@r.tanner.f upgrading without doing a clean install is complete fail if you have any non-trivial programs installed
probability of something installed that breaks the upgrade that isn't detected by the compatibility assistant thing increases with the age of your Win7 install and reaches approximately "1" (guaranteed) after about 2 years' old install
but then I install things like multiple virtualization programs, wireshark, visual studio, multiple versions of sql server and java, steam, origin, etc
@Sathya uhm, I'm not sure, but I don't think Intel has (yet?) released a WDDM 1.2 driver for older than Sandy Bridge graphics, so you'll probably be stuck with WDDM 1.1 running in compatibility mode (if it works at all)
@r.tanner.f Radeon HD4000 is not supported in Windows 8, AFAIK. It isn't supported in Windows 7, either. All Radeon HD4000 and earlier cards are completely unsupported by AMD. The only driver updates you'll ever see for them are likely to be security fixes, if that.
They say they will supposedly push updates as requested by their big dollar customers, but so far there's been very little
AMD lost a lot of engineers in the Catalyst team so they dropped support for cards that were still being sold in new laptops at retail outlets in 2011... it's cool, yo, we don't need no stinkin' drivers (sarcasm)
well... if your hybrid graphics solution has some kind of special sauce that makes it not work with drivers directly from Intel and AMD, then yeah, you're probably pretty hosed
system integrators are known for being extremely sloppy and irresponsible with not supporting their hardware in the long run
oh well; that just means that unsupported hardware works infinity% faster on Linux than it does on Windows :D (since the open source graphics drivers support all the way back to ATI Rage chips, and HD4000 generation is often faster than Catalyst itself)
I am getting so pissed off at work... My computer is working fine for day to day stuff, but, whenever it is at the lock screen, I can come back to my machine and it usually has restarted... if I am logged in remotely, it can also restart - so, it is something about having the login screen locally
if you get an error when trying to run the driver after that then it should be something like it can't find the file or the side by side configuration is incorrect
if you get that then you can start reinstalling only the latest version of the redistribs until you hit gold
@WilliamHilsum sounds stupid but do you have the latest drivers for your mobo and GPU?
sounds like an ACPI issue... what I'd do is lock your screen then sit there and watch it and I'll bet you a $5 game on Steam that it BSODs when it puts the monitor to sleep
@jokerdino on win8 you really don't need them, but some program installers will force them upon you anyway... generally it's pointless to have older versions of them though, you really only need the latest because it ships the older versions too in SxS for backwards compatibility
@allquixotic No, but, My machine has been fine for the past 4-5 months (when I rebuilt it), this has just started in the last week and I haven't done any updates
@allquixotic Ok... Turn off display is set to 15 minutes... I'll try disabling that - now you say it, I have been having a bad connection to my monitor (via HDMI to DVI adapter) - I have had to shake it a bit and vibrations make it go bad (waiting for new one to arrive)... Technically, I don't see why it should cause a reboot, but... bloody computers... maybe that is it!
I hope this is correct, and if it is, thanks so much!
that might be the fix... stupid buggy driver could do something like: (1) wait 15 minutes to turn off display (2) time to turn off the display (3) WTF THE DISPLAY DOESN'T EXIST?!?!? (4) BSOD
@avirk yeah but the term will survive until well after Win8, Win9, Win10, etc... even when the new BSOD screen is a picture of Steve Ballmer crying and the system boots a recovery kernel to play rickroll on youtube to let you know that your system crashed
Yes it does! You can find it under
C:\Windows\winhlp32.exe
Its size is ~11kb and there is winhlp32.mui.exe in the following path separately whose size is 3kb only.
C:\Windows\WinSxS\x86_microsoft-windows-winhstb.resources_31bf3856ad364e35_6.2.9200.16384_en-us_cf05e380ff44d465
The...
@avirk the only time the compiler isn't available is if you have the Client Profile of .NET installed; if you have the "full" profile then you get the compiler, so all you need is a good IDE like SharpDevelop or MonoDevelop, or just use msbuild / csc.exe directly
@Sathya if you have a system restore checkpoint or entire disk image of win7 from backups then yes, otherwise, no... the rollback feature IIRC only works if the upgrade fails during setup
the MonoDevelop project can either run on the Mono framework (which doesn't require any installation; you can just download the binaries and run it) or on top of the .NET framework
SharpDevelop AFAIK is programmed to run with the .NET framework and not really tested on mono
but mono is pretty cool in that you don't have to carry around all the ugly trappings of a .NET install, all you need is the ability to download and execute EXEs and you've got a fully working .NET programming / development / execution environment, even if system doesn't have .NET framework
C# isn't native... that's MSIL... which looks native because it says ".exe" but really it isn't because it's interpreted binaries that depends on .NET assemblies rather than x86 machine code
you have to get Visual Studio C++ or MinGW to get a native compiler to emit x86 machine code (or ARM, for that matter)
the .NET executables and DLLs actually run "natively" on Mono on non-Windows platforms... it's weird because you can wire up a custom handler to the dynamic loader and type ./binary.exe on Linux and it'll wrap it in Mono, and natively execute an .exe depending on .dlls on Linux
not so for x86 native code in the Win32/PE format
because that depends on the Win32 API which is not available on non-Windows platforms
there's actually a bit of conflict there because you have to use a utility that can distinguish between Win32/PE executables and PE/MSIL executables in order to send native code to wine and MSIL to mono... you can't just go on the file extension
unless you want to run mono or the .NET framework in wine, which is weird
@allquixotic All I know is that I can execute .NET binaries on our Ubuntu systems by simply invoking ./awesome.exe and I've never set up anything to achieve it. So I assumed it's done by default with the installation
@OliverSalzburg weird, I thought the default was for wine to do that
maybe they already solved that problem
it gets freaky though because with P/Invoke or certain compiler flags, you can embed native Windows code in an otherwise clean MSIL executable or assembly, which ties it to the Windows platform
so you'd really have to look kind of "deep" into the file format to determine whether to run it in the .NET framework on top of wine (for native code bits or P/Invoke) or on the native mono framework on linux without wine (for pure MSIL code)
that's actually a really good idea if it hasn't been done already: collaboration between wine and mono projects... write one unified executable loader
executable loader relies on a library that analyses the executable to determine whether it's (a) just win32 code, (b) MSIL and win32, or (c) just pure MSIL with no platform dependency, or P/Invokes to native Linux
in case (a), run it in wine; in case (b), run it in mono on top of wine (and ship mono default embedded in wine distribution); in case (c), run it in mono
I purchased a new notebook and I'm trying to reinstall Windows. I could start the installation once via external USB DVD Drive. I deleted the old partitions and had to reboot. After that the computer will not boot from the external USB Drive. I also tried the transfer the image to USB stick. Stil...
@Bob I'm actually twice as sadface as Microsoft themselves that more games aren't written in .NET, because the portability story would be a lot better (unless they use DirectX, blah)
thing is, they tried (...oh my god did they try; they tried all the time, in this institution...) and people still didn't want to program games in .NET
game programming or almost anything else with a lot of real time updates and events produces a lot of heap garbage, and garbage collectors are pathetically slow
it's just so much easier to set up a static location for all the game data in memory and continuously bang on those memory addresses rather than having them jump all around all the time and have the heap allocator going nuts trying to size everything up and manage fragments and ugh
easier in terms of performance cost, not design ease
heap allocators are slow; garbage collection routines are slow; combining very frequent heap allocations and very frequent garbage collection as you have to do in most real-time software, and you get fail
you notice it less in throughput computing (server-side, for instance) because you don't have a 16.67 millisecond deadline for pumping through another frame (assuming 60 fps) like you do in real time rendering
heap allocators and garbage collectors very often have to suspend all threads in the application, and for a very dirty virtual address space can exhibit runtimes on the scale of 250 to 1500 milliseconds
that's way more than the deadline of 16.67 milliseconds to keep consistent 60 fps performance
a traditional C/C++ approach involves keeping all your clothes nice and neat, folded, and in drawers... when you use them, you wash them and put them right back in the same place in the same drawer
garbage collection a la virtual machines (python, ruby, .NET, java, etc) involves throwing your laundry in a pile when they're dirty, and then when you wash them you just throw them back in whichever place they fit in in whatever drawer
in terms of understanding the underlying CPU instructions, it's extremely educational for you to write a simple C/C++ program that does heap allocation (e.g. malloc()) and run it in a debugger and step through the C runtime's instructions for allocating a simple block of memory
do the same for a naive Boehm garbage collection implementation and you can see how this can take dizzying amounts of time
the way "low level" C/C++ programs work is, they do allocate memory on the heap, but they do it once at program startup
and they just re-use the existing, already allocated memory hunks to store different data in, as the data changes
but memory managed runtimes like .NET constantly re-allocate everything
moving it around in memory, like scrambling the pieces to a puzzle
the space for memory that the programmer requests is automatically handled both in terms of making space in memory for the data, and also for removing the data when no longer needed
the heap allocator takes care of making the space, and the garbage collection takes care of getting rid of objects that no longer can be referenced, i.e. they are essentially dead
native code that does not involve automatic memory management has to manually keep track of both the allocation and de-allocation of every memory chunk
automatic memory management makes memory leaks much less common, and much easier to find when they do exist, compared to manual memory management
but if you have a native program with manual memory management that does not have any memory leaks, it will always vastly out-perform a program that uses automatic memory management
for server workloads where deadlines are measured in seconds, this rarely matters... who cares if your HTTP response is delayed by 300 msec because the garbage collector was running?
but every single frame in a 3d game has a deadline of 16.67 msec to operate at a consistent 60 fps
and you can only buffer a few frames at a time because the output constantly changes based on input events from the network, keyboard, and mouse
you can't afford to sit around and dawdle for 300 msec, let alone 1500
@MichaelHampton yeah, but compared to network latency, the cost of round trips, and the cost of hitting your database and generating the HTTP response, 300 msec for garbage collection is peanuts
all other things being equal it's just not that important unless your GC is single threaded and blocks the entire jvm for ~10 seconds
Java 6/7's concurrent mark and sweep garbage collector really doesn't take all that long in terms of server workloads
also, server workloads spend significantly less time doing GC than real time rendering, because real time rendering is constantly generating enormous amounts of garbage, whereas servers do fewer and less frequent allocations because you don't need to process an http response every 16.67 msec
you might get one customer in every 1000 getting hit with GC latency on a server, whereas if you're copying tons of pixmaps and stuff in a game, you're going to hit the GC threshold every 2-3 seconds
I'm not so sure that I believe that people really cancel out over 100 ms... that would only be a factor, I think, if the rest of the pipe is already abysmally slow on the order of 8-9 seconds... people would probably click away more often around 9.8 seconds versus 9.5, but not 1.3 versus 1.0
@r.tanner.f one way to do it is to get the Java virtual machine to do what C/C++ programmers do, which is to allocate once and bang on it continuously and re-use the existing allocation
there are ways to mitigate the problems, and minecraft, for example, is a fairly complex 3d game based on Java -- but Minecraft uses a lot of native code to prop itself up; a lot of the "critical path" stuff that it does which involves a lot of memory reads/writes is done in native code
writing a pure Java or pure .NET real-time 3d game is pretty much folly, where I mean "pure" in the sense of, like, using JNA to dynamically bind to the OpenGL runtime and doing all of your processing in Java or managed code, so that you can seamlessly port your game with no changes to any platform with Java and libGL.so
it's possible, but it's just going to be stuttery, and there's not a lot you can do
the only successful Java or .NET based games I've seen use a ton of native code for the tight loops, etc
and imho if you're using native code you may as well not be using a VM to begin with
because you run into integration issues when you pick up a new native library you want to use, but you have to bind it into the VM runtime through some kind of foreign function interface (JNI, JNA, P/Invoke, etc)
I'm trying to create a user using this instructions.
But i'm still getting the login failed message. From local or from a remote machine, when i try connect.
Any help?
nvidia still supports cards that are like 5-6 years old, maybe more, with their legacy driver series that still gets support for new OSes, new Linux kernels, new Xorg, etc
@Hennes they're still pretty okay on the open source side; they expanded their open source team since the initial creation of it, and didn't lay them off in the recent cut
@allquixotic Solar 2 is a game on XNA/.NET. Not sure how much, if any, native code it uses but it does have nice (if slightly simple) graphics, and little to no stutter.
Yes it does! You can find it under
C:\Windows\winhlp32.exe
Its size is ~11kb and there is winhlp32.mui.exe in the following path separately whose size is 3kb only.
C:\Windows\WinSxS\x86_microsoft-windows-winhstb.resources_31bf3856ad364e35_6.2.9200.16384_en-us_cf05e380ff44d465
The...
U want to add my gmail address to this new phone i had my gmail address in a iphone
I'm trying to create a user using this instructions.
But i'm still getting the login failed message. From local or from a remote machine, when i try connect.
Any help?
