int64.org

My tips for efficient I/O are relevant all the way back to coding for Windows 2000. A lot of time has passed since then though, and for all the criticism it got, Windows Vista actually brought in a few new ways to make I/O even more performant than before.

This will probably be my last post on user-mode I/O until something new and interesting comes along, completing what started a couple weeks ago with High Performance I/O on Windows.

Synchronous completion

In the past, non-blocking I/O was a great way to reduce the stress on a completion port. An unfortunate side-effect of this was an increased amount of syscalls — the last non-blocking call you make will do nothing, only returning WSAEWOULDBLOCK. You would still need to call an asynchronous version to wait for more data.

Windows Vista solved this elegantly with SetFileCompletionNotificationModes. This function lets you tell a file or socket that you don’t want a completion packet queued up when an operation completes synchronously (that is, a function returned success immediately and not ERROR_IO_PENDING). Using this, the last I/O call will always be of some use — either it completes immediately and you can continue processing, or it starts an asynchronous operation and a completion packet will be queued up when it finishes.

Like the non-blocking I/O trick, continually calling this can starve other operations in a completion port if a socket or file feeds data faster than you can process it. Care should be taken to limit the number of times you continue processing synchronous completions.

Reuse memory with file handles

If you want to optimize even more for throughput, you can associate a range of memory with an unbuffered file handle using SetFileIoOverlappedRange. This tells the OS that a block of memory will be re-used, and should be kept locked in memory until the file handle is closed. Of course if you won’t be performing I/O with a handle very often, it might just waste memory.

Dequeue multiple completion packets at once

A new feature to further reduce the stress on a completion port is GetQueuedCompletionStatusEx, which lets you dequeue multiple completion packets in one call.

If you read the docs for it, you’ll eventually realize it only returns error information if the function itself fails — not if an async operation fails. Unfortunately this important information is missing from all the official docs I could find, and searching gives nothing. So how do you get error information out of GetQueuedCompletionStatusEx? Well, after playing around a bit I discovered that you can call GetOverlappedResult or WSAGetOverlappedResult to get it, so not a problem.

This function should only be used if your application has a single thread or handles a high amount of concurrent I/O operations, or you might end up defeating the multithreading baked in to completion ports by not letting it spread completion notifications around. If you can use it, it’s a nice and easy way to boost the performance of your code by lowering contention on a completion port.

Bandwidth reservation

One large change in Windows Vista was I/O scheduling and prioritization. If you have I/O that is dependant on steady streaming like audio or video, you can now use SetFileBandwidthReservation to help ensure it will never be interrupted by something else hogging a disk.

Cancel specific I/O requests

A big pain pre-Vista was the inability to cancel individual I/O operations. The only option was to cancel all operations for a handle, and only from the thread which initiated them.

If it turns out some I/O operation is no longer required, it is now possible to cancel individual I/Os using CancelIoEx. This much needed function replaces the almost useless CancelIo, and opens the doors to sharing file handles between separate operations.

Visual Studio 2010 Beta 1 was released yesterday for MSDN subscribers. Probably the most anticipated release in a while for C++ developers, 2010 is Microsoft’s attempt to give C++ first-class support, something which hasn’t been seen since Visual Studio 6.0.

Update: downloads are now available for non-MSDN subscribers.

On the compiler side of things, we get partial C++0x support in the form of lambda expressions, rvalue references, auto, decltype, and static assert. The features are piled on with an improved TR1 library — finally including the much requested stdint.h and cstdint headers, but still lacking inttypes.h.

Also included is the Parallel Patterns Library, a new task-based concurrency library that makes heavy use of the C++0x features for a nice modern design. I mentioned before that on Windows 7 this will make use of a User-mode scheduled thread pool so it should be really efficient. Unfortunately given its proprietary nature I’m not sure how much use it will get.

The first thing you will notice on the IDE side is the inline error checking. Something we’ve enjoyed while editing C# for some time, we now get the red squiggly lines when an error is found. It works fairly well, but support for lambda expressions has not been written yet.

Intellisense has markedly improved since 2008. Using advanced C++ or a Boost library no longer guarantees it breaking. It has worked with nearly all the C++ I’ve thrown at it so far.

You can also see an External Dependencies virtual folder added to your project source, which is dynamically filled with all the files Intellisense will scan. I’ve found it is not terribly useful, though, because even with small projects the header count increases rapidly enough to make the virtual folder become an unintelligible mess.

The problem is only aggravated by libraries like Boost, which have hundreds of headers organized nicely in folders. Putting them into a single virtual folder just doesn’t work.

This release also marks the move to the extensible MSBuild system for C++ projects, which aims to provide functionality similar to GNU make in an XML format.

Perhaps the most obvious change for the overall IDE is that the main UI is now done entirely in WPF. It sounded like a decent plan at first but I’m not too happy with it now. Minor differences from the way native controls behave can be pretty annoying, and the five to twenty second load time makes it less useful for opening random .cpp files, when 2008 would load them in one or two seconds.

Over the past week I’ve been rewriting my rather dated SQLite wrapper to have an efficient, modern C++ feel. The basic wrapper is there, but I was looking for something a little more this time.

While looking at the problem I decided I was spending too much time writing boilerplate SQL for all my types so I decided to look at existing ORM frameworks. I’m pretty picky about my C++ though, and couldn’t find anything I liked so I started writing my own. Instead of creating a tool to generate C++, I wanted to take a pure approach using native C++ types and template metaprogramming.

What I ended up with is not a full ORM framework, and I’m not particularly interested in making it one. All I’m aiming for is removing boilerplate code while leaving it easy to extend it for more complex queries. Here’s what I’ve got so far:

struct my_object
{
  int id;
  std::string value;
  boost::posix_time::ptime time;
};

typedef boost::mpl::vector<
  sqlite3x::column<
    my_object, int, &my_object::id,
    sqlite3x::primary_key, sqlite3x::auto_increment
  >,
  sqlite3x::column<
    my_object, std::string, &my_object::value,
    sqlite3x::unique_key
  >,
  sqlite3x::column<
    my_object, boost::posix_time::ptime, &my_object::time
  >
> my_object_columns;

typedef sqlite3x::table<
  my_object,
  my_object_columns
> my_object_table;

Using it is pretty simple. It uses the primary key as expected, generating the proper WHERE conditions and even extracting the type to let find() and others specify only the primary key:

sqlite3x::connection con("test.db3");

my_object_table my_objects(con, "t_objects");

my_objects.add(my_object());
my_objects.edit(my_object());
my_objects.remove(int());
my_objects.exists(int());
my_objects.find(int());

One benefit of the approach taken is it makes working with single- and multiple-inheritance just as easy:

struct my_derived :
  my_object
{
  float extra;
};

typedef boost::mpl::copy<
  boost::mpl::vector<
    sqlite3x::column<my_derived, float, &my_object::extra>
  >,
  boost::mpl::back_inserter<my_object_columns>
> my_derived_columns;

typedef sqlite3x::table<
  my_derived,
  my_derived_columns
> my_object_table;

The next thing on the list was supporting types not known natively to sqlite3x. I did not want to have the headache of sub-tables, so I took the easy route and implemented basic serialization support:

struct my_derived :
  my_object
{
  std::vector<boost::uuid> uuids;
};

struct uuids_serializer
{
  static void serialize(std::vector<boost::uint8_t> &buffer,
     const std::vector<boost::uuid> &uuids);

  template<typename Iterator>
  static Iterator deserialize(std::vector<boost::uuid> &uuids,
     Iterator first, Iterator last);
};

typedef boost::mpl::copy<
  boost::mpl::vector<
    sqlite3x::column<
      my_derived, float, &my_object::extra,
      sqlite3x::serializer<uuids_serializer>
    >
  >,
  boost::mpl::back_inserter<my_object_columns>
> my_derived_columns;

A few things aren’t finished, like specifying indexes and support for multi-column primary keys.

Overall though, I’m pretty happy with it. The majority of what I use SQLite for doesn’t require many complex queries, so this should greatly help lower the amount of code I have to manage.

Best of all this ORM code is in an entirely isolated header file — if you don’t want it, just don’t include it and you’ll still have access to all the basic SQLite functions. Even with it included I kept to the C++ mantra of “dont pay for what you don’t use” — as it is entirely template-driven, code will only be generated if you actually use it.

Once I’m finished the code will replace what I have up on the SQLite wrapper page, but until then it will exist in the subversion repository only.

There are a few things to keep in mind for I/O that can have pretty incredible effects on performance and scalability. It’s not really any new API, but how you use it.

Reduce blocking

The whole point of I/O completion ports is to do operations asynchronously, to keep the CPU busy doing work while waiting for completion. Blocking defeats this by stalling the thread, so it should be avoided whenever possible. Completion ports have a mechanism built in to make blocking less hurtful by starting up a waiting thread when another one blocks, but it is still better to avoid it all together.

This includes memory allocation. Standard system allocators usually have several points where it needs to lock to allow concurrent use, so applications should make use of custom allocators like arenas and pools where possible.

This means I/O should always be performed asynchronously, lock-free algorithms used when available, and any remaining locks should be as optimally placed as possible. Careful application design planning goes a long way here. The toughest area I’ve discovered is database access — as far as I have seen, there have been zero well designed database client libraries. Every one that I’ve seen has forced you to block at some point.

Ideally, the only place the application would block is when retrieving completion packets.

Buffer size and alignment

I/O routines like to lock the pages of the buffers you pass in. That is, it will pin them in physical memory so that they can’t be paged out to a swap file.

The result is if you pass in a 20 byte buffer, on most systems it will lock a full 4096 byte page in memory. Even worse, if the 20 byte buffer has 10 bytes in one page and 10 bytes in another, it will lock both pages — 8192 bytes of memory for a 20 byte I/O. If you have a large number of concurrent operations this can waste a lot of memory!

Because of this, I/O buffers should get special treatment. Functions like malloc() and operator new() should not be used because they have no obligation to provide page-aligned data. The buffers also shouldn’t be embedded with other data. I like to use VirtualAlloc to allocate large blocks of 1MiB, and divide this up into smaller page-sized (4KiB) blocks to be put into a free list.

Limit the number of I/Os

System calls and completion ports have some expense, so limiting the number of I/O calls you do can help to increase throughput. We can use scatter/gather operations to chain several of those page-sized blocks into one single I/O.

A scatter operation is taking data from one source, like a socket, and scattering it into multiple buffers. Inversely a gather operation takes data from multiple buffers and gathers it into one destination.

For sockets, this is easy — we just use the normal WSASend and WSARecv functions, passing in multiple WSABUF structures.

For files it is a little more complex. Here the WriteFileGather and ReadFileScatter functions are used, but some special care must be taken. These functions require page-sized buffers to be used, and the number of bytes read/written must be a multiple of the disk’s sector size. The sector size can be obtained with GetDiskFreeSpace.

Non-blocking I/O

Asynchronous operations are key here, but non-blocking I/O still has a place in the world of high performance.

Once an asynchronous operation completes, we can issue non-blocking requests to process any remaining data. Following this pattern reduces the amount of strain on the completion port and helps to keep your operation context hot in the cache for as long as possible.

Care should be taken to not let non-blocking operations continue on for too long, though. If data is received on a socket fast enough and we take so long to process it that there is always more, it could starve other completion notifications waiting to be dequeued.

Throughput or concurrency

A kernel has a limited amount of non-paged memory available to it, and locking one or more pages for each I/O call is a real easy way use it all up. Sometimes if we expect an extreme number of concurrent connections or if we want to limit memory usage, it can be beneficial to delay locking these pages until absolutely required.

An undocumented feature of WSARecv is that you can request a 0-byte receive, which will complete when data has arrived. Then you can issue another WSARecv or use non-blocking I/O to pull out whatever is available. This lets us get notified when data can be received without actually locking memory.

Doing this is very much a choice of throughput or concurrency — it will use more CPU, reducing throughput. But it will allow your application to handle a larger number of concurrent operations. It is most beneficial in a low memory system, or on 32-bit Windows when an extreme number of concurrent operations will be used. 64-bit Windows has a much larger non-paged pool, so it shouldn’t present a problem if you feed it enough physical memory.

Unbuffered I/O

If you are using the file system a lot, your application might be waiting on the synchronous operating system cache. In this case, enabling unbuffered I/O will let file operations bypass the cache and complete more asynchronously.

This is done by calling CreateFile with the FILE_FLAG_NO_BUFFERING flag. Note that with this flag, all file access must be sector aligned — read/write offsets and sizes. Buffers must also be sector aligned.

Operating system caching can have good effects, so this isn’t always advantageous. But if you’ve got a specific I/O pattern or if you do your own caching, it can give a significant performance boost. This is the easiest thing to turn on and off, so take some benchmarks.

Update: this subject continued in I/O Improvements in Windows Vista.

Visual Studio 2010 Beta 1 is being launched for MSDN subscribers in four days on the 18th, and for the general public on the 20th!

I described the basics of I/O completion ports in my last post, but there is still the question of what the easiest way to use them. Here I’ll show a callback-based application design that I’ve found make them really simple to use.

I touched briefly on attaching our own context data to the OVERLAPPED structure we pass along with I/O operations. It’s this same idea that I’ll expand on here. This time, we define a generic structure to use with all our operations, and how our threads will handle them while dequeuing packets:

struct io_context
{
  OVERLAPPED ovl;
  void (*on_completion)(DWORD error, DWORD transfered,
                        struct io_context *ctx);
};

OVERLAPPED *ovl;
ULONG_PTR completionkey;
DWORD transferred;

BOOL ret = GetQueuedCompletionStatus(iocp, &transferred,
   &completionkey, &ovl, INFINITE);

if(ret)
{
  struct io_context *ctx = (struct io_context*)ovl;
  ctx->on_completion(ERROR_SUCCESS, transferred, ctx);
}
else if(ovl)
{
  DWORD err = GetLastError();

  struct io_context *ctx = (struct io_context*)ovl;
  ctx->on_completion(err, transferred, ctx);
}
else
{
  // error out
}

With this, all our I/O operations will have a callback associated with them. When a completion packet is dequeued it gets the error information, if any, and runs the callback. Having every I/O operation use a single callback mechanism greatly simplifies the design of the entire program.

Lets say our app was reading a file and sending out it’s contents. We also want it to prefetch the next buffer so we can start sending right away. Here’s our connection context:

struct connection_context
{
  HANDLE file;
  SOCKET sock;

  WSABUF readbuf;
  WSABUF sendbuf;

  struct io_context readctx;
  struct io_context sendctx;
};

A structure like this is nice because initiating an I/O operation will need no allocations. Note that we need two io_context members because we’re doing a read and send concurrently.

Now the code to use it:

#define BUFFER_SIZE 4096

void begin_read(struct connection_context *ctx)
{
  if(ctx->readbuf.buf)
  {
    // only begin a read if one isn't already running.
    return;
  }

  ctx->readbuf.buf = malloc(BUFFER_SIZE);
  ctx->readbuf.len = 0;

  // zero out io_context structure.
  memset(&ctx->readctx, 0, sizeof(ctx->readctx));

  // set completion callback.
  ctx->readctx.on_completion = read_finished;

  ReadFile(ctx->file, ctx->readbuf.buf, BUFFER_SIZE,
           NULL, &ctx->readctx.ovl);
}

void read_finished(DWORD error, DWORD transferred,
                   struct io_context *ioctx)
{
  // get our connection context.
  struct connection_context *ctx =
     (struct connection_context*)((char*)ioctx -
        offsetof(struct connection_context, readctx));

  if(error != ERROR_SUCCESS)
  {
    // handle error.
    return;
  }

  if(!transferred)
  {
    // reached end of file, close out connection.
    free(ctx->readbuf.buf);
    ctx->readbuf.buf = 0;
    return;
  }

  // send out however much we read from the file.

  ctx->readbuf.len = transferred;

  begin_send(ctx);
}

This gives us a very obvious chain of events: read_finished is called when a read completes. Since we only get an io_context structure in our callback, we need to adjust the pointer to get our full connection_context.

Sending is easy too:

void begin_send(struct connection_context *ctx)
{
  if(ctx->sendbuf.buf)
  {
    // only begin a send if one
    // isn't already running.
    return;
  }

  if(!ctx->recvbuf.len)
  {
    // only begin a send if the
    // read buffer has something.
    return;
  }

  // switch buffers.
  ctx->sendbuf = ctx->readbuf;

  // clear read buffer.
  ctx->readbuf.buf = NULL;
  ctx->readbuf.len = 0;

  // zero out io_context structure.
  memset(&ctx->sendctx, 0, sizeof(ctx->sendctx));

  // set completion callback.
  ctx->sendctx.on_completion = send_finished;

  WSASend(ctx->sock, &ctx->sendbuf, 1, NULL, 0,
          &ctx->sendctx.ovl, NULL);

  // start reading next buffer.
  begin_read(ctx);
}

void send_finished(DWORD error, DWORD transferred,
                   struct io_context *ioctx)
{
  // get our connection context.
  struct connection_context *ctx =
     (struct connection_context*)((char*)ioctx -
        offsetof(struct connection_context, sendctx));

  if(error != ERROR_SUCCESS)
  {
    // handle error.
    return;
  }

  // success, clear send buffer and start next send.

  free(ctx->sendbuf.buf);
  ctx->sendbuf.buf = NULL;

  begin_send(ctx);
}

Pretty much more of the same. Again for brevity I’m leaving out some error checking code and assuming the buffer gets sent out in full. I’m also assuming a single-threaded design — socket and file functions themselves are thread-safe and have nothing to worry about, but the buffer management code here would need some extra locking since it could be run concurrently. But the idea should be clear.

Update: this subject continued in Tips for efficient I/O.

I/O completion ports were first introduced in Windows NT 4.0 as a highly scalable, multi-processor capable alternative to the then-typical practices of using select, WSAWaitForMultipleEvents, WSAAsyncSelect, or even running a single thread per client. The most optimal way to perform I/O on Windows — short of writing a kernel-mode driver — is to use I/O completion ports.

A recent post on Slashdot claimed sockets have run their course, which I think is absolutely not true! The author seems to believe the Berkeley sockets API is the only way to perform socket I/O, which is nonsense. Much more modern, scalable and high performance APIs exist today such as I/O completion ports on Windows, epoll on Linux, and kqueue on FreeBSD. In light of this I thought I’d write a little about completion ports here.

The old ways of multiplexing I/O still work pretty well for scenarios with a low number of concurrent connections, but when writing a server daemon to handle a thousand or even tens of thousands of concurrent clients at once, we need to use something different. In this sense the old de facto standard Berkeley sockets API has run its course, because the overhead of managing so many connections is simply too great and makes using multiple processors hard.

An I/O completion port is a multi-processor aware queue. You create a completion port, bind file or socket handles to it, and start asynchronous I/O operations. When they complete — either successfully or with an error — a completion packet is queued up on the completion port, which the application can dequeue from multiple threads. The completion port uses some special voodoo to make sure only a specific number of threads can run at once — if one thread blocks in kernel-mode, it will automatically start up another one.

First you need to create a completion port with CreateIoCompletionPort:

HANDLE iocp = CreateIoCompletionPort(INVALID_HANDLE_VALUE,
   NULL, 0, 0);

It’s important to note that NumberOfConcurrentThreads is not the total number of threads that can access the completion port — you can have as many as you want. This instead controls the number of threads it will allow to run concurrently. Once you’ve reached this number, it will block all other threads. If one of the running threads blocks for any reason in kernel-mode, the completion port will automatically start up one of the waiting threads. Specifying 0 for this is equivalent to the number of logical processors in the system.

Next is creating and associating a file or socket handle, using CreateFile, WSASocket, and CreateIoCompletionPort:

#define OPERATION_KEY 1

HANDLE file = CreateFile(L"file.txt", GENERIC_READ,
   FILE_SHARE_READ, NULL, OPEN_EXISTING,
   FILE_FLAG_OVERLAPPED, NULL);

SOCKET sock = WSASocket(AF_INET, SOCK_STREAM, IPPROTO_TCP,
   NULL, 0, WSA_FLAG_OVERLAPPED);

CreateIoCompletionPort(file, iocp, OPERATION_KEY, 0);
CreateIoCompletionPort((HANDLE)sock, iocp, OPERATION_KEY, 0);

Files and sockets must be opened with the FILE_FLAG_OVERLAPPED and WSA_FLAG_OVERLAPPED flags before they are used asynchronously.

Reusing CreateIoCompletionPort for associating file/socket handles is weird design choice from Microsoft but that’s how it’s done. The CompletionKey parameter can be anything you want, it is a value provided when packets are dequeued. I define a OPERATION_KEY here to use as the CompletionKey, the significance of which I’ll get to later.

Next we have to start up some I/O operations. I’ll skip setting up the socket and go right to sending data. We start these operations using ReadFile and WSASend:

OVERLAPPED readop, sendop;
WSABUF sendwbuf;
char readbuf[256], sendbuf[256];

memset(&readop, 0, sizeof(readop));
memset(&sendop, 0, sizeof(sendop));

sendwbuf.len = sizeof(sendbuf);
sendwbuf.buf = sendbuf;

BOOL readstatus = ReadFile(file, readbuf,
   sizeof(readbuf), NULL, &readop);

if(!readstatus)
{
  DWORD readerr = GetLastError();

  if(readerr != ERROR_IO_PENDING)
  {
    // error reading.
  }
}

int writestatus = WSASend(sock, &sendwbuf, 1, NULL,
   0, &sendop, NULL);

if(writestatus)
{
  int writeerr = WSAGetLastError();

  if(writeerr != WSA_IO_PENDING)
  {
    // error sending.
  }
}

Every I/O operation using a completion port has an OVERLAPPED structure associated with it. Windows uses this internally in unspecified ways, only saying we need to zero them out before starting an operation. The OVERLAPPED structure will be given back to us when we dequeue the completion packets, and can be used to pass along our own context data.

I have left out the standard error checking up until now for brevity’s sake, but this one doesn’t work quite like one might expect so it is important here. When starting an I/O operation, an error might not really be an error. If the function succeeds all is well, but if the function fails, it is important to check the error code with GetLastError or WSAGetLastError.

If these functions return ERROR_IO_PENDING or WSA_IO_PENDING, the function actually still completed successfully. All these error codes mean is an asynchronous operation has been started and completion is pending, as opposed to completing immediately. A completion packet will be queued up regardless of the operation completing asynchronously or not.

Dequeuing packets from a completion port is handled by the GetQueuedCompletionStatus function:

OVERLAPPED *ovl;
ULONG_PTR completionkey;
DWORD transferred;

BOOL ret = GetQueuedCompletionStatus(iocp, &transferred,
   &completionkey, &ovl, INFINITE);

if(ret)
{
  // I/O completed successfully.
}
else if(ovl)
{
  // dequeued successfully but the I/O operation
  // failed, get extended information.
  DWORD err = GetLastError();
}
else
{
  // error dequeuing a packet.
}

This function dequeues completion packets, consisting of the completion key we specified in CreateIoCompletionPort and the OVERLAPPED structure we gave while starting the I/O. If the I/O transferred any data, it will retrieve how many bytes were transferred too. Again the error handling is a bit weird on this one, having three error states.

This can be run from as many threads as you like, even a single one. It is common practice to run a pool of twice the number of threads as there are logical processors available, to keep the CPU active with the aforementioned functionality of starting a new thread if a running one blocks.

Unless you are going for a single-threaded design, I recommend starting two threads per logical CPU. Even if your app is designed to be 100% asynchronous, you will still run into blocking when locking shared data and even get unavoidable hidden blocking I/Os like reading in paged out memory. Keeping two threads per logical CPU will keep the processor busy without overloading the OS with too much context switching.

This is all well and good, but two I/O operations were initiated — a file read and a socket send. We need a way to tell their completion packets apart. This is why we need to attach some state to the OVERLAPPED structure when we call those functions:

struct my_context
{
  OVERLAPPED ovl;
  int isread;
};

struct my_context readop, sendop;

memset(&readop.ovl, 0, sizeof(readop.ovl));
memset(&sendop.ovl, 0, sizeof(sendop.ovl));

readop.isread = 1;
sendop.isread = 0;

ReadFile(file, readbuf, sizeof(readbuf), NULL, &readop.ovl);
WSASend(sock, &sendwbuf, 1, NULL, 0, &sendop.ovl, NULL);

Now we can tell the operations apart when we dequeue them:

OVERLAPPED *ovl;
ULONG_PTR completionkey;
DWORD transferred;

GetQueuedCompletionStatus(iocp, &transferred,
   &completionkey, &ovl, INFINITE);

struct my_context *ctx = (struct my_context*)ovl;

if(ctx->isread)
{
  // read completed.
}
else
{
  // send completed.
}

The last important thing to know is how to queue up your own completion packets. This is useful if you want to split an operation up to be run on the thread pool, or if you want to exit a thread waiting on a call to GetQueuedCompletionStatus. To do this, we use the PostQueuedCompletionStatus function:

#define EXIT_KEY 0

struct my_context ctx;

PostQueuedCompletionStatus(iocp, 0, OPERATION_KEY, &ctx.ovl);
PostQueuedCompletionStatus(iocp, 0, EXIT_KEY, NULL);

Here we post two things onto the queue. The first, we post our own structure onto the queue, to be processed by our thread pool. The second, we give a new completion key: EXIT_KEY. The thread which processes this packet can test if the completion key is EXIT_KEY to know when it needs to stop dequeuing packets and shut down.

Other than the completion port handle, Windows does not use any of the parameters given to PostQueuedCompletionStatus. They are entirely for our use, to be dequeued with GetQueuedCompletionStatus.

That’s all I have to write for now, and should be everything one would need to get started learning these high performance APIs! I will make another post shortly detailing some good patterns for completion port usage, and some optimization tips to ensure efficient usage of these I/O APIs.

Update: this subject continued in I/O completion ports made easy.

A couple weeks before the Windows 7 RC came out, I formatted back to Vista so I could test the upgrade process. I found myself missing various smaller features that made Win7 so much nicer to use. Even the simple feature of moving the “show desktop” feature to the bottom right of the taskbar instead of as a quick launch shortcut.

Well, a little under a week ago the RC was released to testers and I’ve been pretty happy with it. Some features were added but if you’ve used the beta you probably won’t notice many significant changes — it is mostly bug fixes and optimizations. I’ve not only been running it on my desktop, but on my Eee PC where it has been performing quite well with no tweaks.

Today the RC was released for anyone to download, and will be available for the next two months. Developers can also download the SDK to get a head start on writing applications for it.

I’m happy that Windows Media Center, perhaps the most problematic portion of the beta, has got the polish it needed. It is much more stable and finally looks like something other than the TV Pack that was kind of released for Vista. One of the biggest feature complaints was recording to an incompatible “.wtv” format, which has been somewhat alleviated by a “convert to .dvr-ms” option which is enabled for non-DRMed .wtv recordings.

Don’t use threads. Or more precisely, don’t over-use them. It’s one of the first thing fledgling programmers learn after they start using threads. This is because threading involves a lot of overhead. In short, using more threads may improve concurrency, but it will give you less overall throughput as more processing is put into simply managing the threads instead of letting them run. So programmers learn to use threads sparingly.

When normal threads run out of time, or block on something like a mutex or I/O, they hand off control to the operating system kernel. The kernel then finds a new thread to run, and switches back to user-mode to run the thread. This context switching is what User Mode Scheduling looks to alleviate.

User Mode Scheduling can be thought of as a cross between threads and thread pools. An application creates one or more UMS scheduler threads — typically one for each processor. It then creates several UMS worker threads for each scheduler thread. The worker threads are the ones that run your actual code. Whenever a worker thread runs out of time, it is put on the end of its scheduler thread’s queue. If a worker thread blocks, it is put on a waiting list to be re-queued by the kernel when whatever it was waiting on finishes. The scheduler thread then takes the worker thread from the top of the queue and starts running it. Like the name suggests, this happens entirely in user-mode, avoiding the expensive user->kernel->user-mode transitions. Letting each thread run for exactly as long as it needs helps to solve the throughput problem. Work is only put into managing threads when absolutely necessary instead of in ever smaller time slices, leaving more time to run your actual code.

A good side effect of this is UMS threads also help to alleviate the cache thrashing problems typical in heavily-threaded applications. Forgetting your data sharing patterns, each thread still needs its own storage for stack space, processor context, and thread-local storage. Every time a context switch happens, some data may need to be pushed out of caches in order to load some kernel-mode code and the next thread’s data. By switching between threads less often, cache can be put to better use for the task at hand.

If you have ever had a chance to use some of the more esoteric APIs included with Windows, you might be wondering why we need UMS threads when we have fibers which offer similar co-operative multitasking. Fibers have a lot of special exceptions. There are things that aren’t safe to do with them. Libraries that rely on thread-local storage, for instance, will likely walk all over themselves if used from within fibers. A UMS thread on the other hand is a full fledged thread — they support TLS and no have no real special things to keep in mind while using them.

I still wouldn’t count out thread pools just yet. UMS threads are still more expensive than a thread pool and the large memory requirements of a thread still apply here, so things like per-client threads in internet daemons are still out of the question if you want to be massively scalable. More likely, UMS threads will be most useful for building thread pools. Most thread pools launch two or three threads per CPU to help stay busy when given blocking tasks, and UMS threads will at least help keep their time slice usage optimal.

From what I understand the team behind Microsoft’s Concurrency Runtime, to be included with Visual C++ 2010, was one of the primary forces behind UMS threads. They worked very closely with the kernel folks to find the most scalable way to enable the super-parallel code that will be possible with the CR.

I was working on something completely different last night, when an elegant idea came to mind on how to significantly speed up PeerGuardian’s IP searching. It’s funny how an idea can just pop into the mind about a problem that hasn’t been thought of in a long time.

Right now PeerGuardian uses a binary search to match IPs. This is already pretty efficient, running in ⌈log₂ N⌉ — so for 100,000 IP ranges, about 16 tests need to be done. This has the additional advantage of having no memory overhead.

My idea is to use a structure similar to a B+tree, packing as many IPs and branch pointers into a cache line as possible. On today’s architectures, a cache line is typically 64 bytes, so 8 IPs and 9 branch pointers would fit on each node, making it only need to read about ⌈log₉ N⌉ nodes to find a match. So in order to find a match in 100,000 IP ranges, only about 5 nodes would need to be read.

CPUs always read and cache data in blocks (a cache line), so an algorithm that keeps this in mind to minimize memory reads and maximize cache usage should be incredibly fast. Even though this introduces significant overhead for branch pointers (about 2x the storage would be required), it should be far more efficient overall.

But this algorithm improves in another way too: branching. I’m talking in terms of branch instructions, not the branch pointers mentioned above. The fewer branches code takes, the faster a superscalar or pipelined CPU will be able to run your code. For this algorithm, an entire node could be processed (that is, comparing the IPs and determining which branch node to go into) with zero branches using integer SSE2 (PCMPGTD, PMOVMSKB), and bit-scan forward (BSF).

I can’t be sure how much of a speed difference this would make until I code it up, but I bet it would be at least 200% faster. I’ve been too busy to work on PeerGuardian for quite a while, so I don’t know if this will ever make it into PG. We’re looking for a new coder with more time on their hands.