Thanks again Marc. Yes Deserialization code is as you listed below. I was
using protobuf-net 2.4.0.0 but then moved to older version currently being
used across our code base, 2.0.0.668, to see any impact. Where can I send
you mail? My e-mail is [email protected]. I can get these objects
serialized to a binary file and send over to you.
public T Deserialize(byte[] serializedObject)
{
using (var ms = new MemoryStream(serializedObject))
{
return (T)Serializer.Deserialize(ms, null, typeof(T));
}
}
Thanks much!
Shweta
On Wednesday, March 27, 2019 at 7:34:58 AM UTC-7, Marc Gravell wrote:
>
> > using StackExchange.Redis MGET
>
> Yeah, there's really no way for me to dodge this, is there? ;p
>
> Minor note: your parallel code currently doesn't actually allow any
> meaningful parallelism - you *might* want to move the "lock" so that you
> only "lock" around the "Add". You're also currently doing a "sync over
> async" here - I'd probably use:
>
> var results = db.StringGet(rKeys);
>
> but that is unrelated to the real question here. The
> _serializer.Deserialize(RedisValue) API you're using isn't something that
> is directly exposed on protobuf-net, so I'd be grateful if I could see that
> code. I'm *guessing* it is something like:
>
> using (var ms = new MemoryStream((byte[])value))
> {
> return Serializer.Deserialize<MyObject>(ms);
> }
>
> ?
>
> If that is right, then ... it gets tricky; to give a really meaningful
> analysis, I'd kinda have to see something akin to realistic data. Is there
> anything I can do to setup an indicative minimal repro here? Or is there
> anything you can share with me that I can work with to try and repro what
> you're seeing? My preference here is a randomized data creation script like
> my example, but if the data doesn't contain anything privileged, an RDB
> file (from redis) might work, too.
>
> As I say: I can a range of in-progress changes that might help *anyway*,
> but it would be great if I could actually test things to see how it
> performs. If this allows me to confirm that the proposed changes make a
> good improvement, I can always just ship those! But there may be other
> things I can do here too. The key thing is again, to emphasize: me being
> able to reproduce what you're seeing - because right now, I'm getting ~1ms
> times.
>
> Marc
>
--
You received this message because you are subscribed to the Google Groups
"Protocol Buffers" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at https://groups.google.com/group/protobuf.
For more options, visit https://groups.google.com/d/optout.
