New ECMNET Client
 

The new ECMNET Client is nearing a release. There are a number of improvements in the new client, such as:

support for connecting to multiple servers with one .cfg file
support to run additional curves if unable to connect to a server
no more memory leaks
support to do P-1 and P+1 factoring, which will be used when the new ECMNET Server is released

This has been tested against some existing ECMNET servers for a couple of weeks, including Paul Leyland's BMtR factoring project and WIlliam Lipp's Eleven Smooth factoring project. There are no known bugs, in other words, I squashed all bugs that I am aware of.

If you are interested in giving it a spin, send me an e-mail at rogue (at) wi.rr.com and I'll get it off to you. If you need a build, I can build it for Windows (under DOS) and Windows (under Cygwin). Since you get the full source, you would be able to compile under any OS that has the gcc compiler.

Are you working on a new server too?

I'm interested in the client (and new server!) although i won't need it straight away. I guess you'll make it available for general download soon?

Yes, I'm working on a new server, but the server code is much more complex than the client code. It will take a while. I'm only about 25% done with it and don't expect to be done until July at the earliest as I can only work on it in increments.

The new client will be available for general download from Paul Zimmerman's site in the near future.

Good work :)
 

It will be very intressting to see the new client. On the server side. would it be possible to make a sort of template file to use for 'exporting' the stats that are generated by the server. So it would be more easier to import this data to a database or to prepare it for your own code creations ;)

It would be quite handy to read a template ini file or similar that has the global layout. Altho also quite challenging to make it work without problems.

[QUOTE=BotXXX]It will be very intressting to see the new client. On the server side. would it be possible to make a sort of template file to use for 'exporting' the stats that are generated by the server. So it would be more easier to import this data to a database or to prepare it for your own code creations ;)

It would be quite handy to read a template ini file or similar that has the global layout. Altho also quite challenging to make it work without problems.[/QUOTE]

Right now I haven't thought about stats at all. No-one who is currently running a server seems to care much about user stats and user stats on the server have no meaning for clients. They will probably not be included in the next release of the server. The same can be said about client stats. They appear to be fairly meaningless. If someone really wants stats, they can parse the log files, otherwise, I consider them to be a waste of memory. If someone can make a good case for client stats, then I will work with them to add them to the client, but there are no guarantees that they will be in the next release.

On another note for the server, the format of the ini file will be changed, but only slightly so that the current ini file is still compatible with the new server, but I won't support backward compatibility because of the new functionality that is included in the new server.

If you would like to look at the client code, just drop me an e-mail at the above address. The client code is included in the distribution.

Last night i found an annoying problem with the client.

I put up a few dozen numbers on the server some of which contained small factors.

I set recurse=1 to continue factor composite cofactors, but as a side effect, also composite factors found are put back on the server. No problem, but sometimes ecm is not able to factor the composite (to high bound) and only finds the input numbers as a factor and gets stuck in an infinite loop without contacting the server.

[code]
...
[2004-09-03 18:16:18 GMT] Found input number.
[2004-09-03 18:16:19 GMT] Found input number.
[2004-09-03 18:16:20 GMT] Found input number.
[2004-09-03 18:16:21 GMT] Found input number.
...
[/code]

Sander

[QUOTE=smh]Last night i found an annoying problem with the client.

I put up a few dozen numbers on the server some of which contained small factors.

I set recurse=1 to continue factor composite cofactors, but as a side effect, also composite factors found are put back on the server. No problem, but sometimes ecm is not able to factor the composite (to high bound) and only finds the input numbers as a factor and gets stuck in an infinite loop without contacting the server.

[code]
...
[2004-09-03 18:16:18 GMT] Found input number.
[2004-09-03 18:16:19 GMT] Found input number.
[2004-09-03 18:16:20 GMT] Found input number.
[2004-09-03 18:16:21 GMT] Found input number.
...
[/code]

Sander[/QUOTE]

I think I understand what you are saying, but it would be best for me to see your ecmserver.log and ecmclient.log. Could you PM them to me? If I understand your problem correctly from your post, this shouldn't be too difficult to handle in the sever.

--Mark

Hi Mark,

I sent parts of the logs to your wi.rr.com account.

Sander

This is a problem with the old client, but is also a problem with the new client. I will address it with the new client/server.

Linux version?
 

Hi Mark,

I just tried to compile the ecmclient (2.0k) under Linux and ran into difficulties.
It should work with Linux, shouldn't it?

Maybe someone can provide the binaries? :help:

Dennis

Many others have compiled under Linux without problem. What errors are you getting during the compile? Either I or someone else can help you resolve them.

[QUOTE=rogue]Many others have compiled under Linux without problem. What errors are you getting during the compile? Either I or someone else can help you resolve them.[/QUOTE]
gcc -c -O6 -DUNIX thread.cpp
thread.cpp: In function `void MakeThread(_thread_handle*, void (*)(void*),void*)':
thread.cpp:78: error: `usleep' undeclared (first use this function)
thread.cpp:78: error: (Each undeclared identifier is reported only once for each function it appears in.)

I already skiped the compilation of the server as I only need the client...

Prior are some warnings:
[code]
talis@Suse:~/bin/ECMNET/source> make
gcc -c -O6 -DUNIX ecmclient.cpp
gcc -c -O6 -DUNIX common.cpp
common.cpp: In function `unsigned int inet_address(const char*)':
common.cpp:308: Warnung: return of negative value `-1' to `unsigned int'
common.cpp:308: Warnung: argument of negative value `-1' to `unsigned int'
gcc -c -O6 -DUNIX ecm_list.cpp
ecm_list.cpp: In member function `double ecm_list::get_bestb1(int, int*)':
ecm_list.cpp:566: Warnung: assignment to `int' from `double'
ecm_list.cpp:566: Warnung: argument to `int' from `double'
ecm_list.cpp:568: Warnung: assignment to `int' from `double'
ecm_list.cpp:568: Warnung: argument to `int' from `double'
ecm_list.cpp:570: Warnung: assignment to `int' from `double'
ecm_list.cpp:570: Warnung: argument to `int' from `double'
ecm_list.cpp:572: Warnung: assignment to `int' from `double'
ecm_list.cpp:572: Warnung: argument to `int' from `double'
ecm_list.cpp:574: Warnung: assignment to `int' from `double'
ecm_list.cpp:574: Warnung: argument to `int' from `double'
ecm_list.cpp:576: Warnung: assignment to `int' from `double'
ecm_list.cpp:576: Warnung: argument to `int' from `double'
ecm_list.cpp:580: Warnung: assignment to `int' from `double'
ecm_list.cpp:580: Warnung: argument to `int' from `double'
ecm_list.cpp:582: Warnung: assignment to `int' from `double'
ecm_list.cpp:582: Warnung: argument to `int' from `double'
ecm_list.cpp:584: Warnung: assignment to `int' from `double'
ecm_list.cpp:584: Warnung: argument to `int' from `double'
ecm_list.cpp:586: Warnung: assignment to `int' from `double'
ecm_list.cpp:586: Warnung: argument to `int' from `double'
ecm_list.cpp:588: Warnung: assignment to `int' from `double'
ecm_list.cpp:588: Warnung: argument to `int' from `double'
ecm_list.cpp:590: Warnung: assignment to `int' from `double'
ecm_list.cpp:590: Warnung: argument to `int' from `double'
ecm_list.cpp:592: Warnung: assignment to `int' from `double'
ecm_list.cpp:592: Warnung: argument to `int' from `double'
[/code]

Remove thread.o as a requirement to link ecmclient.

[QUOTE=Mystwalker]gcc -c -O6 -DUNIX thread.cpp
thread.cpp: In function `void MakeThread(_thread_handle*, void (*)(void*),void*)':
thread.cpp:78: error: `usleep' undeclared (first use this function)
thread.cpp:78: error: (Each undeclared identifier is reported only once for each function it appears in.)

[/code][/QUOTE]

usleep() sleeps for an integral number of microseconds. Its definition should be in <unistd.h> so make sure that file is included. source/defs.h should #include it.

If usleep() is not defined in your /usr/include/unistd.h you need to go looking for it. grep(1) is your friend.


Paul

[QUOTE=xilman]usleep() sleeps for an integral number of microseconds. Its definition should be in <unistd.h> so make sure that file is included. source/defs.h should #include it.

If usleep() is not defined in your /usr/include/unistd.h you need to go looking for it. grep(1) is your friend.


Paul[/QUOTE]

That is true and that is what I would have posted, but AFAICT, thread.c is not needed in the ECMNet client. The functions it contains are never called. Off the top of my head, I think the code was intended for use with ECMNet as a service for setting its priority. In any case, that code is removed from the new client/server.

Thanks for your help! :bow:

Both suggestions solved that problem, but now I encountered the next one:

[code]gcc -c -O6 -DUNIX users.cpp
gcc -O6 -DUNIX -o ecmclient ecmclient.o common.o ecm_list.o users.o
ecmclient.o(.text+0x3e3): In function `getfactors(char*, char*, char*, ecm_list*, char*)':
: undefined reference to `operator new[](unsigned)'
ecmclient.o(.text+0x507): In function `getfactors(char*, char*, char*, ecm_list*, char*)':
: undefined reference to `operator new[](unsigned)'
ecmclient.o(.text+0x6ce): In function `getfactors(char*, char*, char*, ecm_list*, char*)':
: undefined reference to `operator new[](unsigned)'
ecmclient.o(.text+0xd75): In function `admin(char*, int, char*)':
: undefined reference to `operator new[](unsigned)'
ecmclient.o(.text+0xf8b): In function `admin(char*, int, char*)':
: undefined reference to `operator delete(void*)'
ecmclient.o(.text+0x1430): In function `main':
: undefined reference to `operator new[](unsigned)'
ecmclient.o(.text+0x1442): In function `main':
: undefined reference to `operator new[](unsigned)'
ecmclient.o(.text+0x1453): In function `main':
: undefined reference to `operator new[](unsigned)'
ecmclient.o(.eh_frame+0x12): undefined reference to `__gxx_personality_v0'
common.o(.text+0x521): In function `putstring(int, char*)':
: undefined reference to `operator new[](unsigned)'
common.o(.text+0x54a): In function `putstring(int, char*)':
: undefined reference to `operator delete(void*)'
common.o(.text+0x6e6): In function `getstring(int)':
: undefined reference to `operator new[](unsigned)'
common.o(.text+0x79d): In function `getstring(int)':
: undefined reference to `operator new[](unsigned)'
common.o(.text+0x7cb): In function `getstring(int)':
: undefined reference to `operator new[](unsigned)'
common.o(.text+0x848): In function `getstring(int)':
: undefined reference to `operator new[](unsigned)'
common.o(.text+0x871): In function `getstring(int)':
: undefined reference to `operator delete(void*)'
common.o(.text+0x89f): In function `getstring(int)':
: undefined reference to `operator new[](unsigned)'
common.o(.text+0x9e1): In function `inet_address(char const*)':
: undefined reference to `__h_errno_location()'
common.o(.text+0xcdc): In function `open_server(char*, int, char*)':
: undefined reference to `__h_errno_location()'
common.o(.text+0xf2c): In function `getb1andfacinfo(ecm_list*, int, int, int)':
: undefined reference to `operator new[](unsigned)'
common.o(.text+0xf71): In function `getb1andfacinfo(ecm_list*, int, int, int)':
: undefined reference to `operator delete(void*)'
common.o(.text+0x10ac): In function `returnwork(int, ecm_list*, int, int)':
: undefined reference to `operator new[](unsigned)'
common.o(.text+0x1100): In function `returnwork(int, ecm_list*, int, int)':
: undefined reference to `operator delete(void*)'
common.o(.text+0x1177): In function `returnwork(int, ecm_list*, int, int)':
: undefined reference to `operator new[](unsigned)'
common.o(.text+0x11a6): In function `returnwork(int, ecm_list*, int, int)':
: undefined reference to `operator delete(void*)'
common.o(.text+0x122d): In function `returnwork(int, ecm_list*, int, int)':
: undefined reference to `operator new[](unsigned)'
common.o(.text+0x1280): In function `returnwork(int, ecm_list*, int, int)':
: undefined reference to `operator delete(void*)'
common.o(.text+0x1351): In function `returnwork(int, ecm_list*, int, int)':
: undefined reference to `operator new[](unsigned)'
common.o(.text+0x13a5): In function `returnwork(int, ecm_list*, int, int)':
: undefined reference to `operator delete(void*)'
common.o(.text+0x1474): In function `returnwork(int, ecm_list*, int, int)':
: undefined reference to `operator new[](unsigned)'
common.o(.text+0x14be): In function `returnwork(int, ecm_list*, int, int)':
: undefined reference to `operator delete(void*)'
common.o(.text+0x1570): In function `returnwork(int, ecm_list*, int, int)':
: undefined reference to `operator new[](unsigned)'
common.o(.text+0x15d8): In function `returnwork(int, ecm_list*, int, int)':
: undefined reference to `operator delete(void*)'
common.o(.text+0x1634): In function `returnwork(int, ecm_list*, int, int)':
: undefined reference to `operator new[](unsigned)'
common.o(.text+0x167e): In function `returnwork(int, ecm_list*, int, int)':
: undefined reference to `operator delete(void*)'
common.o(.text+0x174c): In function `returnwork(int, ecm_list*, int, int)':
: undefined reference to `operator new[](unsigned)'
common.o(.text+0x17b4): In function `returnwork(int, ecm_list*, int, int)':
: undefined reference to `operator new[](unsigned)'
common.o(.text+0x1814): In function `returnwork(int, ecm_list*, int, int)':
: undefined reference to `operator delete(void*)'
common.o(.text+0x1879): In function `returnwork(int, ecm_list*, int, int)':
: undefined reference to `operator new[](unsigned)'
common.o(.text+0x195a): In function `nextword(char*)':
: undefined reference to `operator new[](unsigned)'
common.o(.eh_frame+0x11): undefined reference to `__gxx_personality_v0'
ecm_list.o(.text+0x1466): In function `ecm_list::get_allb1info(char*)':
: undefined reference to `operator new[](unsigned)'
ecm_list.o(.text+0x1552): In function `ecm_list::get_allb1info(int)':
: undefined reference to `operator new[](unsigned)'
ecm_list.o(.text+0x1728): In function `ecm_list::get_facinfo(char*, int)':
: undefined reference to `operator new[](unsigned)'
ecm_list.o(.text+0x179c): In function `ecm_list::get_facinfo(int, int)':
: undefined reference to `operator new[](unsigned)'
ecm_list.o(.text+0x1e1f): In function `ecm_list::StoreDetails(char*, char*)':
: undefined reference to `operator new[](unsigned)'
ecm_list.o(.text+0x1e49): more undefined references to `operator new[](unsigned)' follow
ecm_list.o(.text+0x1e73): In function `ecm_list::StoreDetails(char*, char*)':
: undefined reference to `operator new(unsigned)'
ecm_list.o(.text+0x1e9d): In function `ecm_list::StoreDetails(char*, char*)':
: undefined reference to `operator new(unsigned)'
ecm_list.o(.text+0x1f87): In function `ecm_list::StoreDetails(char*, char*)':
: undefined reference to `operator new(unsigned)'
ecm_list.o(.text+0x278c): In function `ecm_list::newnumber(char*, char*)':
: undefined reference to `operator new[](unsigned)'
ecm_list.o(.text+0x27cf): In function `ecm_list::newnumber(char*, char*)':
: undefined reference to `operator new[](unsigned)'
ecm_list.o(.text+0x291d): In function `ecm_list::newnumber(char*, char*)':
: undefined reference to `operator new[](unsigned)'
ecm_list.o(.eh_frame+0x11): undefined reference to `__gxx_personality_v0'
users.o(.text+0x1c0): In function `user::LoadUsers(char*, int)':
: undefined reference to `operator new[](unsigned)'
users.o(.text+0x81f): In function `user::storeuser(char const*, int, double, double, int, char const*, char const*)':
: undefined reference to `operator new[](unsigned)'
users.o(.text+0x83a): In function `user::storeuser(char const*, int, double, double, int, char const*, char const*)':
: undefined reference to `operator new[](unsigned)'
users.o(.text+0x853): In function `user::storeuser(char const*, int, double, double, int, char const*, char const*)':
: undefined reference to `operator new[](unsigned)'
users.o(.text+0x105): In function `user::~user [not-in-charge]()':
: undefined reference to `operator delete(void*)'
users.o(.text+0x195): In function `user::~user [in-charge]()':
: undefined reference to `operator delete(void*)'
users.o(.eh_frame+0x11): undefined reference to `__gxx_personality_v0'
collect2: ld returned 1 exit status
make: *** [ecmclient] Fehler 1[/code]

Unfortunately, I'm quite a newbie when it comes to Linux and C++. :sad:

[QUOTE=Mystwalker]Thanks for your help! :bow:

Both suggestions solved that problem, but now I encountered the next one:

Unfortunately, I'm quite a newbie when it comes to Linux and C++. :sad:[/QUOTE]

Try g++ instead of gcc.

--Mark

Looks a lot better (to me) now:

[code]g++ -O6 -DUNIX -o ecmclient ecmclient.o common.o ecm_list.o users.o
common.o(.text+0x9e1): In function `inet_address(char const*)':
: undefined reference to `__h_errno_location()'
common.o(.text+0xcdc): In function `open_server(char*, int, char*)':
: undefined reference to `__h_errno_location()'
collect2: ld returned 1 exit status[/code]

As a side effect, I learned a lot about Makefiles. :smile:

Ok, figured out the last problem with the help of this site:

[url]http://www.ecompute.org/factors/unix.html[/url]

It is running now! :banana:
Thanks once again for your kind support. :bow:

Need help getting ECMNET to compile
 

Is anyone else still watching this thread? The [URL=http://www.ecompute.org/factors/unix.html]http://www.ecompute.org/factors/unix.html[/URL] site is down and I've gotten to this point:

[CODE]g++ -O6 -DUNIX -o ecmclient ecmclient.o common.o ecm_list.o users.o
common.o(.text+0x9e1): In function `inet_address(char const*)':
: undefined reference to `__h_errno_location()'
common.o(.text+0xcdc): In function `open_server(char*, int, char*)':
: undefined reference to `__h_errno_location()'
collect2: ld returned 1 exit status[/CODE]
Would anyone mind helping me resolve these last compilation problems?

Thanks!

[QUOTE=rpropper]Is anyone else still watching this thread? The [URL=http://www.ecompute.org/factors/unix.html]http://www.ecompute.org/factors/unix.html[/URL] site is down and I've gotten to this point:

[CODE]g++ -O6 -DUNIX -o ecmclient ecmclient.o common.o ecm_list.o users.o
common.o(.text+0x9e1): In function `inet_address(char const*)':
: undefined reference to `__h_errno_location()'
common.o(.text+0xcdc): In function `open_server(char*, int, char*)':
: undefined reference to `__h_errno_location()'
collect2: ld returned 1 exit status[/CODE]
Would anyone mind helping me resolve these last compilation problems?

Thanks![/QUOTE]

What OS is this on?

BTW, if you would like the new client (which shouldn't have any many compile issues), send an email to me at rogue (at) wi.rr.com and I'll send it to you. It is fully compatible with the 2.0 server.

Linux, I'd guess...
I'll look into the source to figure out the changes once I get to that particular machine (1st or 2nd January).
I'm going to post the changes here, as I've seen that this thread is the only reference google finds for that problem. :sad:

[QUOTE=Mystwalker]Linux, I'd guess...
I'll look into the source to figure out the changes once I get to that particular machine (1st or 2nd January).
I'm going to post the changes here, as I've seen that this thread is the only reference google finds for that problem. :sad:[/QUOTE]

Yes, it was linux. I sent the 2.5.5 source to rpropper yesterday and haven't heard from him yet. If this compile issue still exists with linux, I will be more easily able to address it.

BTW, I had asked Paul Zimmerman to host the new ECMNet source a few months ago and he agreed. I sent it off to him today, but have not received a response yet...

Here is the diff:

talis@Suse:~/bin/ECMNET/source> diff common.cpp.original common.cpp
[code]292,294d291
< #ifdef UNIX
< extern int h_errno;
< #endif
306c303
< h_errno);
---
> errno);[/code]

So:
- take out those 3 lines
- change "h_errno" to "errno"

That made it run on my system. :smile:
Hope that helps!