 |
|
|
|
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Hi All,
Just got back from IDF last night. One of intel's booths had an Itanium
system running linux, with the intent of highlighting intel's optimized
compiler (vs standard gcc). It caught my attention because the application
they had compiled and were running in parallel was pov, rendering one of
the metal texture preview scenes (coppers or brasses). They had two
instances of pov running parallel, one built with gcc, the other with the
intel compiler. The intel compile was running about 30% faster. I didn't
actually time the render, and I don't know what clock speed the itanium was
running at, so I can't say much for the overall performance. I just
thought it was cool that they were using pov for their demo.
Ken
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Ken Cecka wrote:
>
> Hi All,
>
> Just got back from IDF last night. One of intel's booths had an Itanium
> system running linux, with the intent of highlighting intel's optimized
> compiler (vs standard gcc). It caught my attention because the application
> they had compiled and were running in parallel was pov, rendering one of
> the metal texture preview scenes (coppers or brasses). They had two
> instances of pov running parallel, one built with gcc, the other with the
> intel compiler. The intel compile was running about 30% faster. I didn't
> actually time the render, and I don't know what clock speed the itanium was
> running at, so I can't say much for the overall performance. I just
> thought it was cool that they were using pov for their demo.
>
Sounds interesting although you can't say much from such a demonstration,
the gcc version could be compiled without any optimization or there could
be other differences.
Christoph
--
Christoph Hormann <chr### [at] gmx de>
IsoWood include, radiosity tutorial, TransSkin and other
things on: http://www.schunter.etc.tu-bs.de/~chris/
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Christoph Hormann wrote:
>
>
> Ken Cecka wrote:
>>
>> Hi All,
>>
>> Just got back from IDF last night. One of intel's booths had an Itanium
>> system running linux, with the intent of highlighting intel's optimized
>> compiler (vs standard gcc). It caught my attention because the
>> application they had compiled and were running in parallel was pov,
>> rendering one of
>> the metal texture preview scenes (coppers or brasses). They had two
>> instances of pov running parallel, one built with gcc, the other with the
>> intel compiler. The intel compile was running about 30% faster. I
>> didn't actually time the render, and I don't know what clock speed the
>> itanium was
>> running at, so I can't say much for the overall performance. I just
>> thought it was cool that they were using pov for their demo.
>>
>
> Sounds interesting although you can't say much from such a demonstration,
> the gcc version could be compiled without any optimization or there could
> be other differences.
>
> Christoph
>
The guy manning the demo told me what the gcc optimization level was (6 I
think), but as you say, there are plenty of other factors that could be
influencing it. I wasn't really trying to evaluate the performance
improvement. It was just cool to see pov show up at something like IDF.
Ken
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
> Ken Cecka wrote:
>The guy manning the demo told me what the gcc optimization level was (6 I
>think), but as you say, there are plenty of other factors that could be
To my knowledge, the maximum is -O3. Either it is a severily patched gcc, or
a completely clueless marketing guy or... just misinformation.
To current knowledge, Itanium is dreadfully slow (it has already been
renamed by writers Itanic...)
Bye!!!
Alessandro Coppo
a.c### [at] iolit
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Ken Cecka wrote:
> Hi All,
>
> The intel compile was running about 30% faster.
Probably because the intel compile was compiled for the new EPIC (explicitly
parrellel ...) architecture.
and that gcc may not yet be capable of compiling with the new instructions.
There was an article linked to from Slashdot a while ago on the new
architecture - sounds exciting.
--
Bye,
Pabs
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Alessandro Coppo wrote:
>
> To my knowledge, the maximum is -O3. Either it is a severily patched gcc, or
> a completely clueless marketing guy or... just misinformation.
>
O6 is legal in gcc, but AFAIK doesn't actually add anything to O3 - at least on
i386. The maximum for pgcc on i386 is O7.
> To current knowledge, Itanium is dreadfully slow (it has already been
> renamed by writers Itanic...)
>
It is very slow with 32 bit code. I haven't seen any benchmark numbers with
native 64 bit code, but it is said to be quite fast.
--
Margus Ramst
Personal e-mail: mar### [at] peakeduee
TAG (Team Assistance Group) e-mail: mar### [at] tagpovrayorg
Home page http://www.hot.ee/margusrt
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
|
|
