Article: MAQSIP-RT: An HPC Benchmark
By: Mark Roulo (nothanks.delete@this.xxx.com), June 27, 2010 10:21 am
Room: Moderated Discussions
Richard Cownie (tich@pobox.com) on 6/27/10 wrote:
---------------------------
>Even if you don't want the hassle of a fully distributed
>compile, now that you can easily get a server with 2 cpus,
>8 cores, and 16 threads, you should be able to get at
>least a 10x speedup with a parallel compile on a single
>machine. Unless your build system or your file IO is
>horribly screwed up.
The build system/dependencies being screwed up is part of the problem. I think.
>This isn't an argument for C++ - if build time is faster
>with Java, that's good - but I don't think what you see
>is inherent to C++.
Actually, I wasn't arguing that the Java build time was faster (although it was). I was arguing that Vincent's claim that very few developers had long/slow build times was rare was an unsubstantiated claim. I still think it is.
However, I also think that C++ inherently takes longer to build than equivalent Java. Two things *should* contribute to this:
(a) The C++ syntax is more complicated, and thus takes more effort to parse. I can't quantify this myself, but I don't think the difference is trivial.
(b) Because "#include" forces I/O and logically a re-parse of the included files (because how the #include parses has a dependency on what include files came first ... Java imports don't have this behavior), C++ should have more disk I/O (logically more, even if the file system caches in RAM) and more parsing of the same files. With cleverness and pre-compiled headers much of this can be mitigated, although back when I was paying more attention the C++ side didn't trust the pre-compiled headers ... they had been nailed by bugs introduced by them too many times, so they didn't use them.
And a third:
(c) Because of the different object models, the C++ can require a full recompile more often than the Java.
But ... basically, I'm arguing that Vincent's claim that slow compiles only effects a small percentage of developers in unfounded. It might, but I don't think Vincent *knows* this (any more than I do). I'd like to see data, if it exists.
-Mark Roulo
---------------------------
>Even if you don't want the hassle of a fully distributed
>compile, now that you can easily get a server with 2 cpus,
>8 cores, and 16 threads, you should be able to get at
>least a 10x speedup with a parallel compile on a single
>machine. Unless your build system or your file IO is
>horribly screwed up.
The build system/dependencies being screwed up is part of the problem. I think.
>This isn't an argument for C++ - if build time is faster
>with Java, that's good - but I don't think what you see
>is inherent to C++.
Actually, I wasn't arguing that the Java build time was faster (although it was). I was arguing that Vincent's claim that very few developers had long/slow build times was rare was an unsubstantiated claim. I still think it is.
However, I also think that C++ inherently takes longer to build than equivalent Java. Two things *should* contribute to this:
(a) The C++ syntax is more complicated, and thus takes more effort to parse. I can't quantify this myself, but I don't think the difference is trivial.
(b) Because "#include" forces I/O and logically a re-parse of the included files (because how the #include parses has a dependency on what include files came first ... Java imports don't have this behavior), C++ should have more disk I/O (logically more, even if the file system caches in RAM) and more parsing of the same files. With cleverness and pre-compiled headers much of this can be mitigated, although back when I was paying more attention the C++ side didn't trust the pre-compiled headers ... they had been nailed by bugs introduced by them too many times, so they didn't use them.
And a third:
(c) Because of the different object models, the C++ can require a full recompile more often than the Java.
But ... basically, I'm arguing that Vincent's claim that slow compiles only effects a small percentage of developers in unfounded. It might, but I don't think Vincent *knows* this (any more than I do). I'd like to see data, if it exists.
-Mark Roulo
Topic | Posted By | Date |
---|---|---|
New article online: MAQSIP RT | David Kanter | 2010/06/21 11:57 AM |
Why no GCC? | Rohit | 2010/06/22 09:25 PM |
Why no GCC? | David Kanter | 2010/06/23 12:45 AM |
sun 's cc better than GCC? | Rohit | 2010/06/23 05:04 AM |
sun 's cc better than GCC? | anon | 2010/06/23 07:49 AM |
Where is the GCC optimization effort directed? | Mark Roulo | 2010/06/23 10:42 AM |
GCC is very ugly bad everywhere in 64 bits | Vincent Diepeveen | 2010/06/23 02:49 PM |
even for 64-bit arch? | anon | 2010/06/23 02:59 PM |
GCC is very ugly bad everywhere in 64 bits | ajensen | 2010/06/23 11:03 PM |
GCC is very ugly bad everywhere in 64 bits | Gabriele Svelto | 2010/06/24 02:33 AM |
GCC is very ugly bad everywhere in 64 bits | ajensen | 2010/06/24 05:32 AM |
GCC is very ugly bad everywhere in 64 bits | Gabriele Svelto | 2010/06/24 07:18 AM |
GCC is very ugly bad everywhere in 64 bits | ajensen | 2010/06/24 09:50 AM |
Why GCC is big and complicated (my guess) | Mark Roulo | 2010/06/24 12:17 PM |
Why GCC is big and complicated (my guess) | Gabriele Svelto | 2010/06/28 04:00 AM |
GCC is very ugly bad everywhere in 64 bits | Bernd Schmidt | 2010/06/24 05:46 AM |
GCC is very ugly bad everywhere in 64 bits | ajensen | 2010/06/24 09:43 AM |
GCC is very ugly bad everywhere in 64 bits | Vincent Diepeveen | 2010/06/26 02:12 PM |
GCC is very ugly bad everywhere in 64 bits | Rob Thorpe | 2010/06/24 07:47 AM |
GCC is very ugly bad everywhere in 64 bits | Anon | 2010/06/24 05:23 PM |
Where is the GCC optimization effort directed? | Gabriele Svelto | 2010/06/23 10:45 PM |
Where is the GCC optimization effort directed? | ? | 2010/06/24 01:48 AM |
Where is the GCC optimization effort directed? | Gabriele Svelto | 2010/06/24 02:29 AM |
Where is the GCC optimization effort directed? | ? | 2010/06/24 03:13 AM |
Where is the GCC optimization effort directed? | Andi Kleen | 2010/06/24 03:15 AM |
Where is the GCC optimization effort directed? | ? | 2010/06/24 04:08 AM |
Where is the GCC optimization effort directed? | Gabriele Svelto | 2010/06/24 03:54 AM |
Where is the GCC optimization effort directed? | ? | 2010/06/24 04:15 AM |
Where is the GCC optimization effort directed? | Gabriele Svelto | 2010/06/24 07:22 AM |
Where is the GCC optimization effort directed? | Rohit | 2010/06/24 03:04 AM |
Placebo effect | ? | 2010/06/24 06:37 AM |
Placebo effect | Rohit | 2010/06/24 08:45 AM |
Placebo effect | Vincent Diepeveen | 2010/06/26 02:50 PM |
Compile time | Mark Roulo | 2010/06/26 05:28 PM |
Compile time | Richard Cownie | 2010/06/27 04:44 AM |
Compile time | Mark Roulo | 2010/06/27 10:12 AM |
Compile time | Mark Roulo | 2010/06/27 10:21 AM |
Compile time | EduardoS | 2010/06/27 11:37 AM |
Compile time | Richard Cownie | 2010/06/27 04:07 PM |
Compile time & efficiency | ? | 2010/06/28 12:03 AM |
Compile time & efficiency | Mark Christiansen | 2010/06/28 06:08 AM |
Compile time & efficiency | Linus Torvalds | 2010/06/28 07:48 AM |
kernel programming language | John Simon | 2010/06/29 06:46 PM |
Compile time & efficiency | Richard Cownie | 2010/06/28 09:29 AM |
Compile time & efficiency | Linus Torvalds | 2010/06/28 11:17 AM |
Compile time & efficiency | Richard Cownie | 2010/06/28 02:16 PM |
Compile time & efficiency | Richard Cownie | 2010/06/28 06:23 PM |
Compile time & efficiency | Mark Roulo | 2010/06/29 08:31 AM |
Compile time & efficiency | Richard Cownie | 2010/06/29 11:48 AM |
Compile time & efficiency | rwessel | 2010/06/29 12:28 PM |
C is a crappy | dev | 2010/06/29 07:12 PM |
C is a crappy, but only when you push it out of it's niche | Rohit | 2010/06/30 02:11 AM |
C is a crappy | anon | 2010/06/30 02:17 AM |
C is a crappy | dev | 2010/06/30 07:59 AM |
C is a crappy | Max | 2010/07/01 04:30 AM |
C is a crappy | Michael S | 2010/07/01 07:00 AM |
C is a crappy | Konrad Schwarz | 2010/07/01 08:02 AM |
C is a crappy | Michael S | 2010/07/01 08:50 AM |
C isn't so crappy | anon | 2010/07/01 10:11 AM |
C isn't so crappy | Mikael Tillenius | 2010/07/01 11:39 AM |
C is a crappy | Konrad Schwarz | 2010/07/01 11:22 AM |
C is a crappy | Max | 2010/07/02 08:44 AM |
C is a crappy | rwessel | 2010/07/02 12:33 PM |
C is a crappy | anon | 2010/07/02 01:17 PM |
C is a crappy | Max | 2010/07/02 02:56 PM |
C is a crappy | Max | 2010/07/02 03:13 PM |
C is a crappy | rwessel | 2010/07/02 03:32 PM |
C is a crappy | Max | 2010/07/02 04:19 PM |
C is a crappy | Gabriele Svelto | 2010/07/05 05:25 AM |
C is a crappy | gallier2 | 2010/07/02 12:14 AM |
C is a crappy | Ian Ollmann | 2010/07/06 03:07 PM |
Portability | Max | 2010/07/06 03:37 PM |
C is a crappy | hobold | 2010/07/07 02:31 AM |
C is a crappy | Ian Ollmann | 2010/07/07 05:18 PM |
failure to standardize types | Carlie Coats | 2010/07/07 04:11 AM |
C is a crappy | Konrad Schwarz | 2010/07/07 08:34 AM |
C is a crappy | Ian Ollmann | 2010/07/07 05:29 PM |
C is a crappy NOT | Konrad Schwarz | 2010/07/08 12:29 AM |
C is a crappy | anon | 2010/07/01 10:40 PM |
C type safety | ? | 2010/07/02 01:10 AM |
C type safety | anon | 2010/07/02 11:02 PM |
C is a crappy | dev | 2010/07/03 04:51 PM |
C is a crappy | anon | 2010/07/03 07:02 PM |
C is a crappy | dev | 2010/07/05 07:27 AM |
C is a crappy | ? | 2010/07/05 09:05 AM |
C is a crappy | anonymous | 2010/07/07 08:32 AM |
C is a crappy | ? | 2010/07/07 10:48 PM |
C is a crappy | Anon | 2010/07/08 12:53 AM |
C is a crappy and a crappie is a fish | anonymous | 2010/07/03 07:24 PM |
Compile time & efficiency | Michael S | 2010/06/29 03:18 AM |
Compile time & efficiency | rwessel | 2010/06/29 12:20 PM |
Compile time & efficiency | someone | 2010/06/30 11:03 AM |
Compile time & efficiency | Jouni Osmala | 2010/07/02 05:29 AM |
Compile time & efficiency | Max | 2010/06/28 05:05 PM |
Compile time & efficiency | EduardoS | 2010/06/28 05:11 PM |
Compile time & efficiency | Michael S | 2010/06/29 03:33 AM |
Compile time | Foo_ | 2010/06/28 09:03 AM |
sun 's cc better than GCC? | Silent | 2010/06/23 06:19 PM |
sun 's cc better than GCC? | Foo_ | 2010/06/23 07:06 PM |
sun 's cc better than GCC? | Andi Kleen | 2010/06/24 02:49 AM |
sun 's versus gcc | Vincent Diepeveen | 2010/06/23 03:07 PM |
Why no GCC? | Carlie Coats | 2010/06/23 05:11 AM |