Board index » cppbuilder » Re: Twine 2 Release Candidate 3

2005-12-17 05:55:34 AM
cppbuilder83
Carmelo Viavattene wrote:
Jonathan
jomitech.com/mtbcc32.shtml - Multi-threaded compilation for BCB
jomitech.com/forum - JomiTech Forums
 

Carmelo Viavattene ha scritto:
Before I must call mtbcc32CompileExecutor?
Regards,
Carmelo Viavattene
 

Carmelo Viavattene wrote:
mtbcc32.exe file name.
Jonathan
jomitech.com/mtbcc32.shtml - Multi-threaded compilation for BCB
jomitech.com/forum - JomiTech Forums
 

{smallsort}

Jonathan Benedicto ha scritto:
MAKE Version 5.2
Fatal: Unable to execute command:
C:\Program~1\Borland\CBuild~1\bin\..\bin\C:\Program~1\Borland\Cbuilder5\Bin\mtbcc...
And now?
Regards,
Carmelo Viavattene
 

Jonathan Benedicto ha scritto:
*** There are errors ***
Carmelo Viavattene
 

Carmelo Viavattene wrote:
Jonathan
jomitech.com/mtbcc32.shtml - Multi-threaded compilation for BCB
jomitech.com/forum - JomiTech Forums
 

Jonathan Benedicto ha scritto:
Regards,
Carmelo
 

Jonathan Benedicto ha scritto:
Carmelo
 

Jonathan Benedicto wrote:
the memory, because I can copy them in less than a second. I was sharing
a hard drive when I started two make processes with bcc32 and I still
got significant benefit from hyperthreading, and nothing with bcc32ide.
I think I was comparing apples to apples. It's just my architecture
reached its limitations with the IDE compiler. It's not that I'm
unhappy, because either bcc32ide or mtbcc32 reduces my compile times
from 6.5 minutes to 26 seconds. When I compile a unit, I no longer wait
7 seconds, I wait half a second, and that's very noticeable.
Yeah, Hyperthreading is not equivalent to two distinct processors. The
main ALU circuits are duplicated, along with the integer registers. In a
dual-core system everything would be duplicated, including the cache and
the floating point unit. They still share the same memory, system bus
and hard disk, so even a dual processor system won't be nearly twice as
fast as a single one. That's when your distributed system will come to
help. I'd be curious to try your solution with a dual-core, but it's not
gonna happen this year, I'm afraid. There's no time, and I don't have
the hardware either. When Delegator becomes available, I'll try it right
away.
Tom
 

Tamas Demjen wrote:
every single file in RAM, on the second build, each executor instance has
all the files in separate RAM blocks, making each have their own RAM disk.
bit on the other, and so on ?
Then, I won't be troubled about HDD speeds, just LAN speeds :-)
Jonathan
jomitech.com/mtbcc32.shtml - Multi-threaded compilation for BCB
jomitech.com/forum - JomiTech Forums
 

Jonathan Benedicto wrote:
but it doesn't show any further information. No compiler errors.
Tom
 

Tamas Demjen wrote:
Jonathan
jomitech.com/mtbcc32.shtml - Multi-threaded compilation for BCB
jomitech.com/forum - JomiTech Forums
 

Jonathan Benedicto wrote:
--
Regards,
Andreas Hausladen
 

Andreas Hausladen wrote:
know if Borland would appreciate by way of getting BCB6 <g>
Jonathan
jomitech.com/mtbcc32.shtml - Multi-threaded compilation for BCB
jomitech.com/forum - JomiTech Forums
 

Hi Jonathan,
I've tried this latest Twine V2 and I get the same. If I use V1.7.2 then it
works, replace with V2 and I get '*** There are errors. ***' like everyone
else.
:(
Simon