Re: [racket-users] Racket v7.3

2019-05-16 Thread Matthew Flatt
At Thu, 16 May 2019 15:28:10 +0100, Laurent wrote:
> On Wed, May 15, 2019 at 7:05 PM 'John Clements' via Racket Users <
> racket-users@googlegroups.com> wrote:
> > - Racket's IO system has been refactored to improve performance
> >  and simplify internal design.
> >
> 
> Is it mild overall performance improvement, or are there cases where there
> can be a big boost? (Like "previous bottlenecks have become highways" kind
> of improvement)

Number parsing in `read` or `string->number` should be faster, but the
main I/O refactoring and improvements were specific to Racket-on-Chez.

In the long run, this new I/O implementation might replace the current
one in the traditional Racket implementation, but I don't know how much
of a performance different it would be.

(I recommended dropping this bullet from the announcement, since the
main changes were mainly for Racket-on-Chez. But I guess that
recommendation was too buried or got lost.)

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/racket-users/5cde00d0.1c69fb81.5743b.16ffSMTPIN_ADDED_MISSING%40gmr-mx.google.com.
For more options, visit https://groups.google.com/d/optout.


Re: [racket-users] Racket v7.3

2019-05-16 Thread Dexter Lagan
  I’ll compare file write times as soon as we’re done testing the tool I’m 
working on.

Cheers

Dex

> On May 16, 2019, at 10:03 PM, Piyush Katariya  
> wrote:
> 
> I am also keen to know the IO perf optimization results in real world.
> 
> Dexter, if possible can you give us some approximate throughput or latency 
> numbers when you switch to v7.3 
> -- 
> You received this message because you are subscribed to the Google Groups 
> "Racket Users" group.
> To unsubscribe from this group and stop receiving emails from it, send an 
> email to racket-users+unsubscr...@googlegroups.com.
> To view this discussion on the web visit 
> https://groups.google.com/d/msgid/racket-users/a2edb6b8-e792-42be-bfc9-63e78ef38e95%40googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/racket-users/3D57811B-57E4-48CB-AAF9-1A307F70E888%40gmail.com.
For more options, visit https://groups.google.com/d/optout.


Re: [racket-users] Racket v7.3

2019-05-16 Thread Piyush Katariya
I am also keen to know the IO perf optimization results in real world.

Dexter, if possible can you give us some approximate throughput or latency 
numbers when you switch to v7.3 

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/racket-users/a2edb6b8-e792-42be-bfc9-63e78ef38e95%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.


RE: [racket-users] Racket v7.3

2019-05-16 Thread Dexter Lagan
A huge thanks to everybody who worked on Racket. 

 

  IO improvements will surely make everybody’s life a little better in my team 
and the whole studio. We’re relying on large (200MB text files), and numerous 
(in the 100,000’s) file reads and writes daily. We’re converting shell scripts 
to Racket and I recon it’s hard to beat some GNU command’s optimizations. I 
have been tempted to try Rust but a) it’s not guaranteed to perform better for 
pure IO, and b) I’d rather code it directly in macro-asm if it meant better 
speed. But so far, Racket’s performance has been more than adequate for our 
tasks, and thanks to this I can do it all in Racket. That puts a big, gleeful 
smile on my face.

 

From: racket-users@googlegroups.com  On Behalf 
Of Laurent
Sent: 16 May, 2019 3:28 PM
To: John Clements 
Cc: Racket Users 
Subject: Re: [racket-users] Racket v7.3.

 

Thank you all for the new release!

 

On Wed, May 15, 2019 at 7:05 PM 'John Clements' via Racket Users 
mailto:racket-users@googlegroups.com> > wrote:

- Racket's IO system has been refactored to improve performance
 and simplify internal design.

 

Is it mild overall performance improvement, or are there cases where there can 
be a big boost? (Like "previous bottlenecks have become highways" kind of 
improvement)

 

 

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com 
<mailto:racket-users+unsubscr...@googlegroups.com> .
To view this discussion on the web visit 
https://groups.google.com/d/msgid/racket-users/CABNTSaEotkhQ_yK8n60fkFgkOmULX061eoj-LmDfhSBopA4h1Q%40mail.gmail.com
 
<https://groups.google.com/d/msgid/racket-users/CABNTSaEotkhQ_yK8n60fkFgkOmULX061eoj-LmDfhSBopA4h1Q%40mail.gmail.com?utm_medium=email_source=footer>
 .
For more options, visit https://groups.google.com/d/optout.

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/racket-users/011301d50c07%244f52f9c0%24edf8ed40%24%40gmail.com.
For more options, visit https://groups.google.com/d/optout.


Re: [racket-users] Racket v7.3

2019-05-16 Thread Laurent
Thank you all for the new release!

On Wed, May 15, 2019 at 7:05 PM 'John Clements' via Racket Users <
racket-users@googlegroups.com> wrote:

> - Racket's IO system has been refactored to improve performance
>  and simplify internal design.
>

Is it mild overall performance improvement, or are there cases where there
can be a big boost? (Like "previous bottlenecks have become highways" kind
of improvement)

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/racket-users/CABNTSaEotkhQ_yK8n60fkFgkOmULX061eoj-LmDfhSBopA4h1Q%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.