westonpace commented on code in PR #13640:
URL: https://github.com/apache/arrow/pull/13640#discussion_r924958407
##########
cpp/src/arrow/io/file.cc:
##########
@@ -378,6 +378,77 @@ Status FileOutputStream::Write(const void* data, int64_t
length) {
int FileOutputStream::file_descriptor() const { return impl_->fd(); }
+// ----------------------------------------------------------------------
+// DirectFileOutputStream, change the Open, Write and Close methods from
FileOutputStream
+// Uses DirectIO for writes. Will only write out things in 4096 byte blocks.
Buffers leftover bytes
+// in an internal data structure, which will be padded to 4096 bytes and
flushed upon call to close.
+
+class DirectFileOutputStream::DirectFileOutputStreamImpl : public OSFile {
+ public:
+ Status Open(const std::string& path, bool append) {
+ const bool truncate = !append;
+ return OpenWritable(path, truncate, append, true /* write_only */, true);
+ }
+ Status Open(int fd) { return OpenWritable(fd); }
+};
+
+DirectFileOutputStream::DirectFileOutputStream() {
+ uintptr_t mask = (uintptr_t)(4095);
+ uint8_t *mem = static_cast<uint8_t *>(malloc(4096 + 4095));
+ cached_data = reinterpret_cast<uint8_t *>(
reinterpret_cast<uintptr_t>(mem+4095) & ~(mask));
Review Comment:
I ran the experiments myself with 1GB of data for timing. Oddly, on an HHD,
I didn't observe much difference. On an SSD, the difference was rather
considerable (I reran the ssd tests a number of times and the difference was
consistent):
Method | Disk | Time (ms)
--- | --- | ---
direct | hhd | 41290.55257
direct | ssd | 4641.38446
fadvise1 | hhd | 41179.51826
fadvise1 | ssd | 6959.85842
The direct I/O results roughly match what I get for a disk benchmark on the
SSD. However, I feel the test is still cheating. The `direct.cpp` example is
assuming the memory is previously aligned. Can you create a version that does
not assume this (e.g. memcpy into a temp buffer before writing)?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
