I'm going mad trying to compile anything. I've followed the instructions and keep getting error message after error message.

Anyone know what this means?

seti_boinc Command line error D2016 : '/GL' and '/YX..\\StdAfx.h' command-line options are incompatible

Thanks in Advance :-)

Linux or Windows? Might be nessicary info for those who can help you.

Windows :-)

---

BORG,

I got that error as well. To remove it (at the cost of a couple % performance), you will have to disable "Whole Program Optimization" under "General Options" in ALL sub-projects EXCEPT seti_boinc itself. This error only occurs when you're using older versions (like the ones I got as trials) of the Intel tools than the commercial ones I'm using now. The current versions I got when I bought the licenses do not give me this error.

HTH,
Simon.

---

Donate to SETI@Home via PayPal!

Optimized SETI@Home apps + Information

BORG,

I got that error as well. To remove it (at the cost of a couple % performance), you will have to disable "Whole Program Optimization" under "General Options" in ALL sub-projects EXCEPT seti_boinc itself. This error only occurs when you're using older versions (like the ones I got as trials) of the Intel tools than the commercial ones I'm using now. The current versions I got when I bought the licenses do not give me this error.

HTH,
Simon.

Thanks KWSN - Chicken of Angnor

Now it runs till the end and give more errors LOL! I'll look at it again later tonight. I'm calling it guits for now >:-(

---

hrm... and here even... http://setiathome.berkeley.edu/forum_thread.php?id=20892

---

OK, got it to build - Ran benchmark - OK so far.

Finnished Benchmark - 50% increase - Then I ran rescmp result.sah.default-515-1.txt result.sah.optimized-1.txt

And I get this message - Cannot open result.sah.default-5.15-1.txt

Anyone else have this? Is it bad news?

---

You most probably did not run the default-515 client (or it's named incorrectly). What the script is telling you is that it cannot find the result data from the default client, which is what it's checking the result against.

Right-click benchmark.cmd, click edit. Then check the lines near the end where it says "Testing default-515" or the like. Check what name the .exe file has and rename accordingly, then run the benchmark again.

Without a result from the default client, you have no valid work to compare against.

HTH,
Simon.

---

Donate to SETI@Home via PayPal!

Optimized SETI@Home apps + Information

You most probably did not run the default-515 client (or it's named incorrectly). What the script is telling you is that it cannot find the result data from the default client, which is what it's checking the result against.

Right-click benchmark.cmd, click edit. Then check the lines near the end where it says "Testing default-515" or the like. Check what name the .exe file has and rename accordingly, then run the benchmark again.

Without a result from the default client, you have no valid work to compare against.

HTH,
Simon.

Thanks again.

File should read result.sah.default-5.15-1.txt

But its showing as result.sah.default5.15-1.txt

Working now.

---

Oops :)

---

Donate to SETI@Home via PayPal!

Optimized SETI@Home apps + Information

Okay,

anyone who's trying to build their own client on Windows -
check out these links:

All Visual C++ compiler options (with ample explanation)

Visual C++ documentation

HTH,
Simon.

---

Donate to SETI@Home via PayPal!

Optimized SETI@Home apps + Information

Okay,

anyone who's trying to build their own client on Windows -
check out these links:

All Visual C++ compiler options (with ample explanation)

Visual C++ documentation

HTH,
Simon.

Simon

I have question about the released windows apts.

I've noticed that the size of the executable is rather large as is mine, compared to the stock app and the former Crunch3r apps.

1 - Stock - 1248KB
2 - Crunch3r - 1616KB
3 - kwsn - 6984KB Mine is around the same KB.

From the 3 - #2 still seems to be the fastest.

It would seem that we have a some ways to go in improving these early releases.

But I'm sure many will appriciate your efforts and that of those that comtributed.

Great work gang :-)

Borg :-)

---

Okay,

anyone who's trying to build their own client on Windows -
check out these links:

All Visual C++ compiler options (with ample explanation)

Visual C++ documentation

HTH,
Simon.

Simon

I have question about the released windows apts.

I've noticed that the size of the executable is rather large as is mine, compared to the stock app and the former Crunch3r apps.

1 - Stock - 1248KB
2 - Crunch3r - 1616KB
3 - kwsn - 6984KB Mine is around the same KB.

From the 3 - #2 still seems to be the fastest.

It would seem that we have a some ways to go in improving these early releases.

But I'm sure many will appriciate your efforts and that of those that comtributed.

Great work gang :-)

Borg :-)

The size issue is interesting. The executable contains huge runs of all zero data with an occasional non-zero run of two or three bytes. I suspect that where ICC sees an array that can be pre-initialized and stored it does so, perhaps that is slightly faster.

Whatever the reason, it may be a difference between ICC 9.0 and ICC 9.1. Crunch3r noted that 9.1 often makes much larger executables in this Einstein@home post.

                                                                Joe

This may to some extent be due to the "optimize for Win98" setting.

If you set it to "default", it will create 4KB-aligned binaries and zerofill the rest, thereby making a much larger executable. Same when you set it to "Optimize for Win98". Only when you set it to not optimize for Win98 (/OPT:NOWIN98) it uses a smaller (512byte-aligned) chunk size and wastes less space.

Still, the difference between the two methods is not that large (in the order of a few hundred KBs), and yes, ICC 9.0/IPP 5.0 made MUCH smaller binaries.

It is already set up to strip unneeded things from the binary, and I haven't found a way to make the binary much smaller. I tried not linking some libs, but get instant linker errors as soon as I do that - and the symbols those errors reference are inside the libraries I just removed. So well, guess I can't link less.

Regards,
Simon.

---

Donate to SETI@Home via PayPal!

Optimized SETI@Home apps + Information

This may to some extent be due to the "optimize for Win98" setting.

If you set it to "default", it will create 4KB-aligned binaries and zerofill the rest, thereby making a much larger executable. Same when you set it to "Optimize for Win98". Only when you set it to not optimize for Win98 (/OPT:NOWIN98) it uses a smaller (512byte-aligned) chunk size and wastes less space.

Still, the difference between the two methods is not that large (in the order of a few hundred KBs), and yes, ICC 9.0/IPP 5.0 made MUCH smaller binaries.

It is already set up to strip unneeded things from the binary, and I haven't found a way to make the binary much smaller. I tried not linking some libs, but get instant linker errors as soon as I do that - and the symbols those errors reference are inside the libraries I just removed. So well, guess I can't link less.

Regards,
Simon.

I did notice the increase with the Win98 optimization, about 3% it seems.

The size will be a problem for very few because the zip compressed download keeps it from being a problem for those on dial-up. I did run into a temporary hitch because the partition containing BOINC on my Win95 system ended up being nearly full after installing the optimized app; the Scheduler wouldn't send work for it until I freed up some space.

I know that some users had problems with the Linux apps compressed with UPX, but I'll note that I UPX compressed the Windows generic SSE2 version and MMX version and tested standalone on this Win98 system. Both ran fine, and the compressed SSE2 version is now running with BOINC on this system. The compressed MMX version is on my Win95 system (making even more free space) and will go active in about 4 hours.

                                                        Joe

Synergy has found the greatest optimized app ever created here.

---

me@rescam.org

Nice catch Misfit :o)

---

Donate to SETI@Home via PayPal!

Optimized SETI@Home apps + Information

Misfit i saw it too.

Maybe a bug in the xml.

Mike

---

With each crime and every kindness we birth our future.

Doesn't look real to me. By the way, it is not Seti@home but vtu@home which I am not familiar with. I would think he has found a crack for it.

Misfit i saw it too.

Maybe a bug in the xml.

Mike

---

Doesn't look real to me. By the way, it is not Seti@home but vtu@home which I am not familiar with. I would think he has found a crack for it.

Misfit i saw it too.

Maybe a bug in the xml.

Mike

I looked at the stats page yesterday.
I didn´t find any person who get this result.

Mike

---

With each crime and every kindness we birth our future.

Hi folks,

I've put a new Windows sources package online. This version is the same that was used to compile my R-1.3 clients.

All source edits now have my name in them - you can easily find all of them by searching for "Simon Zadra" in the Sources. I recommend grep for Windows.

Source Files edited:

analyzeFuncs.cpp
chirpfft.cpp

A lot of project files were edited also, this does not classify as a source edit though. The edits were made so Intel's compiler and libraries are included and linked by default.

Regards,
Simon.

---

Donate to SETI@Home via PayPal!

Optimized SETI@Home apps + Information

Hi Simon


I have The EM64T for the compiler, IPP and MKL. Can these be used for a 64bit seti client?

Borg :-)

---