or is yet another week away) that they will release the source.

The issue why I brought this argument up was actually mainly due to the bogus excuse as to why they hadn't yet released it.
They supposedly completed 2 Avalon's on the 20th of Jan that contain the results of all of their different source code.
They have supposedly since then been making 12 Avalon a day (each with binaries made from that source code) and yet the bogus excuse was that they needed to remove something from it before they could release it.
Of course there are 2 issues: 1) They can't do that with cgminer and keep to the required licence 2) If they are referring to the non-cgminer code, then well that just means they've now made and shipped over 200 Avalons that have this so called problematic debugging code that is required to be removed - unless the "12 Avalons a day" was bogus.

I can tell you up front:
The cgminer code will be using the FPGA serial-USB library.
The cgminer code will have modified the scanhash is some manner to acquire ~24 work items at a time (instead of 1) coz the Avalon requires a group of items sent to it at a time ... no big deal ... and it's not ground breaking coz it's not even a queue as per discussion about improving performance.
Neither of those are any sort of advancement on cgminer, the first is something I've been removing from cgminer, the 2nd is something that is being developed properly now for the BFL SC's if they appear soon and use it in an optimal way ... as a queue.
The other changes they may have made in cgminer will not be ground breaking in any manner or form.
i.e. there really is no excuse to not have already released the cgminer code by a company who promotes themselves as Open Source to the Bitcoin community.
The results of that are that if anyone else does get an Avalon (which is still questionable more than 3 weeks after they shipped the first one) they are unable to deal with any cgminer problems themselves - and also an already known memory problem cannot be implemented by anyone who gets an Avalon, until the source code is released, without having to reinvent the code (which would also normally be done if any of the devs actually had the hardware ... though I'll be the one to help Xiangfu to do that ... if he does get an Avalon)

PS they didn't meet their date they supplied last week of having it this weekend just passed.
My comment a few paragraphs up saying 'a date that either just expired or is yet another week away' was before I checked which weekend he said it would be (on the 8th)

...
Anyhow:

the source code release is set for this weekend on our project timeline at the moment.

...

Even allowing for being in the wrong country (NewYork USA) and adding a whole extra day delay ... it's still past the weekend and no source on the link on the web site or any of the threads I've checked.
I guess it 'could' be a "give the source only to Jeff" ... but as far as Jeff is concerned he doesn't even want it ... and that smells of a rat somewhere with both Jeff and Avalon.

[nathanrees19]

When Avalon is going to disclose their voltage regulator and clock synthesizer programming information

What the actual fuck. Why would fine-grained details about TSMC's manufacturing nodes be present in the modified cgminer?

[kano]

...
I dont think the GPL license says anything about getting free hardware because you commit some code.

Really - you sure of that - you were actually able to read all the words in the GPLv3 license - no words need explaining?
I'm not quite sure why you needed to make an OT post proving that you've learnt to read.
Some sort of milestone you've finally reached?

Also ... let me know when you can actually commit a line of useful code ... some time in the distant future.

Anyway, Xiangfu is getting one as per GitSyncom's statement ... and as I've said a few times ... I don't want one.

[nathanrees19]

Again, I'm just quoting for future reference because

No, you're quoting because you actually think that someone would bother to go back and edit their post after you (literally or figuratively) school them using your vast intellect, and that they wouldn't be able to face being humiliated by you. Since you think that this is reasonable behavior, I shall do the same.

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.

Quoted, because it seems like you actually believe that detailed information about TSMC's processes exists in the modded cgminer.

[nathanrees19]

getting free hardware because you commit some code.

If you keep repeating this, it may become true! Maybe try tapping your heels together to increase the odds.

[tnkflx]

So... Nice weather outside, no?

[punin]

getting free hardware because you commit some code.

If you keep repeating this, it may become true! Maybe try tapping your heels together to increase the odds.

You might wanna be careful with those heel-tricks or you might end up in Kansas

[Bicknellski]

Are you actually complaining about the GPL license? kjj is absolutely correct. Avalon took GPL source, modified it, and failed to offer the source. That is a violation, plain and simple. It doesn't matter if they promised to release it later. It doesn't matter if they are giving up trade secrets to release it. If they don't like the terms of the GPL, they shouldn't have used GPL source. But they did use GPL source, and they have violated the license.

What I'm pointing out is that kjj is preaching from the Free Software Foundation bible in the church of Hardware.

The actions that make sense in software business are frequently suicidal in the hardware business. This is because of the cost structure: hardware is mostly front-end-loaded, whereas software is mostly back-end-loaded.

Yes, Avalon made a mistake by using a code requiring GPLv3 compliance. They should've designed a separation layer like many hardware vendors that support Linux. But the Avalon team is young and inexperienced and they didn't design for that.

My position is that rational supporters of Bitcoin would attempt to come up with some middle-of-the-road solution to safeguard the existence of viable competition of multiple vendors in the Bitcoin ASIC business. What I see is almost exact opposite: they are asking Avalon to nearly commit suicide for the sake of an ilusory freedom. Ilusory, because for the gain of few pages of source the whole Bitcoin ecosystem is paying the price of severely disadvantaging one of the ASIC vendors, to the point that in the next iteration we could have a monopoly.

The rational behaviour would be probably along several possible lines:

a) disclose the code only to Jeff Garzik. He's professionally involved in Linux kernel development and may be able to offer some useful advice on how to both comply with GPLv3 and TSMC/whoever-else NDAs.

b) offer to escrow the code with Bitcoin Foundation and have a programmer at B.F. to produce an obfuscated code that complies both with GPLv3 and NDAs. There are already multiple precedents in escrowing the information with Gavin Andresen, but thus far the escrow was security-related.

c) I personally think that the "binary blob" workaround isn't viable here for purely technical reasons. But I may be wrong. Maybe somebody willing and able to sign the NDA could help Avalon to develop such a solution.

d) take a chill pill and make Avalon folks pinky swar that they disclose the required information after the competition shipped and after the subsequent batches are locked in with TSMC.

But this thread isn't about being rational. It is about militancy, a very short-sighted militancy of playing open-face chinese-rules poker at the traditional-rules poker table.

Think for a moment: What would Richard Stallman do if he had a single tapeout to his name before he received his MacArthur Fellowship? We wouldn't have a Free Software Foundation. We would have maybe a Free Logic Foundation or Free Computation Foundation or something else. But I believe that we would be better off: maybe we could have open-source processors and disk drives in our open-source computers.

+1 Thanks for this clarity.

[organofcorti]

Again, I'm just quoting for future reference because

No, you're quoting because you actually think that someone would bother to go back and edit their post after you (literally or figuratively) school them using your vast intellect, and that they wouldn't be able to face being humiliated by you. Since you think that this is reasonable behavior, I shall do the same.

It is reasonable behaviour. Lots of people edit or delete old or incorrect posts. It's useful to quote for posterity posts a member thinks is important or, for example, we'd never have reeces' wonderful and inventive insults to read.

[2112]

Well, the feedback from the gentle readers

2112's post sounded awfully cryptic

and not-so-gentle readers

Quoted, because it seems like you actually believe that detailed information about TSMC's processes exists in the modded cgminer.

is that my "Teach yourself IC design,fabrication&test in 21 minutes" lecture is too hard and sounds like black magic.

I'm going to quote single linear thought from my post to maybe make it easier to follow.

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.

The technique I'm talking above is called "yield estimation using test data". SHA-2 is 100% self-testing and trivial to reverse-engineer. Competent semiconductor manufacture engineer can with the help of changing clock frequency and supply voltage obtain a highly proprietary data in a completely non-destructive way (no chip desoldering, decaping, etc.)

An observant reader may ask "why neither Intel nor AMD seem to care about chip with unlocked clock-multiplier and voltage identifier". The answer is "binning". A large manufacturer will do an extensive test of their chips and sort them into bins. When they sell the "enthusiast-grade" chips with unlocked clock they sell them from the "fastest process corner" bin. All other bins are clock-locked and sold cheaper into OEM market. By "bin sorting" the manufacturer can completely obfuscate actual process parameters and make competitive yield estimation pointless.

On the other hand Bitcoin ASIC vendors cannot afford detailed chip testing, both because of financial and time constraints. Any chip that passes quick needle-test on the wafer prober will be packaged and mounted in the shipping product. This situation gives the analyst the sampling of an entire defect/yield curve for the fabrication process.

As far as I know the most commercially/competetively interesting information is obtained by testing the worst chips, those from the bin nearest to the "trash bin", the ones that barely met the specifications. In fact the content of the "trash bin" is quite valuable to the competition and therefore each fab carefully destroys the chips that failed the acceptance tests.

I hope that the above addition my posts will look more like grey magic than black magic. Please do some web searches about "yield estimation chip" and read the freely-available information.

[maqifrnswa]

Is 2122's argument, "because it is inconvenient to my business practice I will violate the license agreement?" So if I out-license a patent to you for a fee, will you just stop paying me because it's better for you to not pay me than pay me?

You can take that position, but license violations open you up to lawsuit liability (cease and desist at a minimum, and damages at a maximum).

If you don't believe me, google Aaron Swartz...

Also, Avalon agreed to the terms of the GPL by using the software. No one forced them to use it. If they don't want to disclose code, then just write your own miner software (like BFL did with easyminer...)

If you want to use cgminer, then follow the terms of the license. The alternative is not use cgminer.

EDIT: I'm in small business hardware and software. I sometimes choose GPL and sometimes not, depending on my strategy. Avalon can do the same.

[PuertoLibre]

getting free hardware because you commit some code.

If you keep repeating this, it may become true! Maybe try tapping your heels together to increase the odds.

Selective blindness is always interesting.

[The-Real-Link]

Thanks for the clarification there, 2112.  I didn't type that it was cryptic though because I didn't understand your terms or the concept of binning after the fact, I only put it there as saying that Avalon is not in a good position at the moment with respect to their licenses apparently.

Good post Maqi.

[nathanrees19]

Competent semiconductor manufacture engineer can with the help of changing clock frequency and supply voltage obtain a highly proprietary data in a completely non-destructive way (no chip desoldering, decaping, etc.)

I do see what you're saying now, but you're still assuming that fine-grained control of individual chip frequencies and voltages (since you'll have a hard time getting good data if you can only do a whole board of 80? chips at once) exists within the modded cgminer and not an FPGA/MCU blob and also that this can't be trivially replicated by monitoring the usb-serial comms and also that someone actually has an Avalon unit (lol) that they don't mind risking for the tests (lol) and also that someone actually cares that much about this data to go to the effort at all.

It's a bit of a long limb, especially since Avalon wouldn't have picked GPL code if this was a serious concern (or they're license violating assholes, but PL has assured us that they aren't).

[-ck]

The one thing I don't understand in all this flameage is why Jeff Garzik isn't championing the request for source from Avalon. Jeff?

[jgarzik]

The one thing I don't understand in all this flameage is why Jeff Garzik isn't championing the request for source from Avalon. Jeff?

After certain people made that task considerably more difficult?  It is difficult to summon the motivation.  Ask again in 30 days...