Hi folks,
the summer and the vacations are over so I am back.
my diskless dual 200MMX is now in the case and running again.

pulsefind.cpp ---> its necessary to have a look on it to make some optimizations.

regards heinz~seti_britta

Did you compile whole client under Orcas already?
Any performance comparision?

Compiled now FFTW-3.1.2 under Orcas too
here is the first result ---->
fftw-3.1.2 benchfsse(orcas) started
benchf_sse.exe -opatient 64 128 256 512 1024 2048 4096
Problem: 64, setup: 32.54 ms, time: 2.33 us, ``mflops'': 823.7
Problem: 128, setup: 79.49 ms, time: 6.43 us, ``mflops'': 697.23
Problem: 256, setup: 196.37 ms, time: 14.25 us, ``mflops'': 718.72
Problem: 512, setup: 470.92 ms, time: 31.01 us, ``mflops'': 743
Problem: 1024, setup: 866.86 ms, time: 70.68 us, ``mflops'': 724.4
Problem: 2048, setup: 2.17 s, time: 155.61 us, ``mflops'': 723.88
Problem: 4096, setup: 7.27 s, time: 329.37 us, ``mflops'': 746.15
fftw-3.1.2 benchfsse ended.
-----------------------------------------------------------------------------------------------------
It shows that this complex FFTW project compiles fine and run with Orcas too.
regards heinz   

Okay that's great Heinz, I have many questions about the FFTW build attempts (both Orcas & 2005).  I may have to build this under vs2005 (with or without ICC).  I have had to stop my g++ build three times because it used up my whole machine for too long.  How long did the 100+ .C files take to build, on what kind of machine? do you still have the build log?

Did you choose many options in config.h?  Were they selected for DLL or static lib., single float, single threaded or was that in VS2005 settings or both? Are these the correct ones for seti? (been with Intel libraries until now) Under the MingW g++ build i was able to run the configure script so it chose those for me for the most part, there are still some I am not sure about...

And may have to try a build in VS2005 because ir's too heavy load for my poor little old machine ... strangely the vs2005+ICC seti builds were rather quickly complete and relativelly trouble free.

Jason

sorry forgot to attach the file

the stack overflow in parse_wu_header a history
--------------------------------------
I set a break at seti_header.cpp line 251
251     buffer += buf;
after the break buffer shows
<workunit_header>
  <name>01mr99ab.14893.2848.703400.3.151</name>
  <group_info>
    <tape_info>
      <name>01mr99ab</name>
      <start_time>2451239.5778227</start_time>
      <last_block_time>2451239.5778227</last_block_time>
      <last_block_done>2848</last_block_done>
      <missed>0</missed>
      <tape_quality>0</tape_quality>
      <sb_id>0</sb_id>
    </tape_info>
    <name>01mr99ab.14893.2848.703400.3</name>
    <data_desc>
      <start_ra>8.3103311246938</start_ra>
      <start_dec>27.91552777321</start_dec>
      <end_ra>8.3542955650728</end_ra>
      <end_dec>27.83540577987</end_dec>
      <true_angle_range>0.60488405578199</true_angle_range>
      <time_recorded>Tue Mar  2 01:52:05 1999</time_recorded>
      <time_recorded_jd>2451239.5778356</time_recorded_jd>
      <nsamples>1048576</nsamples>
      <coords>
      <coordinate_t>
        <time>2451239.5778356</time>
        <ra>8.3103311246938</ra>
        <dec>27.91552777321</dec>
      </coordinate_t>
      <coordinate_t>
        <time>2451239.5778934</time>
        <ra>8.3125413024574</ra>
        <dec>27.911197182893</dec>
      </coordinate_t>
      <coordinate_t>
        <time>2451239.5779514</time>
        <ra>8.3147416668476</ra>
        <dec>27.90693256333</dec>
      </coordinate_t>
      <coordinate_t>
        <time>2451239.5780093</time>
        <ra>8.3169450560065</ra>
        <dec>27.902621906048</dec>
      </coordinate_t>
      <coordinate_t>
        <time>2451239.578067</time>
        <ra>8.3191452260533</ra>
        <dec>27.898298175788</dec>
      </coordinate_t>
      <coordinate_t>
        <time>2451239.578125</time>
        <ra>8.3213405864986</ra>
        <dec>27.894014752898</dec>
      </coordinate_t>
      <coordinate_t>
        <time>2451239.5781829</time>
        <ra>8.3217405142933</ra>
        <dec>27.899328820752</dec>
      </coordinate_t>
      <coordinate_t>
        <time>2451239.5782407</time>
        <ra>8.3236184661523</ra>
        <dec>27.896734172274</dec>
      </coordinate_t>
      <coordinate_t>
        <time>2451239.5782985</time>
        <ra>8.325834215088</ra>
        <dec>27.892305831529</dec>
      </coordinate_t>
      <coordinate_t>
        <time>2451239.5783565</time>
        <ra>8.3280041431626</ra>
        <dec>27.888138576843</dec>
      </coordinate_t>
      <coordinate_t>
        <time>2451239.5784144</time>
        <ra>8.3302006840738</ra>
        <dec>27.883798135736</dec>
      </coordinate_t>
      <coordinate_t>
        <time>2451239.5784721</time>
        <ra>8.332395258493</ra>
        <dec>27.879438073973</dec>
      </coordinate_t>
      <coordinate_t>
        <time>2451239.5785301</time>
        <ra>8.3345999455565</ra>
        <dec>27.875038063474</dec>
      </coordinate_t>
      <coordinate_t>
        <time>2451239.5785878</time>
        <ra>8.3367837326947</ra>
        <dec>27.870706325634</dec>
      </coordinate_t>
      <coordinate_t>
        <time>2451239.5786457</time>
        <ra>8.3389835210763</ra>
        <dec>27.866287225636</dec>
      </coordinate_t>
      <coordinate_t>
        <time>2451239.5787037</time>
        <ra>8.3411633518307</ra>
        <dec>27.861977834223</dec>
      </coordinate_t>
      <coordinate_t>
        <time>2451239.5787615</time>
        <ra>8.3433656733927</ra>
        <dec>27.857498886392</dec>
      </coordinate_t>
      <coordinate_t>
        <time>2451239.5788193</time>
        <ra>8.3455480635596</ra>
        <dec>27.853130175559</dec>
      </coordinate_t>
      <coordinate_t>
        <time>2451239.5788773</time>
        <ra>8.3477402284459</ra>
        <dec>27.848707725239</dec>
      </coordinate_t>
      <coordinate_t>
        <time>2451239.5789352</time>
        <ra>8.3499266964689</ra>
        <dec>27.844286361793</dec>
      </coordinate_t>
      <coordinate_t>
        <time>2451239.5789929</time>
        <ra>8.3521124345577</ra>
        <dec>27.839838544895</dec>
      </coordinate_t>
      <coordinate_t>
        <time>2451239.5790508</time>
        <ra>8.3542955650728</ra>
        <dec>27.83540577987</dec>
      </coordinate_t>
    </coords>
  </data_desc>
  <receiver_cfg>
    <s4_id>1</s4_id>
    <name>ao1420</name>
    <beam_width>0.0829999968</beam_width>
    <center_freq>1420</center_freq>
    <latitude>18.3538056</latitude>
    <longitude>-66.7552222</longitude>
    <elevation>497</elevation>
    <diameter>168</diameter>
    <az_orientation>0</az_orientation>
    <az_corr_coeff length=49 encoding="x-csv">
      7.4,5.4,-49.1,-1091,0,0,0,0,0,0,0,0,0
    </az_corr_coeff>
    <zen_corr_coeff length=53 encoding="x-csv">
      -5146.7,66.1,14.3,321.7,0,0,0,0,0,0,0,0,0
    </zen_corr_coeff>
  </receiver_cfg>
  <recorder_cfg>
    <name>sah_ao1420</name>
    <bits_per_sample>2</bits_per_sample>
    <sample_rate>2500000</sample_rate>
    <beams>1</beams>
    <version>1.39999998</version>
  </recorder_cfg>
  <splitter_cfg>
    <version>0.170000002</version>
    <data_type>encoded</data_type>
    <fft_len>2048</fft_len>
    <ifft_len>8</ifft_len>
    <filter>fft</filter>
    <window>welsh</window>
  </splitter_cfg>
  <analysis_cfg>
    <spike_thresh>24</spike_thresh>
    <spikes_per_spectrum>1</spikes_per_spectrum>
    <gauss_null_chi_sq_thresh>2.23918462</gauss_null_chi_sq_thresh>
    <gauss_chi_sq_thresh>1.41999996</gauss_chi_sq_thresh>
    <gauss_power_thresh>3</gauss_power_thresh>
    <gauss_peak_power_thresh>3.20000005</gauss_peak_power_thresh>
    <gauss_pot_length>64</gauss_pot_length>
    <pulse_thresh>19.7613773</pulse_thresh>
    <pulse_display_thresh>0.5</pulse_display_thresh>
    <pulse_max>40960</pulse_max>
    <pulse_min>16</pulse_min>
    <pulse_fft_max>8192</pulse_fft_max>
    <pulse_pot_length>256</pulse_pot_length>
    <triplet_thresh>8.76140499</triplet_thresh>
    <triplet_max>131072</triplet_max>
    <triplet_min>16</triplet_min>
    <triplet_pot_length>256</triplet_pot_length>
    <pot_overlap_factor>0.5</pot_overlap_factor>
    <pot_t_offset>1</pot_t_offset>
    <pot_min_slew>0.00209999993</pot_min_slew>
    <pot_max_slew>0.0104999999</pot_max_slew>
    <chirp_resolution>0.333</chirp_resolution>
    <analysis_fft_lengths>262136</analysis_fft_lengths>
    <bsmooth_boxcar_length>8192</bsmooth_boxcar_length>
    <bsmooth_chunk_size>32768</bsmooth_chunk_size>
    <chirps>
    <chirp_parameter_t>
      <chirp_limit>1</chirp_limit>
      <fft_len_flags>262136</fft_len_flags>
    </chirp_parameter_t>
    <chirp_parameter_t>
      <chirp_limit>2.5</chirp_limit>
      <fft_len_flags>65528</fft_len_flags>
    </chirp_parameter_t>
  </chirps>
  <pulse_beams>1</pulse_beams>
  <max_signals>30</max_signals>
  <max_spikes>8</max_spikes>
  <max_gaussians>0</max_gaussians>
  <max_pulses>0</max_pulses>
  <max_triplets>0</max_triplets>
  <keyuniq>-12348</keyuniq>
</analysis_cfg>
<sb_id>0</sb_id>
</group_info>
<subband_desc>
  <number>151</number>
  <center>1418978879.8359</center>
  <base>1418974607.375</base>
  <sample_rate>9765.625</sample_rate>
</subband_desc>
<sb_id>0</sb_id>
</workunit_header>
------------------------------------------------------------------------
next statements are
252
253    if ( wu ) delete wu;
254    wu = new workunit( buffer ); //seti_britta:here I have a problem

if line 254 executed we come to dbgnew.cpp line 50
50{
51    /* _nh_malloc_dbg already calls _heap_alloc_dbg in a loop and calls _callnewh
52       if the allocation fails. If _callnewh returns (very likely because no
53       new handlers have been installed by the user), _nh_malloc_dbg returns NULL.
54     */
55    void *res = _nh_malloc_dbg( cb, 1, nBlockUse, szFileName, nLine );
    RTCCALLBACK(_RTC_Allocate_hook, (res, cb, 0));

    /* if the allocation fails, we throw std::bad_alloc */
    if (res == 0)
    {
        static const std::bad_alloc nomem;
        _RAISE(nomem);
    }

    return res;
}
---------------------------------------------------------------------------------------
if we execute line 55 we come to dbgheap.c line 294
294{
295        int errno_tmp = 0;
296        void * pvBlk = _nh_malloc_dbg_impl(nSize, nhFlag, nBlockUse, szFileName, nLine, &errno_tmp);

        if ( pvBlk == NULL && errno_tmp != 0 && _errno())
        {
            errno = errno_tmp; // recall, #define errno *_errno()
        }
        return pvBlk;
}
-----------------------------------------------------
if we execute line 296 we goto dbgheap.c line 232
232{{
        void * pvBlk;

        for (;
        {
            /* do the allocation
             */
239            pvBlk = _heap_alloc_dbg_impl(nSize, nBlockUse, szFileName, nLine, errno_tmp);
we came back and run into dbgheao.c line 296
296  void * pvBlk = _nh_malloc_dbg_impl(nSize, nhFlag, nBlockUse, szFileName, nLine, &errno_tmp);
if we execute we come to dbgnew.cpp line 55
55    void *res = _nh_malloc_dbg( cb, 1, nBlockUse, szFileName, nLine );
if we execute we came back to seti_header.cpp line 254
254wu = new workunit( buffer ); //seti_britta:here I have a problem
if we execute with F10  it switches to sah_gfx.cpp line 287
286        ra_string( wu.start_ra, ra_buf );
287        dec_string( wu.start_dec, dec_buf );  <--- here it crashes
        sprintf(
            buf,
            "From: %s, %s\nRecorded on: %s\nRecorded at: %s\nBase frequency: %.9f GHz",
            ra_buf,
            dec_buf,
            short_jd_string( wu.time_recorded ),
            locations[s4_id],
            wu.subband_base / 1e9 );
        }
-----------------------------------------------------------------------------------------------------------------
>   seti_boinc.exe!SAH_GRAPHICS::get_data_info_string(char * buf=0x044360d8)  Line 287 + 0x21 bytes   C++
First-chance exception at 0x7c93a321 in seti_boinc.exe: 0xC00000FD: Stack overflow.
First-chance exception at 0x7c812a7b in seti_boinc.exe: Microsoft C++ exception: std::bad_alloc at memory location 0x000337e4..
First-chance exception at 0x7c91eddd in seti_boinc.exe: 0xC0000005: Access violation writing location 0x00030ffc.
Unhandled exception at 0x7c91eddd in seti_boinc.exe: 0xC0000005: Access violation writing location 0x00030ffc.
--------------------
dec_buf ---->looks like well prepared
sb_id>
</group_info>
<subband_desc>
  <number>151</number>
  <center>1418978879.8359</center>
  <base>1418974607.375</base>
  <sample_rate>9765.625</sample_rate>
</subband_desc>
<sb_id>0</sb_id>
</workunit_header>
------------------------------------------------------------------
but wu looks bad --->
-      wu   {data_type=??? data_class=??? splitter_version=??? ...}   SETI_WU_INFO
      track_mem<SETI_WU_INFO>   {...}   track_mem<SETI_WU_INFO>
      data_type   CXX0030: Error: expression cannot be evaluated   
      data_class   CXX0030: Error: expression cannot be evaluated   
      splitter_version   CXX0030: Error: expression cannot be evaluated   
      start_ra   CXX0030: Error: expression cannot be evaluated   
      start_dec   CXX0030: Error: expression cannot be evaluated   
      end_ra   CXX0030: Error: expression cannot be evaluated   
      end_dec   CXX0030: Error: expression cannot be evaluated   
      angle_range   CXX0030: Error: expression cannot be evaluated   
      true_angle_range   CXX0030: Error: expression cannot be evaluated   
      time_recorded   CXX0030: Error: expression cannot be evaluated   
      subband_center   CXX0030: Error: expression cannot be evaluated   
      subband_base   CXX0030: Error: expression cannot be evaluated   
      subband_sample_rate   CXX0030: Error: expression cannot be evaluated   
      fft_len   CXX0030: Error: expression cannot be evaluated   
      ifft_len   CXX0030: Error: expression cannot be evaluated   
      subband_number   CXX0030: Error: expression cannot be evaluated   
+      receiver_cfg   {id=??? s4_id=??? name=0x2f3c2130 <Bad Ptr> ...}   receiver_config
      nsamples   CXX0030: Error: expression cannot be evaluated   
      bits_per_sample   CXX0030: Error: expression cannot be evaluated   
+      position_history   ...   std::_Vector_const_iterator<coordinate_t,std::allocator<coordinate_t> >
      num_positions   CXX0030: Error: expression cannot be evaluated   
+      tape_version   0x2f3c2290 <Bad Ptr>   char [16]
+      analysis_cfg   {id=??? spike_thresh=??? spikes_per_spectrum=??? ...}   analysis_config
      num_fft_lengths   CXX0030: Error: expression cannot be evaluated   
+      analysis_fft_lengths   0x2f3c2384   int [32]
      beam_width   CXX0030: Error: expression cannot be evaluated   
+      chirp_fft_table   0x2f3c2410 {MaxChirpRate=??? DoFft=0x2f3c2418 <Bad Ptr> }   ChirpFftTable_t [2]
      wu   CXX0030: Error: expression cannot be evaluated   
----------------------------------------------------------------------------------------------------------------------------
I have no idea why wu is not allocated
your hints are welcome
heinz
 

so far we measured the modified PFCASE  speeds up ~13%
some further measuring is necessary to manifest it
regards 

Its angry, but be cool... can you restore the files before the installation in any way..with systemrestore.
If not install a tool like Norton Ghost. I have installed and configured it so that once a week it makes a full backup and every day  a incremental backup of my system on a second disk. And all the time you can work....no problems.
Yesterday evening I restored a directory, after some mishappen changes on some config files of the pulstime project.
It works very well
regards heinz