Spork Boards

Swiffies for...

johnny k – December 27, 2007 02:31PM Reply Quote
Let's balance out every milkshake with a swiffie. Everyone else on the internet only complains.

-

So I bitched about AT&T DSL and it's only fair that I throw them a bone when deserved. In the process of transferring service to my roommate (fairly straightforward), I asked that a 411 charge be removed from my bill because we only have the landline to get DSL, and there's not even a phone hooked up. They credited my account without a question.

btw, why do all these phone trees ask for your account number before they direct you, and then when you're talking to a person, they just ask for it again? I was transferred twice, and I had to give the number 4 times.



Edited 2 time(s). Last edit at 12/27/2007 02:46PM by johnny k.

tliet – August 04, 2009 02:24PM Reply Quote
DPBD

Although I'm not sure just how long Sun under Oracle can afford to crank out a product like this, but for the price VirtualBox is a pretty good piece of software, you just can't argue with free. I've been running it quite successfully on my frankenmac and it's been really, rock solid. Not sure about the speed as I haven't run VMWare nor Parallels.

ghidorah – October 07, 2009 09:34AM Reply Quote
Raise taxes on cavemen. --jw
http://www.productwiki.com/

although unbiassed reviews might be stretching it

John Willoughby – December 03, 2009 02:53PM Reply Quote
Cyberdyne Systems Customer Support
Naked Mole Rats, once again!

http://news.yahoo.com/s/livescience/nakedmoleratssurviveextremeoxygendeprivation

Hey, there doesn't seem to be a fecal-chirp tag defined in HTML. How odd.

YDD – February 16, 2010 05:13PM Reply Quote
Am I totally l33t or what...
$ ./meanfilter_test --width=34
Mean Filtering Tester
=====================

CPU Version (Xeon W5580 @ 3.20GHz)

Filtering complete in 3.69872e+06 ms
Writing output file: cpublur.out



$ ./meanfilter_test_cuda --width=34
Mean Filtering Tester
=====================

Acquiring CUDA device
Using default device
CUDA device: Tesla C1060

Filtering complete in   136.063 ms
Writing output file: gpublur.out


$ vol_diff cpublur.out gpublur.out
No difference
$

Simon – February 18, 2010 05:34AM Reply Quote
I have no idea what any of that means so I'm gonna say yes.

tomierna (Admin) – February 18, 2010 09:30AM Reply Quote
Hideously Unnatural
YDD is showing us that, while his CPU takes a long time to render a specific blur function, his CUDA-enabled GPU (the hotshit NVidia Tesla) runs that specific blur function much much faster, with identical results.

YDD – February 18, 2010 09:51AM Reply Quote
Quote

while his CPU takes a long time to render a specific blur function
I take offense at that statement! ;-)

It's not my blur function, which is why I got the ludicrous speed up (on the order of 25,000x). I was given the CPU version - a bad implementation of an O(n3) algorithm, and told to make it run on the GPU. I changed it to an equivalent O(n) one when doing my CUDA version. Based on other tests, the GPU speed up is only about 30x (give or take a few PCIe transfers). All the rest of it comes from using an algorithm of much lower complexity.

tomierna (Admin) – February 18, 2010 11:35AM Reply Quote
Hideously Unnatural
Have you re-worked the bad implementation of the algorithm for the CPU as well to see how much improvement you'd get there?

YDD – February 18, 2010 12:33PM Reply Quote
Quote

Have you re-worked the bad implementation of the algorithm for the CPU as well to see how much improvement you'd get there?
No, although I have pointed out the problem in the bug tracking database (such as it is). Part of the problem is the arr[j][k] semantics which pervade the CPU code. For the rest.... it's C code which is trying to be templated, which means that there are switch statements all over the place. I generally prefer to avoid touching that without a very long pole.

John Willoughby – February 18, 2010 01:34PM Reply Quote
Cyberdyne Systems Customer Support
I keep running into code like this:

if (computationally_complex_condition)
{
a = b;
}
a = b;


Who writes this stuff?

YDD – February 18, 2010 01:37PM Reply Quote
And I forgot to mention.... on the CPU side, some arrays are arr[j][k] while others are arr2[k][j]. Cache coherency anyone?

Cloudscout – February 18, 2010 01:47PM Reply Quote
Det finnes ikke dårlig vær, bare dårlige klær!
Quote
John Willoughby
Who writes this stuff?

Did you guys hire Wipro or something?

John Willoughby – February 18, 2010 02:27PM Reply Quote
Cyberdyne Systems Customer Support
Actually, I got stuck doing code reviews for a newly acquired subsidiary whose products were having a lot of problems in the field. The horror... the horror...

stan adams – February 18, 2010 02:32PM Reply Quote
Without code like that it would be too easy to spot the "rounding errors of a cent going to a secret account" sub routines, and without THAT neither Lex Luther nor the dudes from Office Space would have had a chance...

John Willoughby – February 18, 2010 03:20PM Reply Quote
Cyberdyne Systems Customer Support
I especially love code like this, which was EVERYWHERE:

int8 retCode = NO_ERROR;

retCode = DoSomething();
retCode = DoSomethingElse();
retcode = DoYetAnotherThing();

if (retCode == ERROR)
{
return SOMETHING_WENT_WRONG;
}
else
{
return EVERYTHING_IS_SUPER;
}

How hard is the flaw here to see? I mean, people made a lot of money writing this code...

YDD – February 18, 2010 03:27PM Reply Quote
I hope that for a bugfix you suggested overloading the '=' operator so that if the LHS value was not NO_ERROR, the old value would be left ;-)

And some time I'll proofread before hitting 'post'



Edited 3 time(s). Last edit at 02/18/2010 03:29PM by YDD.

John Willoughby – February 18, 2010 04:06PM Reply Quote
Cyberdyne Systems Customer Support
I don't favor overloaded operators for anything but very obvious usages. Too much confusion in group code. If they wanted that structure, they could have done something with |= and used different bits for different errors, but I'd have been happier if they just handled the errors as they detected them. Especially since some of the later functions depended on preceding functions completing successfully.



Edited 1 time(s). Last edit at 02/18/2010 04:08PM by John Willoughby.

YDD – February 18, 2010 04:53PM Reply Quote
Quote

I don't favor overloaded operators for anything but very obvious usages
But how're going to confuse people if you do that? I'd like to submit another little gem from the codebase I'm working on right now:
/*  compute the triple scalar product v1 x v2 . v3 */
float VectorTripleProduct( const VECTOR *v1, const VECTOR *v2, const VECTOR *v3)
Yes folks. The routine "VectorTripleProduct" does not, in fact, calculate the Vector Triple Product, but the Scalar Triple Product. Just a slight difference there....

Mokers (Moderator) – February 21, 2010 04:53PM Reply Quote
Formerly Remy Martin
Microsoft silverlight. A lot of streaming video sites use it now, including Netflix. And although Mac support lags on some of the latest features, the browser plugin is much less buggy and much more compact than Flash.

Cloudscout – February 25, 2010 05:31AM Reply Quote
Det finnes ikke dårlig vær, bare dårlige klær!
There's a neat site here in India called Burrp! which I came across when I was trying to find a good place to get local TV listings. I have been using it for a couple of weeks and noticed that vertical line they draw through the listing grid to show where the current time lines up was always slow by 10 minutes or so. I finally decided to click their Feedback link and told them that I found their site really useful and mentioned that I thought the clock might be off on their server. The next morning, the problem was corrected and today I got an email from them:

Quote

Hi,

Thanks a ton for your feedback and the great catch with the time lag. We have made the change. Please take a look and let us know if we're still not on track.

Thanks and regards,

Aarti Jesrani
Director - Content & Communications
burrp!
http://www.burrp.com/

That kind of friendly response is good for business. I've already recommended it to some of my colleagues here who didn't realize there was a place they could look up free TV listings in a convenient interactive grid like that without having to pay for digital TV service. I've also started poking around the rest of the site with their restaurant reviews, etc.

Sorry, only registered users may post in this forum.

Click here to login