Yea! I finally found the tutorial on how to use Xilinx's ISE program to turn VHDL into bitfiles that can run on the avnet Spartan 3A Evaluation Kit. I followed the included tutorial that sets up a counter and sets the LEDs to display the top bits of the count, except I made some simple changes to make sure I understood everything. The only material change I made was to use a clock multiplier to increase the frequency from the 16 MHz oscillator to a much nicer 100 MHz. I'm so used to python and C++ that it took me a while to get used to the pascal-like syntax of VHDL, but aside from that (and the fact that the ISE software is not netbook friendly - way too much HD access and the UI doesn't fit on the 600-pixel-high screen), the process and results were pretty cool. Ok, there is one more annoyance - that I can't find any drafts (or other legally free versions) of the 2008 VHDL standard. I like having the standard for a language I'm working in, but I don't quite like VHDL enough yet to spend nearly $300 as IEEE, ANSI, et al are asking for the pdf.
I've forgotten pretty much everything I knew about digital logic design, but I located my college text on the subject (one of the very few good books from my classes) and have been reading through it. I also picked up an interesting-looking book on processor design in a brick-and-mortar store, and then found out it was one of the best-rated books on the subject on amazon, which is nice to know. I need to start reading it sometime soon.
I think I've decided on my first big project to do on the FPGA in order to really familiarize myself with the hardware design and development process - Core War! I've already done a lot of software work with the game, so I'm pretty familiar with the instruction set and mechanics. I'm thinking I'll make a CPU that runs a game of Core Wars using the 'nanohill' rules - mainly because the board I'm using doesn't have SRAM so I need to make the whole game state fit into the BRAM in the FPGA. While almost any rule set would fit into the BRAM once, I'd really like to make a large array of Core Wars CPUs running simultaneously. Once I get that, I can load some known warriors from nano hills into the flash ram, and compete my creations against a full match (with all 142 starting positions against every opponent), which gives me a good way to benchmark a given warrior. Finally, with that capability, I should be able to make an evolver that can (by rough calculations) go through as many as a billion evaluations a day. If I'm lucky, I'll rule the nano hills a few days after I finish the project =-)
 It is the least expensive FPGA board I've ever seen. It's missing the fun things the more expensive kits have, like ports for VGA and audio, but for a mere $50, it's a great way to see if you're interested enough to spend more money on such hardware (and you can always use the pins it exposes to connect your own ports). It also comes with a a programmable-system-on-a-chip and a programmer for it, which is a nice bonus (especially considering that the board just for that PSoC is almost $200).
Lately, work has been satisfying my software development drive, so I've started getting back into another hobby - electronics. I dug out an old 'lots of projects in 1' kit from my closet and sent out a few orders to various places (primarily mouser, amazon, and ebay) to start creating a full kit to work from.
After receiving the first few things I ordered, I started playing around and wired up a 'white noise' generator based on avalanche noise. I built the circuit 5 times, but it never seemed to work for some reason when I made it on the kit's breadboard. I finally got out my separate breadboard and power supply, and lo and behold, it worked wonderfully (apparently when a schematic says you need a voltage source of a certain strength, it might say that because fewer volts will not work properly =-).
Yesterday my oscilloscope arrived, and it was rather awesome to example the white noise wave on it. Since my oscilloscope isn't officially calibrated, I decided I'd like to generate some signals from my sound card and use the oscilloscope's measurement capabilities to check the frequency, and all seems well so far (at least up to 8khz, which isn't far for a 250MHz oscilloscope). Now I'm working on a signal generator program in python that will also give me something awesome to look at on my oscilloscope. After that, maybe I'll start working on some kind of game on the oscilloscope. Analog computers are neat =-)
">oscilloscopes can be vector displays
Now that this is a new month (and my "fun money" budget reset), I'm considering buying an FPGA board. I can't decide whether I cant to spend a bit more and get a fully functional board (from diligent) that has RAM etc already built in, or whether I'd like to get a bare bones board (specifically, the $50 "Spartan 3A evaluation board" board from avnet that has a few buttons, a bit of flash memory, and lots of headers). Since I have no soldering experience and SRAM doesn't seem to offer many DIP options, I'll probably go with the more expensive board, but, then again, learning to solder would probably help quite a bit with some of the projects I want to do eventually - breadboards have tons of parasitic problems with high-frequency circuits, and with the FPGA, I have several video projects I'll get around to eventually.
Life is good, but busy. The new job is working out great. Working on interesting, challenging problems is much better than tedium.
On June 8, 2008, I met an amazing woman and spent over 4 hours enjoying conversation with her over lunch. On June 18, 2009, I presented her with a ring, and she accepted it. On September 6, 2009, she becomes my wife.
Also, I just recently discovered an awesome programming game titled Core Wars. It is loads of fun, yet frustrating. I'm building my own MARS and an evolver based on it, but MSVC bugs are getting in my way. At least, I believe it is an MSVC bug - g++ and comeau both compile the code just fine in strict standards modes.
Near the end of last year, I found out that my employer was closing the branch I was working in, and that I could continue on until the office was actually closed in December. After we (the employees of the closing branch) found out, we made an argument for a later date that coincided with a contract deadline, and it was accepted by management. The office still had to close because the lease was up, but they allowed us to take what equipment we required and to continue working from home. I've been doing that since mid December, but the extended deadline is fast approaching.
For that reason, I've been looking on various job listing sites for a while now, and found many jobs I could do but only a few that sounded interesting. I'm more interested in low- and intermediate-level programming, while most jobs were for developing database front-ends and other simple business GUIs in Java or .Net languages. Of the few jobs that sounded interesting, several required high level security clearances that I don't have. Just recently, I found a listing that combined the best of all worlds - a software reverse engineering position that didn't require security clearance. After reading the posting, this posting instantly rose to the top of my list as an interesting, challenging position.
I submitted my resume to this new-found first choice on Feb 2, heard back from them on Feb 4 with a skill assessment test, and returned the test on Feb 6. The next week, I received the response that they would like to fly me half-way across the country to DC for an interview on Feb 19. I was impressed with their interview technique and blown away by the description of the working environment in the office I'd be joining. I also felt I did well and made a good impression on the many people I spoke to. At the end of the interview, I was told I'd hear back within a week one way or another. Yesterday, Feb 25, I got an offer letter that includes a >50% raise from my current position, a week more vacation, and several other improved benefits. Today, I signed the forms and gave my current employer notice of my resignation in 2 weeks.
A while back, there was a big post in the lounge about the fun game Fantastic Contraption. I liked it enough to register and created a rather challenging level titled, creatively, Extrarius 1. The next thing I did was to solve my own level to be sure it could be done. My solution is fairly simple, but I want to see some other solutions (if there are any), and that is where you come in. You have to register to access custom levels, so try out the free levels and find out you like it enough to help me out =-)
Note: I'm not affiliated with the game, it's authors, etc, so the only benefit I get out of you registering is (hopefully) some more solutions for my map =-)
Postscript: If you insist on seeing my design, you can message me here on GDNet. I don't want to publish it because people would then publish copies of the same solution, and that is boring. I might publish it eventually, but not today.
I realized when making this post that I'd hadn't announced that I graduated here. On Dec 14, 2007, I got my B.Sc. in Computer Science. Well, actually, I don't have the degree yet - an ID was required before they would hand over the degree after the ceremony, and I forgot to bring one, but I did graduate =-)
I'm not really sure why I feel it neccessary to announce, but I do, so I did, and here it is.
If you've kept up with various threads over the years, you might have noticed that I repeatedly mention a boycott of games, movies, music, etc for various reasons and that the only way I get such things is as gifts.
Well, I finally found an artist awesome enough to buy - the amazing "Weird Al" Yankovic. I noticed some of his videos on YouTube, then I noticed he was coming here on his tour and I liked his stuff so I went to see him. Next day I had to get his latest album and his video collection. His are the only songs I've listened to repeatedly as far back as I can remeber. All the other music I have is classical, and I'm really tired of songs about cliche subjects, but his lyrics are just so amazing that I couldn't help myself.
I am starting a Weird Al collection, but I'm still not buying anything from anybody else.
Hello, World. I'm Extrarius, and I'm a domain reserver. That's right - I get a "good idea" for a domain name and then register it, even though I know I'll never have time to work on the project (if I even have a project in mind).
For example, I recently made a joke to a friend about making reviews for movies/games/etc using RPG mechanics to explain the story. That led to me registering RollReviews.com for the purpose of creating such a review site. Then the friend remarked that everybody 'not in the know' would hear the name as "RoleReviews", so I registered RoleReviews.com too and made it redirect to the former site.
Another example came to me after I made a post in a forum (this one?) pointing out that an argument was about "BadWrongFun" - thus, BadWrongFun.com. I have no idea what project would fit with this domain, but there has to be something appropriate.
Oh, and of all these domains, not a single one uses any space or bandwidth on my web hosting. I'd really feel like my money was wasted if not for a few other sites I host. ClanDV.com doesn't use any more resources than my own domains, but defmacro.org (run by CoffeeMug) actually uses a tiny bit of the resources I'm alloted so my hosting gets a bit of use =-)
I was just going through some old code when I found a 'TrigTest' program I made to benchmark various trig approximations for a thread here a while back. Since I originally compiled it using VS 2003, I now have VS 2005, and I've heard many complaints about VS 2005's optimizer, I decided to recompile it with the same settings and see how much changed.
Before recompiling it, I ran the existing executable that was made with VS 2003, and the timing portion of the program produced the following results:
While the code does use templates a bit, it doesn't use anything besides simple calculation operations and a few functions from the Windows API for timing (so any new 'security features' should not affect the timing code). Thus, I was quite surprised to find that the results for the VS2005-compiled version were so vastly different:
The built-in 'sinf' function almost doubled in execution time, but 'WildMagicApprox0' is over 18 times slower!
If anybody is interested, you can find everything related (source, projects, executables, and assembly output for both vs 2003 and vs 2005) using this link. If there is a way to get the VS 2005 version as fast as the VS 2003 version (without modifying the vs 2003 project to make it worse =-), I'd love to know about it. The settings seem to have converted correctly as far as I could tell, but I avoid vs2005 for the most part so there could be something simple that I'm missing.
After buying the Asomiv collection "I, Robot" twice, I decided it'd probably be a good idea to catalogue all the books (and the short stories in those books, where applicable) that I own. I had a few spare minutes this morning, so I went over one shelf and have recorded the title and ISBN (or, for those that lack an ISBN, misc other info that might help you find the book) of each book. They are listed below in order of their position on the shelf from left to right.
Brock's Biology of Microorganisms (0-13-520875-0) Semantics: A Reader (0-19-513698-5) "You And The Law"; 1971; Readers' Digest Abraham Licoln (1-56619-360-5) New York Times Book of Science (0-8129-1880-0) A Wider View of the Universe: HENRY THOREAU'S STUDY OF NATURE (0-252-06620-0) US Army Survival Manual (1-56619-022-3) Strategy of Process Engineering (0-471-74455-7) Engineering Graphics (0-02-342620-9) Software Engineering: A Practitioner's Approach (0-07-365578-3) Tab Electronics Guide to Understanding Electricity and Electronics (0-07-136057-3) The Art of Electronics (0-521-37095-7) "Basic Electronics"; Copyright 1994; Radio Shack / Master Publishing, Inc.; Evans & McWorter Grob Basic Electronics (0-07-024928-8) Microsoft Windows Internals (0-7356-1917-4)
There are actually a few more books on the shelf, but I wouldn't have them unless somebody else put them there, so I excluded them because they don't really tell you anything about me (and aren't "my books" so much as "books in my shelf space").
Only 10 more shelves and 3 small floor piles to go (plus 1 additional floor pile per visit to Half-Price books between now and whenever I finish) !
C++ templates are fun. I hadn't gotten "warning C4503: decorated name length exceeded, name was truncated" in MSVS since I stopped using 6 years ago, but just now I had a project gain 85 such warnings after adding some classes like the following:
Is there a better way to make such comparison functors? I looked around in MSDN and Boost for such wrapper classes, but I couldn't find them so I created the above, another that derefs it's arguments and compares the result (for iterators), and a third that does a lexicographical_compare on two provided containers. Of course, in some cases, they are chained, which causes the problem:
typedef TIteratorCompare TSymbolRefCompare; typedef TContainerCompare TSymbolRefListCompare; typedef TMemberCompare TExpressionCompare; typedef TIteratorCompare TExpressionRefCompare; typedef TContainerCompare TExpressionRefTableCompare; //TExpressionRefTableCompare is used in a custom comparison functor for TAssertionCompare to make TAssertionTable
Last night I received and set up my new machine, and even though I don't have much installed, it already seems MUCH faster. Previously, I had an Athlon 2400+ (2.0 GHz) with 640MB of PC2100 RAM. Now I'm running an Athlon 64 X2 3800+ (2.0 GHz) with 2.0GB (Dual Channel 2x 1GB) of DDR400 RAM and the machine boots in under 5 seconds. Yes, I know that a fresh install of the OS might help, but there isn't anything that autostarts on my old machine that isn't on my new machine also (basically just BitDefender).
It's amazing how much faster things can get when all the buses widen and the cpu gains a 'few' transistors =-)
If you ever decide it'd be a good idea to implement your own memory heap system for whatever reason, make sure you do it correctly, which includes combining multiple small neighboring blocks into a single larger block at least occasionally if you allow blocks to be split. Also, while a linked list is a convenient way to store a 'free list', it's going to end up quite slow if there is a single unordered list that you search for every call to 'MyAlloc' and 'MyFree' (especially if you don't combine blocks but do split them).
Yay! I finally get to sleep after 48 hours of last-minute-start-and-finish-a-major-school-project. Over the last 48 hours (during which I had no sleep), I wrote 65KB (counting end file size, which excludes the initial work I did since I started over) of barely commented C++ code that implements the essentials of an intel 8086 assembler to the point that it can generate a correct(as far as I'm aware - I can't look closely as tired as I am) listing file of a test assembly program.
I can see why the average programmer would slow down to only write an average of 3 lines a day. 65Kb of code is too much too manage when you wrote it while half asleep, and I'm still not quite sure how I got it working. I just hope the professor appreciates the work I put into - maybe not much time, but a lot of just about everything else there is to give.
Normally this teacher asks for code in hard copy, but I'm fairly certain she'll make an exception just this once =-)
This concludes my second 'crunch time' session, with the first being for work (see a previous journal entry of mine) and this one lasting much longer (last time I actually slept before I finished!) and produced MUCH more code (byte-size-wise).
Desipite the fact that school computers are underpowered (128MB of RAM isn't enough for a near-default WinXP Pro install) and filled with crap (despite all kinds of 'protective measures' that are instantly nullified by the fact that the 'guest' account is a local admin on every machine), they usually run decently. Sometimes, however, they come with interesting errors. Windows is great though - a reboot fixed the problem right up (no, they don't boot from any kind of network image) and the computer worked fine afterwards.
I'm just glad I have a dictatorship on my machine.
I just had an awesome idea for a website: InaneRavings.Com Note that it's all just rhetoric meant to gather a 'community' (so much as such a site could)
It will take a lot of work, but I think it will be quite popular once I get it up and running.
I'm wondering if the page looks the same on other browsers, but I'd rather not install them myself because I hate installing programs I don't use (with all the junk they put in the registry etc) so here you can see an image of what it looks like to me in IE6. If it looks different for you in your browser, would you please list browser, version, etc and either describe the difference or post a screencap of what it looks like?
Checking my web logs, I noticed somebody had not only 'stole' some of my art (below), but they had forums linking straight to my site so they were leeching bandwidth too =-( It isn't a lot, but it was kind of a slap in the face to see that kind of thing going on. It's nice to see that the person liked what little I've done, but they could have at least asked if it was ok to use. If you go to the www subdomain (or leave off a subdomain), you are taken to a forum where you could easily ask such things.
The image they were using is
so I had to write a mod_rewrite rule to change it into:
for the domains the person was using it on.
I wouldn't really have minded if it wasn't the one work of art I use to identify myself (that figure is in every avatar I've made since I created it), but it's kind of like somebody tried to steal my identity =-/
Well, today I (hopefully) finished my first ever crunch time. Lucky for me, it only lasted a day and that only meant putting in 15 hours in 24. Unfortunately for me, I'm working so few hours(part time) that those extra 11 won't put me up to overtime pay.
Basically, I found out thursday that some of my code (that had to be designed long before all the requirements were known and thus wasn't well engineered code) needed major modifications to work with a new file format and it needed to work with that format by today (or I'd have to come in saturday to manually do it's work). I had to entirely rewrite the subsystem because hacking it to work wasn't really possible. It's now properly engineered, but needs several more layers to be as functional as the previous library I had written. Luckily, the program that was needed only required the base level of functionality from the library.
It was an interesting experience to be coding at 4AM and then find a complete lack of caffiene in the house =-/ Oh well, I worked much better after some sleep anyways. It cost me was a few points in various classes that I missed because of the sleep, but I think it was worth it to finish it 'early' so others(users of the program) might be able to avoid overtime if they want and I really didn't want to work on saturday (might still have to, but hopefully the program is bug free enough for now).
I've really been wanting to get into VHDL, despite the poor quality of Xilinx's software mentioned several pages back, so I decided to try the latest version of the free Xilinx software to see if they fixed the many problems I noticed. Surely they would have, since it would be a big selling point to get students into their stuff which would subsequently move businessess toward it also, right? Apparently not. I downloaded the latest version, and the ~5MB installer took around 6 hours to install the program. The whole time it was using practically 0% CPU, so I'm not sure what it was doing but it was progressing very slowly. After that, I went to download the tuturials and noticed there was one specifically designed for the latest free version of the software. Awesome! It must have fixed all the problems where the tutorial didn't cover steps required and did cover impossible steps. So I thought. I load up the tutorial and start following the directions, only to find out after a few steps that I couldn't proceed because an important option on a dialog isn't actually in the program. I look around in an attempt to be able to continue, but I'm unable to find the option anywhere so I can't do the tutorial. Not only that, but really the tutorial starts with a 90% finished project so I'm not sure how much it would have helped to be able to complete it since I'm not sure I could redo the work already done in the project.
Oh well. Xilinx 0 - Extrarius 0 VHDL will have to wait another while at least I would go with Altera, because from what I hear they actually have superior parts, but I haven't seen a board with similar functionality to the Spartan 3 Board for a similar price and I'm not willing to pay anything more because it'd just be a hobby.
After all that work for nothing, I checked my email and I noticed 1&1 finally canceled my account with them after I emailed them something like 30 times each getting the same response of "Mail or Fax us the form" even though the idea was I didn't want to spend any more money on those domains. I'm pretty sure the catalyst was that the only CC# they had expires before the next billing cycle. If they had just canceled them when I asked the first time, they would still have my business (even though I hated their control panel because it was so slow) but because of their hostility I transferred all my domains to the much more expensive but much more friendly GoDaddy.
Don't you just hate it when your professor doesn't know what he/she is talking about? It seems that some of my professors haven't touched a computer as a 'power user' in 20+ years, which makes them give all kind of bad advice. =-(
Semi-related, I finally got a decent assignment during my 'college career' - I have to make (over the course of the semester) a full x86 macro assembler and linker. Luckily, I just have to make it extendable and not actually create a full table of every instruction.
On an unrelated note, gas prices jumped $0.60 today from $2.39/gallon to $2.99/gallon for the least expensive grade at the least expensive station around here due to the hurricane that recently hit some major oil rigs and refineries. =-(
I look on the dell site to check on some stuff in my account, and I notice there is a new special - get a free Windows Mobile 5.0 CD with the purchase of a new Dell Axim X50v. I thought I might as well try to get it since I JUST bought mine, so I go to online support and talk to them, and it should suffice to say they didn't like that idea. After going through 3 support personel, I realized I could just return my Axim (21 day return period) and get a new one under the exact same deal (except for the bonus CD). I tried using that to leverage some people, but no dice on that either.
Now, don't misunderstand me and think I'm just giving them hell - I was willing to buy the CD (the promotion says $39 value), but a short talk with a sales representative quickly revealed that the CD was only for the promotion and was not for sale seperately.
After a few more people, I finally got a number with extension that I'll try when I get home, but I'd just like to say that I'm rather disappointed that it looks like I'm actually going to have to send it back and get a new one to get the windows upgrade for my Dell Axim X50v. It seems like it would have to be costing them more money for all the support I've gone though, to process the return and new order etc than it would for them to just give me the '$39 value' promotion.
Update: The number and extension took me to the same place I'd already been (but last time it was with an entirely different number and no extension).
Update #2: After talking to 9 people (no exaggeration) without success, I finally decided to go ahead and get the return process started so I could get the free CD. That person (the 10th at Dell I've talked to) immediately told me that I didn't need to return it and that he would update my account and email me the info for the free CDs. If only more people over there had that kind of sense =-|