HNHacker News
TopNewBestAskShowJobs

tcas

445 karma · joined November 22, 2011

submissionscomments
tcas··on Amazon Pairs Print and Digital Books With New Program
I hope this applies to technical books as well. I would love to have things like my old DSP textbook and others available with me wherever I go, but still have the hardcopy I'm used to at home.
tcas··on The one HR benefit every startup should add
Just double checked. It's own occupation for 24 months, then it becomes reasonable occupation. I have a feeling that that term is something that when 24 months approach, you want a good lawyer for.

Still, considering I'm paying less than a coffee per month for LTD, and personally knowing someone who was once working as a top tier developer and got into a serious biking accident that left him with brain injuries, I'm going to buy into it. Next year I'm going to push HR about that 24 month clause.

tcas··on The one HR benefit every startup should add
My LTD is 60% of my income, tax free (very important, always pay for any disability post-tax, if you don't your income from it is taxable), until I am able to make over 80% of my pre disability income.

It provides me benefits from 6 months after my disability (I decided not to pay for short term which starts at 2 weeks until 6), and will keep going until I am ~65 years old.

tcas··on On Apple’s A7 Processor
Reading over the article, it doesn't really say anything of technical content other than "Apple has complete control and can vertically integrate, applications have different needs per thread, if you can tightly couple hardware and software, it'll be fast." Computer architects have known that for decades by now.

Apple isn't custom designing the IPCores for the processor or graphics. I'm willing to bet that almost all of this chip is off the shelf stuff. What they can do is choose parameters like number of cores, cache sizes, CPU options (like you said, stuff that is available with existing suppliers) and also include additional units that they can license separately (e.g. earSmart)

The blog post mentions stuff like Frameworks and Apps. Unless I am very out of touch with SoC offerings these days, you don't get a MapKit.framework accelerator. You get things like H264 encoding/decoding and AES, or DSP functions which these frameworks can use. However, the faster performance you want with one function, the more locked in you get. An example of this is comparing encryption speeds on an Intel processor with AES-NI. AES is an order of magnitude faster than just about everything else. Everything else however, uses the same vector extension set that were designed specifically to speed up sub operations of common algorithms.

The other advantage they get by doing an in house chip is that they can choose the on board peripherals they want (possibly remove some chips that would normally be external), and the exact number of external busses they need for all the external peripherals, which can lead to a smaller, cheaper, more power efficient chip.

tcas··on JavaScript right on the hardware
Also: peripheral muxing, interrupts in input mode, connecting a timer as PWM output to a pin (which is basically what blink is probably doing). In input mode it might also have schmitt triggering.

Where's the blink getting the timer from? What happens if I update the main PLLs? I'll refrain from making assumptions however until I can see their driver libraries. I'm hopeful, but having played around with the 1830 it's not a simple chip.

tcas··on Google, IBM, Mellanox, Nvidia, Tyan form OpenPower Consortium
The RTL source was released under the GPLv2, GPL and hardware is a bit iffy since synthesis isn't the same as compilation, but depending on who owns the IP (in this case Oracle) they could make the case you need to open source your RTL source as well. I personally wouldn't invest the high hundreds of thousands to millions of dollars needed to produce a chip without getting a license agreement first with the ambiguity of the license, and which point I would probably consider the POWER architectures (like many FPGAs vendors used to) or ARM. In the case of pushing it to FPGAs, the OpenSPARC would use a ton of resources to the point where I'd rather have a dedicated off chip processor, or a FPGA with a built in processor.

I know people have put the OpenSPARC code on FPGAs, but I don't know of any use in industry in FPGAs or ASICS. I'd be interested if anybody knows any.

tcas··on Feds put heat on Web firms for master encryption keys
YubiHSM[1] seems much more reasonable at $500, but it's still not "cheap."

A COTS implementation would be interesting. Securing SSL keys would require very quick public key crypto and probably a PCI-E interface to talk to the server, the PCI-E interface making it much more expensive (FPGAs with PCI-E hard blocks have gone down in price, but are by no means cheap, not to mention the increased cost of the PCB). Prototypes probably wouldn't need VC for a proof of concept, but after that you typically need to put up a bit of money for any manufacturing.

I'd be interested in a way to securely generate and store a Root CA certificate and sign other keys with it. Ideally using something like secret sharing so that no single person can access the Root CA as well without having n out of m people.

[1] http://www.yubico.com/products/yubihsm/

tcas··on Feds put heat on Web firms for master encryption keys
Any info on using a HSM for front end SSL? I know of crypto accelerators, but nothing like using a Luna HSM to store the keys.
tcas··on Apple Developer Website Update
That's what I thought -- but it was also down the 19th yet 2 were revoked.

Probably means nothing however. I doubt that anybody with the ability to get into the system would want to get only developer certificates.

tcas··on Apple Developer Website Update
I downloaded the CRL for developer certificates [1] and quickly looked at it using grep:

  grep -E "Revocation Date: Jul 17 .{8} 2013" wwdrccrl.txt | wc -l
      3065
  grep -E "Revocation Date: Jul 18 .{8} 2013" wwdrccrl.txt | wc -l
      2289
  grep -E "Revocation Date: Jul 19 .{8} 2013" wwdrccrl.txt | wc -l
         2
  grep -E "Revocation Date: Jul 20 .{8} 2013" wwdrccrl.txt | wc -l
         0
  grep -E "Revocation Date: Jul 21 .{8} 2013" wwdrccrl.txt | wc -l
         0
These are the two certificates that were revoked on the 19th

  grep -A 3 -B 1 -E "Revocation Date: Jul 19 .{8} 2013" wwdrccrl.txt
      Serial Number: 2628C7F90970D227
          Revocation Date: Jul 19 03:14:04 2013 GMT
          CRL entry extensions:
              X509v3 CRL Reason Code: 
                  Key Compromise
  --
      Serial Number: 1A51ABFA4844BD45
          Revocation Date: Jul 19 03:24:03 2013 GMT
          CRL entry extensions:
              X509v3 CRL Reason Code: 
                  Key Compromise
To generate the wwdrccrl.txt file I used:

  openssl crl -inform DER -text -noout -in wwdrca.crl > wwdrccrl.txt
Just to be clear -- every entry there I see lists the reason as Key Compromise, just interesting that they usually seem to revoke at least 2000 certificates a day but suddenly stopped on the 19th with just revoking 2.

[1]http://www.apple.com/certificateauthority/

tcas··on Apple Developer Website Update
Which wouldn't take very long. I'd be amazed if Apple wasn't using a hardware security module to store their intermediate signing keys and logging the keys generated from it.

I wonder where the CRL for the dev certificates is -- we might see an update to it soon.

tcas··on Dennis Ritchie's first C compiler on Github
Doh, I completely glossed over the readme and went straight to the code. That makes sense -- Thanks!

Cool to think that that waste function can still compile with todays compilers. A quick disassembly it seems to take up 751 bytes compiled on x64 using clang on O0.

tcas··on Dennis Ritchie's first C compiler on Github
Looks like a very early dialect. C assumes everything is an int unless specified otherwise. You can declare parameter types after the function name. So:

  init(s, t)
  char s[]; {
would be equivalent to:

  int init(char s[], int t) {
This still works with modern compilers.

I'd be interested if anyone has any more info about this:

  waste()		/* waste space */
  {
  	waste(waste(waste),waste(waste),waste(waste));
  	waste(waste(waste),waste(waste),waste(waste));
  	waste(waste(waste),waste(waste),waste(waste));
  	waste(waste(waste),waste(waste),waste(waste));
  	waste(waste(waste),waste(waste),waste(waste));
  	waste(waste(waste),waste(waste),waste(waste));
  	waste(waste(waste),waste(waste),waste(waste));
  	waste(waste(waste),waste(waste),waste(waste));
  }
Found in last1120c/c10.c
tcas··on \d less efficient than [0-9]
Also, when using Python 3.2 it seems to be the default behavior

  Python 3.2.3 (default, Oct 19 2012, 20:10:41) 
  [GCC 4.6.3] on linux2
  Type "help", "copyright", "credits" or "license" for more information.
  >>> import re
  >>> re.match(r'\d', '੧')
  <_sre.SRE_Match object at 0x7f188f6d4850>
tcas··on Homemade GPS Receiver
Speed of the FPGA? FPGAs are given speed grades, but the actual design dictates the max speed. Different logic sections, depending on pipelining, size, which pins they are connected to can run at different speeds. You might have an overall max speed/critical path of say 50mhz, but you might be able to safely run a small section of the design at 200mhz.
tcas··on Four companies will produce the microchips on which the global economy depends
This is in regards to the highest end fab technologies, which doesn't even include IBM who apparently have a 22nm fab. http://en.wikipedia.org/wiki/List_of_semiconductor_fabricati...

With the most advanced fab technologies there's a much smaller number of companies that can both afford the initial capital investment (these plants are _insanely_ expensive) and the engineering expertise / tools to even design at these processes. Meanwhile, the $3 microcontroller I buy that runs as 12mhz fits every requirement I have while being manufactured at a cheaper, but much larger process. The high end processors and FPGAs however, benefit much more from smaller feature sizes (and those FPGAs cost accordingly).

As far as I know, when these 4 companies make advancements, the entire industry benefits. For example, Nikon comes out with a more advanced stepper for 14nm processes, Intel is the one paying the premium for them, and later the rest of the industry can benefit from a more reliable 65/45/32nm process after the cutting edge companies have moved on. I believe this is true for the EDA tool space as well.

The last paragraph is a bit of speculation on my part. I'm not in the semiconductor field but I know a bit about it, and logically it makes sense for tools / equipment. It does not include things like internal knowledge about manufacturing reliably on a certain process however.

tcas··on 'We thought we'd sell 1,000': The inside story of the Raspberry Pi
The power distribution on a board is essential to it running properly. The 5v needs to go into 3.3v and probably 1.2-1.8v as well. You need to make sure that any loads on the power net won't cause a momentary voltage drop below any component's tolerance. You also want to make sure that the components on the far end of the board will receive the same voltage that everything else is receiving. Next you want to make sure that any digital chips won't inject noise on the line and have their own power reserve via decoupling caps. These capacitors need to be speced out to block specific frequencies and placed accordingly. Also, grounding is super important to consider as you want to have a stable "0V" reference (for digital only designs this means having a dedicated ground plane, mixed analog and digital can have separate ground planes).

Now, for most hobbyist designs the rule of thumb is use a simple LDO and the recommended input/output caps (1-10uF) for it and use wide traces to route the power in a star like fashion. Also put a 100nF capacitor next to each power pin for noise and to serve as a power reserve.

More complicated circuits / circuits that will need to go through FCC will require more complicated power distributions. Most professional boards (with high frequency signals on it) that will meet FCC will be 4+ layers, with a dedicated ground and power planes. A PCB I just designed with an FPGA on it had something like 50 capacitors for just the 1 chip, with physical size requirements for each type (the smallest capacitors were 1mm x 0.5mm to block the highest frequency noise, anything bigger won't be as effective).

Don't let this deter you however, looking back on my first schematic/PCB designs I'm amazed at how much better I've gotten. Microcontroller designs with a few chips are very, very lenient. If you mess something up it will likely still work, but your performance won't be as good (i.e. if you use a built in ADC it may not be as accurate, or you might not be able to reliable run it at it's max clock speed)

tcas··on Your Very First Microprocessor
You're right. I just tested it in Xilinx ISE 13.4 and it seemed to work without using any assign as both a single wire and a bus. I just recall seeing an error message at some point in the past which was about something similar.
tcas··on Your Very First Microprocessor
Didn't expect to see Verilog on the front page of HN.

1 nit pick, I think wherever you use a wire assignment you need the assign keyword. e.g

  wire [1:0] command = mem_out[7:6];
  wire [2:0] destination_addr = mem_out[6:3];
  wire [2:0] source_addr = mem_out[2:0];
should be

  assign wire [1:0] command = mem_out[7:6];
  assign wire [2:0] destination_addr = mem_out[6:3];
  assign wire [2:0] source_addr = mem_out[2:0];
Or maybe it's wire assign. I usually first declare the wire in a single statement then use assign on a new line.
tcas··on $.url() - A simple, lightweight url parser for jQuery
Looking at the source, it looks like there's only one line that calls jQuery on line 37 (and I don't think it should be using the short calling convention without a wrapper -- if you rename jQuery to '_' or something else it won't work):

    else if($.isNumeric(arg)
There's no native JavaScript equivalent to it if I recall correctly, but there's a bunch of alternatives to it as well.

It should be easy enough to extract out the logic and repackage it. Maybe I'll fork it tomorrow and do the changes myself.

tcas··on Red Bull Stratos Skydive Rescheduled for today
It's definitely not full multicast. Flash doesn't support multicast (if I recall correctly).

They're probably doing a single stream to CDNs which then fan them out.

tcas··on ShowHN: GetURL - A CLI tool to get a public link for any file
I think the term you're looking for is UDP hole punching, port knocking is performing a special sequence of connections (i.e. try TCP on 8100 then UDP on 4000 then TCP on 2000) to open up an additional port (like SSH) to a certain IP.

I'd be curious if there exists such a company or even any good open source libraries that can tie into other servers.

tcas··on Apple A6 Teardown
Does anyone know what the two horseshoe looking things are in the bottom left of the RTR8600 die? I want to say inductors, but they aren't in the shapes I'm familiar with (that being said, my VLSI knowledge is extremely rusty)
tcas··on Mars to Earth at an average of 29 kbits/s
Encoding != Encryption. Encoding is on a line level, so for example, 8b/10b is used in HDMI to prevent a DC bias on the line. The way it works is by making sure within a certain number of 8 bit inputs (I think 3), the resulting 10b outputs will have an equal number of 1s and 0s, therefor preventing the physical wire from charging itself closer to V+ or V-, which in turn helps prevent bit errors.

8b/10b has exactly 25% overhead, this is the difference between baud per second (what the line is capable of transmitting, which is on the 10b side), and bits per second (which is the decoded 8b data). If you have a 10kilobaud/second line, you'll have an effective data rate of 8kbit/second.

Encryption is done before the encoding stage, the encoding scheme doesn't care and isn't affected by the encryption.

tcas··on Emacs packages for programmers
It's pretty easy using package. Here's what I just copied and pasted from my init.el

  (require 'package)
  (dolist (source '(("technomancy" .   "http://repo.technomancy.us/emacs/)
                    ("elpa" . "http://tromey.com/elpa/)))
    (add-to-list 'package-archives source t))
  (package-initialize)

  (defvar my-packages '(starter-kit zenburn-theme auctex color-theme-solarized csharp-mode ecb_snap find-file-in-project flymake-css flymake-php idle-highlight ido-ubiquitous ipython python-mode magit markdown-mode paredit pastels-on-dark-theme php-mode rainbow-mode smex solarized-theme starter-kit-js zenburn-theme)
  "A list of packages to ensure are installed at launch.")

  (dolist (p my-packages)
    (when (not (package-installed-p p))
      (package-install p)))
Where you add the package name to my-packages. It's kinda cool to see it download and compile everything on a new install.
tcas··on How SR-71 Blackbird Designers Overcame Canonical Jet Engine Limitations
On modern jet engines: yes. There's a documentary / video about the manufacture of the Rolls-Royce Trent engine that mentions this fact and shows a CAD drawing / part of the manufacture process of the blades right after the combustion stage.
tcas··on It's Dangerous, a Python cryptographic signing module
You shouldn't use urandom for crypto purposes. /dev/random is generated (on most platforms) as cryptographic strength numbers (usually from hardware), but can block if it runs out of data. /dev/urandom was created with the guarantee to never block and will use /dev/random's pool of numbers initially but can start outputting lower entropy numbers if /dev/random blocks.
tcas··on Amazing optical illusion or "glitch in the Matrix"?
There's an exhibit at the Boston Aquarium that is exactly this with a strobe light. It's really cool to see in person.
tcas··on Take Over The Galaxy with GitHub (DCPU16 support)
I'm working on a multicycle implementation that I will be able to push to a FPGA. There's no way I'll be able to get the same cycle timing as the specs indicate however. 3 cycles for a divide is very, very generous for such a simple CPU, I'll probably either end up implementing a shift and subtract algorithm (will take more than 3 cycles), or using huge look up tables (probably too big fit in a single blockRAM as well...) to try and achieve it. On the other hand, SHR and SHL are trivial to do in hardware via a barrel shifter, but he assigned 2 cycles for them.

The [next word + register] instruction is also a bit annoying to deal with in the given time tables and a simple register file design, though I haven't thought about the design of that too much.

tcas··on 2012’s top ten tech. April Fools’ Day stories
It's not. If you have a jailbroken iPhone there is an a mod you can download to enable a morse code keyboard. http://iditdahtext.com/iDitDahText.html

At this point however I became so used to the virtual keyboard on my iPhone that I doubt I would get any speed increases by switching to morse code. The only reason why it would be more convenient is to ensure typing accuracy when not looking at the phone.

← PreviousPage 5 of 6Next →