Pieter Wuille [ARCHIVE] on Nostr: 📅 Original date posted:2012-02-02 📝 Original message:> You will also find the ...
📅 Original date posted:2012-02-02
📝 Original message:> You will also find the RPC server in libcoin blistering fast compared to
the Satoshi client. (It was actually what got me to write libcoin in the
first place...). The Satoshi client HTTP server executes all rpc commands
in its own thread, but to do so, it needs to stop the thread of the Node,
even though the command executed is just a query (i.e. not a SendTo), you
hence have two threads blocking each other and when they wait, you wait...
In libcoin all the query methods access the blockChain as a const object
and they can hence safely query it without intervening the work of the Node
thread. The exception are the SendTo methods that first query if a
transaction can take place, then pushes it to the work-queue of the Node
thread and again exits immediately. The actual execution then follows once
the Node has finished its current tasks (e.g. validating a block).
Hello Michael,
I'm impressed by your refactorings, and hope some of them can make it into
the Satoshi codebase. I am however not sure what you've said above is safe.
In particular, how do you guarantee that no other thread modifies the
blockchain structure while you are performing your query on it? Does the
query code operate on a const copy of the structure, or is there guaranteed
only one thread accessing it?
I've been thinking about moving to read-write locks that allow multiple
threads reading the datastructure simultaneously, but removing the locking
all together sounds wrong to me.
--
Pieter
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linuxfoundation.org/pipermail/bitcoin-dev/attachments/20120203/270a6572/attachment.html>;
📝 Original message:> You will also find the RPC server in libcoin blistering fast compared to
the Satoshi client. (It was actually what got me to write libcoin in the
first place...). The Satoshi client HTTP server executes all rpc commands
in its own thread, but to do so, it needs to stop the thread of the Node,
even though the command executed is just a query (i.e. not a SendTo), you
hence have two threads blocking each other and when they wait, you wait...
In libcoin all the query methods access the blockChain as a const object
and they can hence safely query it without intervening the work of the Node
thread. The exception are the SendTo methods that first query if a
transaction can take place, then pushes it to the work-queue of the Node
thread and again exits immediately. The actual execution then follows once
the Node has finished its current tasks (e.g. validating a block).
Hello Michael,
I'm impressed by your refactorings, and hope some of them can make it into
the Satoshi codebase. I am however not sure what you've said above is safe.
In particular, how do you guarantee that no other thread modifies the
blockchain structure while you are performing your query on it? Does the
query code operate on a const copy of the structure, or is there guaranteed
only one thread accessing it?
I've been thinking about moving to read-write locks that allow multiple
threads reading the datastructure simultaneously, but removing the locking
all together sounds wrong to me.
--
Pieter
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linuxfoundation.org/pipermail/bitcoin-dev/attachments/20120203/270a6572/attachment.html>;