Monday, February 23, 2015

Why Open Source Has Not Taken Over EDA


We all know that in the world of general software Open Source has all but won. Linux is everywhere. Nearly all programming languages are Open Source (meaning their compilers/interpreters). IDE's, build tools, revision control, syntax highlighters, refactoring tools, you name it, if it's a development tool it's Open Source. Major infrastructure is Open Source too. I mentioned the Linux operating system, but all the major applications are Open Source too: web servers, databases, queueing systems, messaging systems, load balancing, caching systems, GUI frameworks, encryption, authentication, email servers, instant messaging servers, blog engines, you name it.

In the world of Electronic Design Automation (EDA), Open Source has not won. We do now develop on Linux and we use a lot of ancillary tools from the software world that are Open Source such as scripting languages, text editors, databases, web frameworks, build systems, and so forth, but our core tools are still very much closed source and commercial, namely simulation and synthesis tools.

Why haven't those tools followed the same trend that the general software world has seen? Why don't we have a solid Open Source simulator that every one in the industry rallies around, similar to the way gcc is the defacto standard C compiler? Every time I have posed this question there are a few answers that always come up. One is that our industry is small. There are far more software developers, the argument goes, and so the likelihood of a Linus Torvalds or Richard Stallman like leader emerging is small. Another answer is that we aren't software developers. Our primary skill is designing circuits, not software, therefore we just don't have the skills or drive to write our own tools like software developers have done.

While the logic in those answers seems sound, I don't believe they are the primary reason we still lack quality Open Source tools. There have been hobbyists like Linus and RMS that have started Open Source EDA tool projects that are probably at least as functional and easy to use as early versions of Linux or gcc were, but people just haven't flocked to them. Yes, the number of EDA people that are capable of writing a synthesis tool is small, but how many software developers really have the chops to write operating systems or compilers? Also not very many. There are far more consumers than producers of general Open Source software. I don't believe that's the reason there isn't any successful Open Source EDA tools.

I believe the real reason we don't have Open Source tools is because we haven't had an oppressive monopoly in EDA like Microsoft was in its day. Microsoft was so good at locking people in to their proprietary tools and crushing competition that the only answer was to not play their game. You had two choices, go the Microsoft route or go Open Source. We in the EDA world, on the other hand, actually have a (somewhat) functioning capitalist software economy. There are three (three! not just two like Coke and Pepsi) big EDA companies that sell competing software (these companies are lovingly referred to as, "the big three"). Unlike the days of Microsoft, these companies actually have to listen to their customers[1] and compete on the merits and price of their products[2]. Since we have been able to play the big three against each other and generally get at least the bare minimum of what we need, there hasn't been a strong need for a linux or gcc like project to set us free.

Hooray for capitalism! Yet somehow I'm not satisfied. I am very pro capitalism in general, don't get me wrong. I think my dissatisfaction comes from the fact that I am even more pro freedom, and EDA tools are not Free. They are not Open Source. I'm not free to dig into the code if I desire to. I'm not even free to use the tools as I chose. They are offered under onerous license agreements that don't even allow us to publish performance numbers or talk about their costs publicly[3]. When they don't perform or I encounter bugs, my only recourse is beg for mercy from the vendor or embark on the onerous task of switching vendors (just because it's possible doesn't mean it's easy). If it were open source there would be mailing lists and forums where not just the vendors customer support people participated, but other tool users and the tool developers themselves would be available to collaborate with on solutions.

This would sound like an outrageous and silly utopian dream if it wasn't already working in the software world. It's not just young single dudes living in their parents basement or slumming it in graduate cubicles at MIT anymore either. Red Hat Software is a publicly traded company making 11% profit margin on about $1.8 billion in revenue, which puts it right in among the Big Three. There are plenty of other businesses that make real profits doing various combinations of producing, supporting, and providing services based on Open Source software. It can be done. In fact I'm pretty sure if one of the Big Three even released just their simulator as Open Source they would quickly grab all of the market share for simulators and a whole lot of new customers for their other tools. They would also make the world a much better place for all of EDA.

Footnotes

1. Well, at least their biggest customers.

2. Well, not publicly, because their license agreements don't allow us to publish their tools' merits or prices.

3. See footnote above.

Friday, February 20, 2015

Bitcoin vs. Credit Cards

Stripe announced bitcoin support.  There has been much discussion on hacker news about it.  Lots of people are not seeing the benefits of bitcoin over credit cards for individual shoppers. They talk about the fraud protection that credit cards provide them as customers and the 1% or whatever cash back they get on each purchase from the credit card companies that they would be giving up.  I don't think they fully understand the trade-offs here.

First of all, the cash back.  Credit card companies are charging merchants a percentage of each transaction larger than the 1% whatever they are giving you back (how else do they stay in business?).  What people may be forgetting is that retailers are not just eating that percentage cost that the credit cards are charging them.  They are most definitely passing that cost on to all of us consumers in the form of higher prices.  If enough transactions happened with bitcoin that would lower prices for all of us.  Merchants could also offer a cash/bitcoin discount.

Next is the fraud protection.  There are two kinds of fraud protection that the credit card companies provide.  The first kind is if if someone steals my credit card and starts making charges that I didn't authorize.  The credit card companies protect me from the fraudulent charges the thief makes.  Hopefully we all realize that if it weren't for the credit card in the first place this kind of fraud wouldn't be possible.  If Target, for example, hadn't had a bunch of their customers' credit card information stored on their servers when they got hacked their customers wouldn't have needed this kind of protection.

The thing to understand here is that paying with a credit card is a pull operation and paying with bitcoin is a push operation.  To pay a merchant I have to give them my credit card information and they pull money from my card.  Anyone that has that information can pull money from my card, making this kind of fraud easy.  With bitcoin a merchant gives me their bitcoin address and I send (push) bitcoin to them.  I don't have to fork over any sensitive information. The merchant is safe too, by the way, because that bitcoin address they gave me cannot be used to withdraw funds from them, only to give them funds.

Another cost of this kind of fraud protection that I discovered recently is that once it is discovered that someone has stolen your credit card information you lose the use of that credit card until a new one can be issued.  That can take up to two weeks.  Over this last Christmas both of my credit cards were cancelled due to fraud.  I was leaving on a trip soon and didn't have a credit card to take with me.  I really didn't like the idea of carrying a bunch of cash around on a trip instead of a credit card (fortunately I still had a debit card).

The other kind of fraud protection that credit cards provide is chargebacks.  If someone doesn't give me the thing I thought I had paid for with my credit card I can essentially take my money back by asking the credit card company for a chargeback.  You can't do this with bitcoin because it works like cash.  Once the bitcoin leaves your hand (so to speak) you don't have any control over it anymore.  The think to keep in mind again is, this fraud protection does not come for free.  Those fees I already talked about are partly there to pay for this[1].  Now think about it, how often are you required to resort to using a chargeback?  I personally have never needed to.  I have had disputes with merchants where I thought I might need to, but I have always been able to resolve the problem by talking things out with them.  Merchants are very motivated to protect their reputation and their relationship with their customers.  Since this is the case, why are we all paying for the ability to request a chargeback with every transaction?  Could I say to Visa when buying from someone I trust such as Amazon, "hey, I don't want to pay for chargeback protection for this purchase?"  The answer, unfortunately, is no.

I do understand that there will be times when buyers will want or even need chargeback protection.  You could simply use your credit card for times like that, but could we ever get chargeback fraud protection on a purchase made with bitcoin?  I don't personally know of an easy way to do that right now, but escrow services have existed since before credit cards were invented precisely for this reason.  Escrow services could easily work with bitcoin.

I titled this "Bitcoin vs. Credit Cards" but I'm not trying to say that There Can Only Be One.  More options is better.  I see a lot of people dismissing bitcoin out of hand due to misunderstandings about the costs and benefits of bitcoin when compared to other forms of moving money around.  Hopefully this helps clear things up for a few people.

Footnote:

1. Another costly thing about chargebacks is people can use them to defraud merchants. They can ask for a chargeback even when the merchant did deliver the thing they promised.  This is another cost the merchant bears when accepting credit cards.  A cost that, surprise, gets passed back to you as a customer.


Sunday, December 21, 2014

Adventure in Building a Home Gym

Last year I started assembling a little garage gym so I could do some weight lifting in the comfort of home. I scoured the online classifieds and found good deals on used power racks, a bench, and some weights. I was doing pretty well finding these deals, I thought. The one oddity was with the weights. The seller wouldn't sell them separate from this monster:



I had no use for this because I had bought into the idea that compound lifts were the best and this enormous and heavy thing served only to isolate your shoulders. I took it anyway because the guy was nice and he was giving me a good deal on the weights whether this was included or not. I was also thinking in the back of my head that I have a friend that likes metalworking that could probably help me turn it into something more useful. Nearly a full year later my hope was finally realized.

I stored it as is in my garage for about 6 months. My basement was being remodeled so there were lots of things in the garage with it (on it, under it, around it) so that wasn't a big deal. Once the basement was finished and we moved everything back in, it was painfully in the way. I moved it to the side of the house where it sat for a couple months, and then finally my next-door neighbor showed me how his acetylene torch worked and this thing was now in a much more compact state:


The flat bench I originally bought was a little rickety and cheap and my idea was to turn this into a nice sturdy one. A couple weeks ago my schedule and my metalworking friends schedules finally synced up and he graciously helped me use his chop saw, MIG welder, and drill press to reform the original beast into this:



It turns out welding is pretty dang fun. Once this metal frame was done I cut out and sanded some 1/2 inch plywood for the actual bench and attached it using bolts and t-nuts. My wife graciously painted the metal with some Rustoleum, found some padding and vinyl and helped me cover the wood. Here are the last few progress pictures:




Monday, October 6, 2014

More SystemVerilog Streaming Examples

In my previous post I promised I would write about more interesting cases of streaming using a slice_size and arrays of ints and bytes.  Well, I just posted another set of streaming examples to edaplayground.  I'm going to mostly let you look at that code and learn by example, but I will take some time in this post to explain what I think is the trickiest of the conversions.  When you go to edaplayground, choose the Modelsim simulator to run these.  Riviera-PRO is currently the only other SystemVerilog simulator choice on edaplayground, and it messes up on the tricky ones (more on that in a bit).

These examples demonstrate using the streaming operator to do these conversions:
  • unpacked array of bytes to int
  • queue of ints to queue of bytes
  • queue of bytes to queue of ints
  • int to queue of bytes
  • class to queue of bytes
Of all those examples, the queue of ints to queue of bytes and the queue of bytes to queue of ints are the tricky ones that I want to spend more time explaining.  They are both tricky for the same reason.  If you are like me, your first thought on how to convert a queue of ints to a queue of bytes is to just do this:

byte_queue_dest = {<< byte{int_queue_source}};

Before I explain why that might not be what you want, be sure you remember what "right" and "left" mean from my previous post.  The problem with the straightforward streaming concatenation above is it will start on the right of the int queue (int_queue_source[max_index], because it's unpacked), grab the right-most byte of that int (int_queue_source[max_index][7:0], because the int itself is packed), and put that byte on the left of byte_queue_dest (byte_queue_dest[0], because it is unpacked).  It will then grab the next byte from the right of the int_queue (int_queue_source[max_index][15:8]) and put it in the next position of byte_queue_dest (byte_queue_dest[1]), and so on.  The result is that you end up with the ints from the int queue reversed in the byte queue.  If that doesn't make sense, change the code in the example to the above streaming concatenation and just try it.

To preserve the byte ordering, you do this double streaming concatenation:

byte_queue_dest = {<< byte{ {<< int{int_queue_source}} }};

Let's step through this using the literal representation of the arrays so that rights and lefts will be obvious.  You start with an (unpacked, of course) queue of ints:

int_queue_source = {'h44332211, 'h88776655};

And just to be clear, that means int_queu_source[0][7:0] is 'h11 and we want that byte to end up as byte_queue_dest[0].  The inner stream takes 32-bits at a time from the right and puts them on the left of a temporary packed array of bits.  That ends up looking like this:

temp_bits = 'h8877665544332211;

Now the outer stream takes 8 bits at a time from the right of that and puts them on the left of a queue.  That gives you this in the end:

byte_queue_dest = {'h11, 'h22, 'h33, 'h44, 'h55, 'h66, 'h77, 'h88};

Which, if you wanted to preserve the logical byte ordering, is correct.  Going from bytes to ints, it turns out, is pretty much the same: reverse the queue of bytes and then stream an int at a time.

So what happens with Riviera-PRO?  If you try it in edaplayground you see that the resulting queue of bytes in the int-to-byte conversion ends up with a whole bunch of extra random bytes on the right (highest indexes of the queue).  8 extra, to be exact.  Same for the int queue result in the byte-to-int conversion.  I think Riviera-PROP must be streaming past the end of the temporary packed array (that I called temp_bits above) of pulling in bytes from off in the weeds.  Pretty crazy.  That's all done behind the scenes so I don't really know, but that's sure what it looks like.  Hopefully they can fix that soon.

Well, I hope I've helped clear up how to use the streaming operators for someone.  If I haven't, I have at least helped myself understand them better.  Ask any questions you have in the comments.

Friday, October 3, 2014

SystemVerilog Streaming Operator: Knowing Right from Left

SystemVerilog has this cool feature that is very handy for converting one type of collection of bits into another type of collection of bits.  It's the streaming operator.  Or the streaming concatenation operator.  Or maybe it's the concatenation of streaming expressions (it's also called pack/unpack parenthetically).  Whatever you want to call it, it's nice.  If you have an array of bytes and you want to turn it into an int, or an array of ints that you want to turn into an array of bytes, or if you have a class instance that you want to turn into a stream of bits, then streaming is amazing.  What used to require a mess of nested for-loops can now be done with a concise single line of code.

As nice as it is, getting the hang of the streaming operator is tough.  The SystemVerilog 1800-2012 LRM isn't totally clear (at least to me) on the details of how they work.  The statement from the LRM that really got me was this, "The stream_operator << or >> determines the order in which blocks of data are streamed: >> causes blocks of data to be streamed in left-to-right order, while << causes blocks of data to be streamed in right-to-left order."  You might have some intuitive idea about which end of a stream of bits is on the "right" and which is on the "left" but, I sure didn't.  After looking at the examples of streaming on page 240 of the LRM I thought I had it, and then none of my attempts to write streaming concatenations worked like I thought they should.  Here's why: "right" and "left" are different depending on whether your stream of bits is a packed array or an unpacked array.

As far as I can tell, "right" and "left" are in reference to the literal SystemVerilog code representations of an arrays of bits.  A literal packed array is generally written like this:

bit [7:0] packed_array = 8'b0011_0101;

And packed_array[0] is on the right (that 1 right before the semicolon).  A literal unpacked array is written like this:

bit unpacked_array[] = '{1'b1, 1'b0, 1'b1, 1'b0};

unpacked_array[0] is on the left (the first value after the left curly brace).  I don't know about you, but I'm generally more concerned with actual bit positions, not what is to the right and left in a textual representation of an array, but there you have it.

Once I got that down, I still had problems.  It turns out the results of streaming concatenations will be different depending on the variable you are storing them in.  It's really the same right/left definitions coming into play.  If you are streaming using the right-to-left (<<) operator, the right-most bit of the source will end up in the left-most bit of the destination.  If your destination is a packed array then, just as I explained above, "right" means bit zero and left means the highest index bit.  If, your destination is an unpacked array, your right-most source bit will end up as bit zero of the unpacked array (which is the "right" bit according to the literal representation).

Got all that?  If not, I put a code example on edaplayground that you can run and examine the output of.  The examples are all streaming bits one at a time.  It gets a little harder to wrap your head around what happens when you stream using a slice_size and when your source and/or destinations array is an unpacked array of bytes or ints. I'll write another post explaining some tricks for those next (UPDATE: next post is here).