Topic: Recieved my Jalapeno! Order # 3486 (Read 4812 times)

Josh stated that they will use a firmware applied to selected chips.
From Shoutbox:

I don't disagree here either. But I doubt that they're taking the time to 'tweak' firmware in each unit. I believe that end users should theoretically be able to tweak their firmware for their individual units. However, the tweaking will be individual to each unit (just like buying multiple GPUs and getting different hash rates out of all of them). I don't think that there's any one specific firmware flash that can just be flashed on everybody's Jalapenos, and achieve the same hash rate.

Again, I fully admit that it's POSSIBLE that it's merely a firmware mod that they're using to make these things hash at 7G/H, I just don't think it's conclusive, or should be relied upon as being such.

I am 99.9% certain that the chips have a programmable clock rate. So the firmware is indeed setting the clock. I am also fairly certain that the firmware can accept a clock rate from external to the device (otherwise testing would be horrifically awful). So if someone reverse engineers what those inputs are (no inputs = default clock rate) then everyone should be able to tweak their Jala's to a higher hash rate.

Another rare point where K9 and I agree. We don't know what the yield of wafers was regarding grades, we only know what the average yield would be in raw chip orders, of 100 units. Not much of a sample size to go on. They're obviously going to be binning these things BEFORE they ship, with such small order sizes. I think if they were shipping lot sizes of 10k chips, right outta the foundry, your numbers would be a reliable benchmark. However, in this instance, the low numbers of Grade D chips could easily indicate that they're being held back and utilized because they come up to a nice 5g/h range.

Further, if they were just limiting hash rate intentionally somewhere in the firmware, we'd see Jalapenos in the wild all hashing right at exactly 5.0g/h. I haven't run across any yet that are hashing at less than 5.0, but I've seen several that were hashing at more. One of the more well known ones being the one that was removed from the case, and is hashing at 5.9, as long as it's kept cool (and seems to slow down, to I believe 5.4g/h if it gets warm).

This indicates to me, 2 things:

1. There's a hardware limitation on how fast these chips can hash, based on the temp they're cooking at.
2. There's nothing specifically setting a hashrate ceiling at 5.0, per se.

Now, you will probably say that they're likely limiting clock speed in the firmware. OK, that's POSSIBLE, but if it were as simple as simply upping the clock speed, they'd be introducing more heat, which would in turn have a NEGATIVE effect on things, due to the fact that a guy running a bone-stock Jalapeno is losing hashing speed when he puts his unit back in the case, moderately reducing the cooling efficiency of the thing.

Of course, we're all just speculating here, and my guess is as good as any, but I think people here are seeming just a little too eager to jump on the idea that this is simply a firmware mod, and not a chip selection issue. There's a fair amount of evidence to support speculation that this is a hardware limitation, rather than an in-built firmware limitation.

Although I'm entirely open to the possibility that it's a poorly-implemented firmware limitation, I'm not leaning towards it, and I think people ought to be cautious about accepting the firmware explanation as gospel, and/or repeating it, as though it's conclusive.

Doubt it. I am not totally sure that both chips are active, or if one is a dud for balancing the heatsink.

They were originally designed to do 5GH, most likely with some overhead. It could be this overhead that is taken out(if single chip able to do 7GH).

The other scenario is that the chips are underperforming, and did not do 5GH. In this scenario, they are still likely to be able to do 4GH or more each.

Methink that soon someone will have overclocked their jally and we'll see what it can do with a little tweaking.

Edit: My points did not come trough.

Unless that is the distribution for sale as opposed to the yield of the binning. For instance, they might be reserving D lot chips to go into Jalapeno's because they add up nicely to 5 GH/s.

Okay, I am back.... Check out the specs on BFLs chips.

https://products.butterflylabs.com/65nm-asic-bitcoin-mining-chip.html

" Chip grades:  Chips come in four grades of performance.  Chips are sold in mixed grade lots.  A grade has 16 engines, B grade has 15 engines, C grade has 14 engines and D grade has no less than 12 engines.  All chips run at a minimum of 250 mhz.  Higher grade chips will run up to 294mhz.  The percentage distribution in each lot is 60% Grade A, 20% Grade B, 15% Grade C and 5% Grade D."

You have a 90% chance of having ZERO Grade D chips in the standard Jalapeno.
Remember Jallys now have 2 chips.

Running them both at 250 mhz will mean that there is an average output of at least 7.5 GH/s.
(I used an average of 15 engines per chip, at 250,000 hz per engine)

BFL has the Jally cranked down to run at 5.5 GH/s.

Mmmm, lunch. I think I'll do the same.

Wrenchmonkey,

From my thinking, due to recollection, which could be faulty, it is in all likelihood a firmware upgrade.

:-)

I will have to do math in public, but I will investigate and report back.

Gimme a little time, as I am eating lunch at the moment.

What makes you so certain it's a firmware? How do you know they didn't select the higher performing chips for this 'upgrade'? It would seem to make sense, considering that they're now using two chips instead of one, that it could be unlocked with firmware, but I haven't seen it confirmed that it IS firmware. It could very well be that they're just using higher-grade chips to achieve the better hash rate.

Has anyone hacked the 7GH firmware yet?

Congratulations, you are a lucky miner.

Handy little tool!

Thanks for the time you put into it.

That was one of the funniest things I've heard in a while.

I'd tip you some but you don't seem to have a BTC address…

I maintain this list manually by monitoring the forums over at BFL_Jody's blog and then marking them delivered when she announces which dates have shipped.  So looks like I got one right. 

But still gonna hate and try to mud every BFL thread out there 

Let me just mix a little data in from http://bfl.ptz.ro/

Total SC Jalapeno	Total Qty Little Single	Total Qty SC Single	Total Mini Rig	Total
2588	934	1970	138	5630
11,646 GH/s	28,020 GH/s	118,200 GH/s	207,000 GH/s	364,866 GH/s
$385,612	$606,166	$2,559,030	$4,126,062	$7,676,870

364,866 GH/s owed by BFL before this thread.
364,861 GH/s owed by BFL after this thread.

I think the people getting pissed are the investors who are owed Little Singles, SC Singles, and Mini-Rigs.
According to Josh in 90 days 3% of outstanding BFL orders will be filled when they finish shipping Jalapenos. 

Congrats. Nice to see more of these things in the wild. The haters are getting soooooo pissed. 

Congratulations!
My Jala is still under delivery by USPS.

that is not correct at all.

he claimed to get about 200 mhs more (5.6 - 5.7 to 5.8 - 5.9 ghs). check here https://bitcointalksearch.org/topic/m.2336656

I did the same and there is no difference at all with mine. its 5.5 steady.