kango_v

We generate a seed based upon the Epoc (date period) that the user signs up on. This seed is stores in a seperate service (geographically different). So if you did get hold of the hashes of the passwords, there are thousands of different seeds, each of which is cryptographically secure. So, good luck trying to brute force that.

But, the point is that the RTX cards are a fudge. You still have to do all the tricks that the game designers do for the rasterisation step. Using Path Tracing you get rid of all of those "tricks" so that the designer can concentrate on the scene instead of telling the GPU to work around rasterisation's shortcomings. Ray tracing traces a single ray until it hits an object. Path tracing, on the other hand, bounces a ray around until it cannot be seen. To get better quality, you just up the samples per pixel. This is why I'd like others to see what they get on their GPUs using path tracing. We would then get an idea of how far the current cards are away from using this in games. Path tracing is not perfect, but it's the best way to get to near photo realism.

Hi all, I've been playing around with path tracing lately and have run across the Octane Renderer 4. This is a brilliant piece of software that uses your GPU to perform path tracing on a scene. Path tracing is way superior than Ray Tracing. There is a demo to download which can be run on Mac, Windows and Linux along with a benchmark. I've only got a lowly GTX760. What I'd like is for as many people on here to download it and give it a go so we can post some raw path tracing numbers for comparison. Maybe, just maybe we could get Linus to adopt it when testing all the graphics cards. Remember an RTX card only uses ray tracing, and it does this only on shadows (black and white). It also only does it at a resolution of 1 sample per pixel (the denoiser then kicks in). I'd like to see what path tracing looks like and how long it takes at 1000 samples per pixel What do you all think? Can this be achieved? Thanks all :)

Is there anywhere that documents the back-end architecture of Floatplane? I'm a technical architect for an on-line gambling game provider and as you can imagine, we have to handle a huge amount of bets (securely). I'm kind of a back-end kind of guy, having never ventured into Javascript etc, but am hot on server side stuff. Just kind of wondering about DB servers, web servers, messaging, distributed transactions, datacenter failover etc. All that jazz. Thanks.

Stick Ubuntu Server on it and then install Docker. Go to Docker Hub and start experimenting. Cool stuff. It's what I'm doing at home. Actually I have a Google Kubernetes Container Cloud at home. Can spin up VMs at will and can run practically anything!!!!

RSX-11M-PLUS

Bloody hell, i feel like an old man on this forum. Anyway, I've been coding since the age of 7 on a Sinclair ZX81 (assembler), then on to a BBC Micro (6502) assembler. Once I got my first PC it was into C, C++, DBase III+, FoxPro, Turbo Pascal, COBOL, Basic etc. My first job with using RPGIII/CLP on an IBM System 38 (what a beast), then onto an AS/400 (System i now). From there I've used VB, ASP, C#, SQL, Bash etc. Now do Java full time. I now design large distributed systems to handle on-line betting using highly async/concurrent servers using Ratpack (way faster than NodeJS). We use the same Disributed NoSQL db that Netflix uses (Cassandra) which is the dogs nuts. Oh as for IDE, we use Eclipse and IDEA with a sprinkling of VIM and Emacs for good measure. Not a Windows machine in site! LOL

I'd just like to know why NodeJS is being used and for what? We just discounted it in favour of Ratpack. Way faster and easier. Lots of better stuff out now. Also, when the code base is small it's ok, but when it gets large, trying to manage the JS with version, deps etc gets to be a real nightmare. For server side stuff, I'd go with Java, Scala, Kotlin, Groovy.

You can edit the Linux /etc/fstab file and mark the windows partition read only.

I work for a games developer and it's all written and hosted on Linux. Software runs from Ubuntu, Blender, Gimp, Inkscape, DarkTable (awesome), render farm etc. Software dev is javascript (nodejs, 3d libs), Eclipse, IDEA, VSCode, Java, etc. Back end is Apache HTTPD, NGinx, Java, JBoss, Ratpack, Coherence data grid, PostgreSQL, Apache Cassandra, CentOS, Ubuntu Server plus loads more The one thing on Linux that has changed everything for us is Docker. This has transformed our development. We now run our software in Docker containers orchestrated with Kubernetes. This is what Google uses to shuffle software around its million plus servers. We would never even think of using Windows for any of this stuff. This is all battle tested in production. We also have centrally managed software updates. We have all sorts of hardware connected to our Linux desktops from printers, scanners, video capture, some huge monitors for the 3D stuff. SANs using 40Gbe adapters getting insane throughout. When I hear people saying that Linux is not ready, I just utter "bollox" under my breath. I've been using Linux since Ubuntu 5.10 (nearly 12 years) at home and at work and would NEVER switch back. I've never paid for an OS in my life (money wise). I do contribute back to open source projects when I can though (code, testing, docs).

I've now been using Ubuntu Linux since version 5.10 (nearly 12 years!) at home and at work. I last used Windows when XP had only been out around a year (ish). When I now go back to Windows I'm lost :). So to install the Gimp: 1. Click "Software" icon 2. Click Gimp 3. Click "Install" 4. Run or from the command line: $ sudo apt install gimp $ gimp Now, that is NOT difficult. My machine at home gets used for games, software developing, video editing, kids homework. I do lots of work stuff like testing clusters of Cassandra nodes (Netflix use this), PostgreSQL, Jboss, Ratpack, various data grids etc. I've never run anti-virus and never had any malware or viruses. I've always built my own PC's so that Linux friendly hardware can be selected. Dive in! Linux today is a doddle to use. If my mother-in-law can use it, I'm pretty sure any one on this forum can

We started out with C/C++ for quite large project, but due to many hassles we ended up using Java and the SWT libraries (from Eclipse). It looked and acted like a native app on Windows, Linux and Mac OS, loading fast and proved to be very usable and stable. The users could not notice a difference between a truly native app whatever the platform. This was a massive win for us as it was truly multi platform. Oh and a new user could install it and the JVM via a HTTP URL. Such is the power of JNLP. Something I've learned over the years is to split your app (small or large) into service and client pieces. Think about the service API and test it before even writing a gui. This will enable you to expose it via a remote (e.g. REST, ProtoBuf, Netty etc) interface later if needed with hardly any client changes. This is the same way Linux tools/apps are written. You will always have a libXXXX portion and then the Python/Ruby/C++/Bash front end using the library. Makes sense.

LOL. It's been my experience that people who say that have absolutely no grasp of the problem domain. All you know is 4 to 6 tables. Our database has 6 tables, but will be using Apache Cassandra (as used by Netflix/Apple) as the RDBMS that we are using just cannot handle the load. The the RDBMS based system we use currently is created with Hibernate/Java to create the classes and then the build (Maven) can create all the SQL statements and even connect to an existing database and update it. Good for unit tests. The same would be with NHibernate. You can be up and running in Hibernate within a day.