Whoever wants to run my superfast (maybe the fastest) LZSS decompression benchmark will help me a lot to learn more about supremacy of 512bit registers.
Seeing how Haswell boasts 1TB/s L1 cache speeds made me curious how close to that amazing bandwidth one well-written memory etude can come.
The benchmark package includes 2 executables, first compiled as 64bit using 64bit GP registers, second as 32bit using 512bit ZMM registers.
The 807MB file included in the test is compressed down to 249MB (ZIP's maximum mode gives 77MB), the decompression speed is 956MB/s on my laptop with Core2 T7500.
Given that my 'memcpy()' works at 1950MB/s and i7-4770K's at 13211MB/s, I expect on Haswell speeds exceeding 6x956MB/s (for one thread), is my estimation correct?
The package: Fastest strstr-like function in C!?[^]
I will be glad to see how both Intel & AMD i.e. Haswell & Excavator perform.
Someone came up with the idea that 14th of March (written as 3.14 in US) should be celebrated as Pi Day. Of course, since in many countries, such as in India, people write the date as 14/3 or 14.3, no word on why it would be called Pi Day there.
But no harm in harmless celebration, and I thought why not use this occasion to upload a major update of my hexpi (pronounced as 'Hex-Pie') including 'Swapnajit's Pi Webservice' on Sourceforge (http://hexpi.sourceforge.net). Soon enough, the revision 0.3 of hexpi was uploaded yesterday.
Hexpi has three parts. First, it just lists first 62,500 and 1,000,000 digits of Pi in hexadecimal. Download it and use it whichever way you like. The second part is the C program pi-in-hex.c that lists an arbitrary number of digits of Pi in hexadecimal from an arbitrary location specified by the user. The program is an extension of original work by David H Bailey in early 2000 and uses the same BBP algorithm that he co-developed.
But the real intention of Hexpi is to provide a webservice that returns user specified number of hex digits from a specified location. For example, if you want to generate first 200 hexadecimal digits of Pi right after the decimal point (meaning starting from 0th location), you simply load the following web address: http://hexpi.sourceforge.net/webservices/index.php?s=0&n=200
The idea is that since Pi is an irrational transcendental number (i.e. its digits do not repeat itself in any pattern), it must be possible to use consecutive sets of n such digits as random number. This idea, as it turned out, was not new - some cryptographic algorithms (including blowfish) already uses this concept. NIST published tests for checking randomness of digits in Pi. However, such a webservice will provide a tool that provides those random sets starting from s=s1 and n=n1 and then s=s1+n1 and n=n1 on the next call and so on.
So, there is my small gift to all of you on this Pi day. Enjoy! And do not forget to file bug if you find one. Your help can only make the service better.
PS: The concept can be extended to other transcendental irrational numbers, such as sqrt(2), e and many others.
Recently while looking for Braille characters as images, I discovered a rather strange thing - there were many, many websites that taught Braille, but I could not find any that had all of its characters as separate image files, let alone in one place.
In order to fill this gap, I went ahead and created an open source project on SourceForge BrailleAlphabetGenerator hoping it would be useful for someone. The intention was to keep the image parameters customizable.
I am seeking feedback from all of you on how to improve the project - code, visual, or anything else (except perhaps, 'Why not use Unicode?').
I built a Windows 8 app and there's apparently a design problem that my computer doesn't show. I'm losing users left and right but I can't get a good answer from them exactly what's wrong.
If just one or two of you were able to tell/show (screenshots) me what's wrong by looking at my app, I would appreciate that more than words can say.
The app is a time-lapse app so it requires a webcam. Anyone with 8/8.1 should be able to run it. I don't believe you have to spend long with it. All I'm trying to figure out is why I got one (and only one) report about the "stop" button being missing during recording.
The app is here for those who can help. Thank you!
I have two major projects on my hands that I need some help with. I've spec'd them out so that they are WPF based (for a lot of reasons). Unfortunately I don't have a lot of WPF experience and I'm looking for somebody to help work on the WPF side of these projects with me. I've read some good books on WPF but with a kid on the way I'd like to get these projects moving a little faster.
The first project is a visual scripting system similar to FlowHub[^] but designed to be integrated into other applications. I have the first version working and I'm working on version 2. This is the first project that I need help with the designer side.
The second one is also a designer in WPF but much closer to a full forms designer.
If you are interested in helping out let me know (either post here or hit the Email button below my post).
I need a tester running Windows 8.1 to test an application I've rewritten that keeps track of electrical components. Basically just to make sure it Initializes, creates DB and can access Config file and Registry, shouldn't take more than about 15-20 minutes of your time.
I don't have and or have access to a Weight machine (I like how that sounds )
I should be able to take a shot at it next weekend (I run Windows 8.1), just answer as private email to this post.
If you are interested you can compile it to run on W8.1 RT as well, I could test that too.
Veni, vidi, caecus | Everything summarizes to Assembly code
Thanks I've run into a few problems and I've been delayed on releasing the Beta. Got a tooth that needs pulled and I've not been able to get it pulled until tomorrow because it was/is infected. I don't have much to do but have not been able to concentrate.