Forks/spawn test using ~latest~ CVS source 20031217 10:00AM (GMT+1)
I've tested the CVS source 20031217 10:00AM (GMT+1) to see if fork/spawn works on Multi CPU (4 Xeon) with Windows 2003 Server (see old message about it in the mailing list). Unfortunatly It doesn't ! Below a test script, use it by running run_t.sh. After some time ( 1 minute) one or more of the 5 sub-shell stop. This batch works perfectly using a single P4 cpu on Windows XP ! Thanks in advance, Philippe. PS 1 : 4-5th attempt to send a mail to [EMAIL PROTECTED] using three different mail account, gmane newgroup ! Anti-spam ? :( PS 2 : Vital for me that this works before end of the month ! Scripts cat 'EOF' t.sh #!/bin/bash i=0 while true do A=$(basename /bin/sh) last_exec=$? i=$(($i+1)) echo Instance $1, loop $i, status $last_exec if [ $last_exec -ne 0 ]; then echo ERROR fi done EOF chmod a+rwx t.sh cat 'EOF' run_t.sh #!/bin/bash t.sh 1 t.sh 2 t.sh 3 t.sh 4 t.sh 5 wait EOF chmod a+rwx run_t.sh ./run_t.sh -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: Forks/spawn test using ~latest~ CVS source 20031217 10:00AM (GMT+1)
On Fri, Dec 19, 2003 at 12:04:02AM +0100, Philippe Torche wrote: I've tested the CVS source 20031217 10:00AM (GMT+1) to see if fork/spawn works on Multi CPU (4 Xeon) with Windows 2003 Server (see old message about it in the mailing list). Unfortunatly It doesn't ! Below a test script, use it by running run_t.sh. After some time ( 1 minute) one or more of the 5 sub-shell stop. This batch works perfectly using a single P4 cpu on Windows XP ! I don't know how many people out there have Multi CPU Xeons, but I suspect there aren't many. This works fine in my single CPU system, as you note. Thanks in advance, Philippe. PS 1 : 4-5th attempt to send a mail to [EMAIL PROTECTED] using three different mail account, gmane newgroup ! Anti-spam ? :( Yes. ZIP file attachments. PS 2 : Vital for me that this works before end of the month ! Sorry, but why would your urgency have any bearing on a community, volunteer project? cgf -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: Forks/spawn test using ~latest~ CVS source 20031217 10:00AM (GMT+1)
Christopher Faylor wrote: On Fri, Dec 19, 2003 at 12:04:02AM +0100, Philippe Torche wrote: I've tested the CVS source 20031217 10:00AM (GMT+1) to see if fork/spawn works on Multi CPU (4 Xeon) with Windows 2003 Server (see old message about it in the mailing list). Unfortunatly It doesn't ! Below a test script, use it by running run_t.sh. After some time ( 1 minute) one or more of the 5 sub-shell stop. This batch works perfectly using a single P4 cpu on Windows XP ! I don't know how many people out there have Multi CPU Xeons, but I suspect there aren't many. This works fine in my single CPU system, as you note. Thanks in advance, Philippe. PS 1 : 4-5th attempt to send a mail to [EMAIL PROTECTED] using three different mail account, gmane newgroup ! Anti-spam ? :( Yes. ZIP file attachments. PS 2 : Vital for me that this works before end of the month ! Sorry, but why would your urgency have any bearing on a community, volunteer project? Sorry too, We have been surprised by our first test on a multi CPU engine, after one year of development without big problem (only a tee problem and rsync not very usable (now using robocopy instead)). And now we are close to the delivery on 3 Multi CPU Xeon. But probably like you, we have only mono processor PC, Pentium 3, Pentium 4 and Athlon XP. cgf I know that threaded applications are difficult to develop. I think you are our chance to correct this. I can do test for you, but I can't give you access to this system ! I'm going to bed, thank you. -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: Forks/spawn test using ~latest~ CVS source 20031217 10:00AM (GMT+1)
On Thu, Dec 18, 2003 at 06:09:54PM -0500, Christopher Faylor [EMAIL PROTECTED] wrote: On Fri, Dec 19, 2003 at 12:04:02AM +0100, Philippe Torche wrote: I've tested the CVS source 20031217 10:00AM (GMT+1) to see if fork/spawn works on Multi CPU (4 Xeon) with Windows 2003 Server (see old message about it in the mailing list). Unfortunatly It doesn't ! Below a test script, use it by running run_t.sh. After some time ( 1 minute) one or more of the 5 sub-shell stop. This batch works perfectly using a single P4 cpu on Windows XP ! I don't know how many people out there have Multi CPU Xeons, but I suspect there aren't many. This works fine in my single CPU system, as you note. Thanks in advance, Philippe. PS 1 : 4-5th attempt to send a mail to [EMAIL PROTECTED] using three different mail account, gmane newgroup ! Anti-spam ? :( Yes. ZIP file attachments. PS 2 : Vital for me that this works before end of the month ! Sorry, but why would your urgency have any bearing on a community, volunteer project? Maybe he's implying there is funding available for working on this if necessary? -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: Forks/spawn test using ~latest~ CVS source 20031217 10:00AM (GMT+1)
On Fri, Dec 19, 2003 at 12:32:10AM +0100, Philippe Torche wrote: Sorry, but why would your urgency have any bearing on a community, volunteer project? Sorry too, We have been surprised by our first test on a multi CPU engine, after one year of development without big problem (only a tee problem and rsync not very usable (now using robocopy instead)). And now we are close to the delivery on 3 Multi CPU Xeon. But probably like you, we have only mono processor PC, Pentium 3, Pentium 4 and Athlon XP. I think you really need to read this: http://www.catb.org/~esr/faqs/smart-questions.html The fact that this is really important to you does not make me reorient my priorities. That's how it works. I'm doing this for free so I get to choose what I want to do. Engaging in a laborious Now try this testing scenario through the next week and into the Christmas holiday was not how I planned to spend my time. I know that threaded applications are difficult to develop. I think you are our chance to correct this. AFAICT, you are talking about fork, not thread. Just as my usual obligatory aside: You do realize that if/when you release your product it will have to be GPLed right? That means you release the source code of the product and the source code of the cygwin DLL and any utilities you bundle with it. -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: Forks/spawn test using ~latest~ CVS source 20031217 10:00AM (GMT+1)
Christopher Faylor wrote: On Fri, Dec 19, 2003 at 12:32:10AM +0100, Philippe Torche wrote: Sorry, but why would your urgency have any bearing on a community, volunteer project? Sorry too, We have been surprised by our first test on a multi CPU engine, after one year of development without big problem (only a tee problem and rsync not very usable (now using robocopy instead)). And now we are close to the delivery on 3 Multi CPU Xeon. But probably like you, we have only mono processor PC, Pentium 3, Pentium 4 and Athlon XP. I think you really need to read this: http://www.catb.org/~esr/faqs/smart-questions.html The fact that this is really important to you does not make me reorient my priorities. That's how it works. I'm doing this for free so I get to choose what I want to do. Engaging in a laborious Now try this testing scenario through the next week and into the Christmas holiday was not how I planned to spend my time. It's Christmas holiday for me too since next wednesday ! And have some nice time with my family. I know that threaded applications are difficult to develop. I think you are our chance to correct this. AFAICT, you are talking about fork, not thread. Just as my usual obligatory aside: You do realize that if/when you release your product it will have to be GPLed right? That means you release the source code of the product and the source code of the cygwin DLL and any utilities you bundle with it. Yes, we will and we do that already ! -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: Forks/spawn test using ~latest~ CVS source 20031217 10:00AM (GMT+1)
Yitzchak Scott-Thoennes wrote: On Thu, Dec 18, 2003 at 06:09:54PM -0500, Christopher Faylor [EMAIL PROTECTED] wrote: On Fri, Dec 19, 2003 at 12:04:02AM +0100, Philippe Torche wrote: I've tested the CVS source 20031217 10:00AM (GMT+1) to see if fork/spawn works on Multi CPU (4 Xeon) with Windows 2003 Server (see old message about it in the mailing list). Unfortunatly It doesn't ! Below a test script, use it by running run_t.sh. After some time ( 1 minute) one or more of the 5 sub-shell stop. This batch works perfectly using a single P4 cpu on Windows XP ! I don't know how many people out there have Multi CPU Xeons, but I suspect there aren't many. This works fine in my single CPU system, as you note. Thanks in advance, Philippe. PS 1 : 4-5th attempt to send a mail to [EMAIL PROTECTED] using three different mail account, gmane newgroup ! Anti-spam ? :( Yes. ZIP file attachments. PS 2 : Vital for me that this works before end of the month ! Sorry, but why would your urgency have any bearing on a community, volunteer project? Maybe he's implying there is funding available for working on this if necessary? Yes, I can (C, C++ is in my knowledge, debug, strace, ...). Thanks a lot, Philippe. -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
