|
[Rivet] alpgen+herwig in agile 1.2.2Gavin Hesketh gavin.hesketh at cern.chThu Apr 28 17:54:15 BST 2011
Hello, Although I'm a bit worried I might be gaining a reputation as an alpgen defender, there is a bug in the new agile which means alpgen+herwig will not run correctly. There are two small tweaks, calling: > FC_HWEVNT.IERROR = -1; and > continue; (instead of return) when UPVETO fails. Calling IERROR ensures the event count is correct at the end of the run (mostly cosmetic, unless people use this to normalise). Calling continue ensures agile does not spit out empty events if upveto fails (which can be very often) - this is more serious, and basically makes it unusable as (eg) rivet aborts on seeing an empty event. The two patched files are attached, the diff is below. cheers, Gavin > diff AlpGenFHerwigJimmy.cc build/AGILe-1.2.2/src/AlpGen/AlpGenFHerwigJimmy.cc 70d70 < FC_HWEVNT.IERROR = -1; 72c72 < continue; --- > return; > diff AlpGenFHerwig.cc build/AGILe-1.2.2/src/AlpGen/AlpGenFHerwig.cc 57,70c57,65 < < while (true) { < FC_HWUINE(); // Initialize event < FC_HWEPRO(); // Generate hard subprocess < // Call the matching routine < int imatch = 0; < FC_UPVETO(&imatch); < if (imatch != 0) { < MSG_DEBUG("Event killed by UPVETO"); < FC_HWEVNT.IERROR = -1; < FC_HWUFNE(); < continue; < } < break; --- > FC_HWUINE(); // Initialize event > FC_HWEPRO(); // Generate hard subprocess > // Call the matching routine > int imatch = 0; > FC_UPVETO(&imatch); > if (imatch != 0) { > MSG_DEBUG("Event killed by UPVETO"); > FC_HWUFNE(); > return; -------------- next part -------------- A non-text attachment was scrubbed... Name: AlpGenFHerwig.cc Type: text/x-c++src Size: 2502 bytes Desc: not available URL: <http://www.hepforge.org/lists-archive/rivet/attachments/20110428/9cdbbf71/attachment.cc> -------------- next part -------------- A non-text attachment was scrubbed... Name: AlpGenFHerwigJimmy.cc Type: text/x-c++src Size: 3684 bytes Desc: not available URL: <http://www.hepforge.org/lists-archive/rivet/attachments/20110428/9cdbbf71/attachment-0001.cc>
More information about the Rivet mailing list |