My problem is in regards file copying performance. We have a media management system that requires a lot of moving files around on the file system to different locations including windows shares on the same network, FTP sites, AmazonS3, etc. When we were all on one windows network we could get away with using System.IO.File.Copy(source, destination) to copy a file. Since many times all we have is an input Stream (like a MemoryStream), we tried abstracting the Copy operation to take an input Stream and an output Stream but we are seeing a massive performance decrease. Below is some code for copying a file to use as a discussion point.
public void Copy(System.IO.Stream inStream, string outputFilePath)
{
int bufferSize = 1024 * 64;
using (FileStream fileStream = new FileStream(outputFilePath, FileMode.OpenOrCreate, FileAccess.Write))
{
int bytesRead = -1;
byte[] bytes = new byte[bufferSize];
while ((bytesRead = inStream.Read(bytes, 0, bufferSize)) > 0)
{
fileStream.Write(bytes, 0, bytesRead);
fileStream.Flush();
}
}
}
Does anyone know why this performs so much slower than File.Copy? Is there anything I can do to improve performance? Am I just going to have to put special logic in to see if I'm copying from one windows location to another--in which case I would just use File.Copy and in the other cases I'll use the streams?
Please let me know what you think and whether you need additional information. I have tried different buffer sizes and it seems like a 64k buffer size is optimal for our "small" files and 256k+ is a better buffer size for our "large" files--but in either case it performs much worse than File.Copy(). Thanks in advance!
File.Copy was build around CopyFile Win32 function and this function takes lot of attention from MS crew (remember this Vista-related threads about slow copy performance).
Several clues to improve performance of your method:
Like many said earlier remove Flush method from your cycle. You do not need it at all.
Increasing buffer may help, but only on file-to-file operations, for network shares, or ftp servers this will slow down instead. 60 * 1024 is ideal for network shares, at least before vista. for ftp 32k will be enough in most cases.
Help os by providing your caching strategy (in your case sequential reading and writing), use FileStream constructor override with FileOptions parameter (SequentalScan).
You can speed up copying by using asynchronous pattern (especially useful for network-to-file cases), but do not use threads for this, instead use overlapped io (BeginRead, EndRead, BeginWrite, EndWrite in .net), and do not forget set Asynchronous option in FileStream constructor (see FileOptions)
Example of asynchronous copy pattern:
int Readed = 0;
IAsyncResult ReadResult;
IAsyncResult WriteResult;
ReadResult = sourceStream.BeginRead(ActiveBuffer, 0, ActiveBuffer.Length, null, null);
do
{
Readed = sourceStream.EndRead(ReadResult);
WriteResult = destStream.BeginWrite(ActiveBuffer, 0, Readed, null, null);
WriteBuffer = ActiveBuffer;
if (Readed > 0)
{
ReadResult = sourceStream.BeginRead(BackBuffer, 0, BackBuffer.Length, null, null);
BackBuffer = Interlocked.Exchange(ref ActiveBuffer, BackBuffer);
}
destStream.EndWrite(WriteResult);
}
while (Readed > 0);
Three changes will dramatically improve performance:
Increase your buffer size, try 1MB (well -just experiment)
After you open your fileStream, call fileStream.SetLength(inStream.Length) to allocate the entire block on disk up front (only works if inStream is seekable)
Remove fileStream.Flush() - it is redundant and probably has the single biggest impact on performance as it will block until the flush is complete. The stream will be flushed anyway on dispose.
This seemed about 3-4 times faster in the experiments I tried:
public static void Copy(System.IO.Stream inStream, string outputFilePath)
{
int bufferSize = 1024 * 1024;
using (FileStream fileStream = new FileStream(outputFilePath, FileMode.OpenOrCreate, FileAccess.Write))
{
fileStream.SetLength(inStream.Length);
int bytesRead = -1;
byte[] bytes = new byte[bufferSize];
while ((bytesRead = inStream.Read(bytes, 0, bufferSize)) > 0)
{
fileStream.Write(bytes, 0, bytesRead);
}
}
}
Dusting off reflector we can see that File.Copy actually calls the Win32 API:
if (!Win32Native.CopyFile(fullPathInternal, dst, !overwrite))
Which resolves to
[DllImport("kernel32.dll", CharSet=CharSet.Auto, SetLastError=true)]
internal static extern bool CopyFile(string src, string dst, bool failIfExists);
And here is the documentation for CopyFile
You'll never going to able to beat the operating system at doing something so fundemental with your own code, not even if you crafted it carefully in assembler.
If you need make sure that your operations occur with the best performance AND you want to mix and match various sources then you will need to create a type that describes the resource locations. You then create an API that has functions such as Copy that takes two such types and having examined the descriptions of both chooses the best performing copy mechanism. E.g., having determined that both locations are windows file locations you it would choose File.Copy OR if the source is windows file but the destination is to be HTTP POST it uses a WebRequest.
Try to remove the Flush call, and move it to be outside the loop.
Sometimes the OS knows best when to flush the IO.. It allows it to better use its internal buffers.
Here's a similar answer
How do I copy the contents of one stream to another?
Your main problem is the call to Flush(), that will bind your performance to the speed of the I/O.
Mark Russinovich would be the authority on this.
He wrote on his blog an entry Inside Vista SP1 File Copy Improvements which sums up the Windows state of the art through Vista SP1.
My semi-educated guess would be that File.Copy would be most robust over the greatest number of situations. Of course, that doesn't mean in some specific corner case, your own code might beat it...
One thing that stands out is that you are reading a chunk, writing that chunk, reading another chunk and so on.
Streaming operations are great candidates for multithreading. My guess is that File.Copy implements multithreading.
Try reading in one thread and writing in another thread. You will need to coordinate the threads so that the write thread doesn't start writing away a buffer until the read thread is done filling it up. You can solve this by having two buffers, one that is being read while the other is being written, and a flag that says which buffer is currently being used for which purpose.
Related
I am reading all files(around 3000 files and size is 50 GB) from specified path with 4k bytes at a time. Below is the code for the same. My query is when i see the CPU and Memory of the application in task manager i could see that the IO Reads are gradually increasing to high level, i can understand that it might be because of 4k read but does that affect to other things or its ok to increase the IO Read. Also is FileStream the optimum way to read the file as it does not load the entire file in memory?
fileStream = new FileStream(filePath, FileMode.Open, FileAccess.Read)
do
{
BytesRead = fileStream.Read(Buffer, 0, MAX_BUFFER);
}
while (BytesRead != 0);
fileStream.Close();
Check Hans Passant's answer about this issue, i find it very clear.
Files are already buffered by the file system cache, You just need to
pick a buffer size that doesn't force FileStream to make the native
Windows ReadFile() API call to fill the buffer too often. Don't go
below a kilobyte, more than 16 KB is a waste of memory.
Take a look at this post too, it provides some benchmarking code.
Occasionally we need to copy huge files from one bucket to another in AWS S3. Whenever possible we use the CopyRequest to handle this operation all on AWS (since no round trip required back to the client). But sometimes we do not have the option to do this because we need to copy between 2 completely separate accounts which requires a GET and then a PUT.
Problems:
The response stream returned from the GET is not seekable so it cannot be passed to the PUT request and have it stream seamlessly from one to the other
Copying the response stream to an intermediary stream (MemoryStream) using CopyTo() and then passing that to the PUT operation works well but doesn't scale (large files will throw OutOfMemory exceptions)
So basically I need an intermediary stream that I can read/write to at the same time, basically I would read a chunk from the response stream and write it to my intermediary stream, meanwhile the PUT request is reading out the content and its just a seamless pass-thru sort of scenario.
I found this post on stackoverflow and it seemed promising at first but it still throws an OutOfMemory exception with large files.
.NET Asynchronous stream read/write
Anyone ever had to do something similar to this? How would you tackle it? Thanks in advcance
It's not clear why you would want to use MemoryStream. The Stream.CopyTo method in .NET 4 doesn't need to use an intermediate stream - it will just read into a local buffer of a fixed size, then write that buffer to the output stream, then read more data (overwriting the buffer) etc.
If you're not using .NET 4, it's easy to implement something similar, e.g.
public static void CopyTo(this Stream input, Stream output)
{
byte[] buffer = new byte[64 * 1024]; // 64K buffer
int bytesRead;
while ((bytesRead = input.Read(buffer, 0, buffer.Length)) > 0)
{
output.Write(buffer, 0, bytesRead);
}
}
I found this, but it uses a Queue internally, which the author notes is an order of magnitude slower than a MemoryStream.
http://www.codeproject.com/Articles/16011/PipeStream-a-Memory-Efficient-and-Thread-Safe-Stre
I keep hoping I'll find an official MS library solution, but it seems that this wheel hasn't been properly invented yet.
everyone,i have a lot of files write to disk per seconds,i want to disable disk cache to improve performance,i google search find a solution:win32 CreateFile method with FILE_FLAG_NO_BUFFERING and How to empty/flush Windows READ disk cache in C#?.
i write a little of code to test whether can worked:
const int FILE_FLAG_NO_BUFFERING = unchecked((int)0x20000000);
[DllImport("KERNEL32", SetLastError = true, CharSet = CharSet.Auto, BestFitMapping = false)]
static extern SafeFileHandle CreateFile(
String fileName,
int desiredAccess,
System.IO.FileShare shareMode,
IntPtr securityAttrs,
System.IO.FileMode creationDisposition,
int flagsAndAttributes,
IntPtr templateFile);
static void Main(string[] args)
{
var handler = CreateFile(#"d:\temp.bin", (int)FileAccess.Write, FileShare.None,IntPtr.Zero, FileMode.Create, FILE_FLAG_NO_BUFFERING, IntPtr.Zero);
var stream = new FileStream(handler, FileAccess.Write, BlockSize);//BlockSize=4096
byte[] array = Encoding.UTF8.GetBytes("hello,world");
stream.Write(array, 0, array.Length);
stream.Close();
}
when running this program,the application get exception:IO operation will not work. Most likely the file will become too long or the handle was not opened to support synchronous IO operations
later,i found this article When you create an object with constraints, you have to make sure everybody who uses the object understands those constraints,but i can't fully understand,so i change my code to test:
var stream = new FileStream(handler, FileAccess.Write, 4096);
byte[] ioBuffer = new byte[4096];
byte[] array = Encoding.UTF8.GetBytes("hello,world");
Array.Copy(array, ioBuffer, array.Length);
stream.Write(ioBuffer, 0, ioBuffer.Length);
stream.Close();
it's running ok,but i just want "hello,world" bytes not all.i trying change blocksize to 1 or other integer(not 512 multiple) get same error.i also try win32 WriteFile api also get same error.someone can help me?
CreateFile() function in No Buffering mode imposes strict requirements on what may and what may not be done. Having a buffer of certain size (multiple of device sector size) is one of them.
Now, you can improve file writes in this way only if you use buffering in your code. If you want to write 10 bytes without buffering, then No Buffering mode won't help you.
If I understood your requirements correctly, this is what I'd try first:
Create a queue with objects that have the data in memory and the target file on the disk.
You start writing the files first just into memory, and then on another thread start going through the queue, opening io-completion port based filestream handles (isAsync=True) - just don't open too many of them as at some point you'll probably start losing perf due to cache trashing etc. You need to profile to see what is optimal amount for your system and ssd's.
After each open, you can use the async filestream methods Begin... to start writing data from memory to the files. the isAsync puts some requirements so this may not be as easy to get working in every corner case as using filestream normally.
Whether there will be any improvement to using another thread to create the files and another to write to them using the async api, that might only be the case if there is a possibility that creating/opening the files would block. SSD's perform various things internally to keep the access to data fast, so when you start doing this sort of extreme performance stuff, there may be pronounced differences between SSD controllers. It's also possible that if the controller drivers aren't well implemented, OS/Windows may start to feel sluggish or freeze. The hardware benchmarks sites do not really stress this particular kind of scenario (eg. create and write x KB into million files asap) and no doubt there's some drivers out there that are slower than others.
I'm trying to write a class that will copy a file from one location to another and report progress. The problem that I'm having is that when the application is run, the progress will shoot from 0 to 100% instantly, but the file is still copying in the background.
public void Copy(string sourceFile, string destinationFile)
{
_stopWatch.Start();
_sourceStream = new FileStream(srcName, FileMode.Open);
_destinationStream = new FileStream(destName, FileMode.CreateNew);
read();
//On a 500mb file, execution will reach here in about a second.
}
private void read()
{
int i = _sourceStream.Read(_buffer, 0, bufferSize);
_completedBytes += i;
if (i != 0)
{
_destinationStream.Write(_buffer, 0, i);
TriggerProgressUpdate();
read();
}
}
private void TriggerProgressUpdate()
{
if (OnCopyProgress != null)
{
CopyProgressEventArgs arg = new CopyProgressEventArgs();
arg.CompleteBytes = _completedBytes;
if (_totalBytes == 0)
_totalBytes = new FileInfo(srcName).Length;
arg.TotalBytes = _totalBytes;
OnCopyProgress(this, arg);
}
}
What seems to be happening is that FileStream is merely queuing the operations in the OS, instead of blocking until the read or write is complete.
Is there any way to disable this functionality without causing a huge performance loss?
PS. I am using test source and destination variables, thats why they dont match the arguments.
Thanks
Craig
I don't think it can be queuing the read operations... after all, you've got a byte array, it will have some data in after the Read call - that data had better be correct. It's probably only the write operations which are being buffered.
You could try calling Flush on the output stream periodically... I don't know quite how far the Flush will go in terms of the various levels of caching, but it may well wait until the data has actually been written. EDIT: If you know it's a FileStream, you can call Flush(true) which will wait until the data has actually been written to disk.
Note that you shouldn't do this too often, or performance will suffer significantly. You'll need to balance the granularity of progress accuracy with the performance penalty for taking more control instead of letting the OS optimize the disk access.
I'm concerned about your use of recursion here - on a very large file you may well blow up with a stack overflow for no good reason. (The CLR can sometimes optimize tail-recursive methods, but not always). I suggest you use a loop instead. That would also be more readable, IMO:
public void Copy()
{
int bytesRead;
while ((bytesRead = _sourceStream.Read(_buffer, 0, _buffer.Length)) > 0)
{
_destinationStream.Write(_buffer, 0, bytesRead);
_completedBytes += bytesRead;
TriggerProgressUpdate();
if (someAppropriateCondition)
{
_destinationStream.Flush();
}
}
}
I hope you're disposing of the streams somewhere, by the way. Personally I try to avoid having disposable member variables if at all possible. Is there any reason you can't just use local variables in a using statement?
After investigating I found that using "FileOptions.WriteThrough" in a FileStream's constructor will disable write caching. This causes my progress to report correctly. It does however take a performance hit, the copy takes 13 seconds in windows and 20 second in my application. I'm going to try and optimize the code and adjust the buffer size to see if I can speeds things up a bit.
I've been playing around with what I thought was a simple idea. I want to be able to read in a file from somewhere (website, filesystem, ftp), perform some operations on it (compress, encrypt, etc.) and then save it somewhere (somewhere may be a filesystem, ftp, or whatever). It's a basic pipeline design. What I would like to do is to read in the file and put it onto a MemoryStream, then perform the operations on the data in the MemoryStream, and then save that data in the MemoryStream somewhere. I was thinking I could use the same Stream to do this but run into a couple of problems:
Everytime I use a StreamWriter or StreamReader I need to close it and that closes the stream so that I cannot use it anymore. That seems like there must be some way to get around that.
Some of these files may be big and so I may run out of memory if I try to read the whole thing in at once.
I was hoping to be able to spin up each of the steps as separate threads and have the compression step begin as soon as there is data on the stream, and then as soon as the compression has some compressed data available on the stream I could start saving it (for example). Is anything like this easily possible with the C# Streams? ANyone have thoughts as to how to accomplish this best?
Thanks,
Mike
Using a helper method to drive the streaming:
static public void StreamCopy(Stream source, Stream target)
{
byte[] buffer = new byte[8 * 1024];
int size;
do
{
size = source.Read(buffer, 0, 8 * 1024);
target.Write(buffer, 0, size);
} while (size > 0);
}
You can easily combine whatever you need:
using (FileStream iFile = new FileStream(...))
using (FileStream oFile = new FileStream(...))
using (DeflateStream oZip = new DeflateStream(outFile, CompressionMode.Compress))
StreamCopy(iFile, oZip);
Depending on what you are actually trying to do, you'd chain the streams differently. This also uses relatively little memory, because only the data being operated upon is in memory.
StreamReader/StreamWriter shouldn't have been designed to close their underlying stream -- that's a horrible misfeature in the BCL. But they do, they won't be changed (because of backward compatibility), so we're stuck with this disaster of an API.
But there are some well-established workarounds, if you want to use StreamReader/Writer but keep the Stream open afterward.
For a StreamReader: don't Dispose the StreamReader. It's that simple. It's harmless to just let a StreamReader go without ever calling Dispose. The only effect is that your Stream won't get prematurely closed, which is actually a plus.
For a StreamWriter: there may be buffered data, so you can't get away with just letting it go. You have to call Flush, to make sure that buffered data gets written out to the Stream. Then you can just let the StreamWriter go. (Basically, you put a Flush where you normally would have put a Dispose.)
Unless you're reading in streams bigger than your hard drive, I don't think you'll run out of memory:
http://blogs.msdn.com/ericlippert/archive/2009/06/08/out-of-memory-does-not-refer-to-physical-memory.aspx