Level 1.5 Calorimeter Trigger Data Block --------------------------------------------- Rev 20-DEC-1995 The L15 Cal Trig Data Block is made up from a number of sections. The first view shown below is the overall layout. Following this the individual sections are presented in detail. Overall Layout of the L15 Cal Trig Data Block from a Crate of L15 Cal Trig --------------------------------------- Crate Header Frame Code Section Frame Parameter Section Tool Parameter Section Local DSP Section Global DSP Section DeBug Section (meaningful for mark and pass events) If there is more than one crate of L15 Cal Trig, then each crate will have its own separate L15 Cal Trig Data Block. The crate ID in the Crate Header section will identify which crate of L15 Cal Trig this data is coming from. As is the D-Zero standard, each of the sections in the L15 Cal Trig Data Block begins with a "Longwords to Follow" Word Count. The L15 Cal Trig Data Block has a fixed format (i.e. the sections will not move around from one event to the next). The length of the sections will be constant from event to event, but might change from run to run depending on the number of terms defined. The length of the sections might also change with later revisions of the L1.5 Caltrig software, and/or new sections might be added. All sections EXCEPT for the DeBug section are of fixed length from one event to the next (although the amount of valid data in some sections, e.g. the local DSP entry list, will change from one event to the next). The DeBug section will vary in length from event to event. For NON Mark and Pass events, the DeBug section will contain a minimal amount of data. For Mark and Force Pass events, the DeBug section will be expanded. The DeBug section contains a flag which indicates which "flavor" of DeBug section is present. Crate Header ------------ The Crate Header follows the January 1990 D-Zero standard. The layout as implemented for the L15 Cal Trig is shown below. 1st Longword Header "Longwords to Follow" Word Count (6 as shown) 2nd Longword SYNC Word 3rd Longword Controller Word 4th Longword Version Number 5th Longword Revision Number 6th Longword Data Validity Word 7th Longword Mark and Force Pass Mask The 32-bit longword #1 is the header length count (not including this word) which will be 6 for a the Level 1.5 Calorimeter Trigger Crate. The longword #2 is the SYNC word. The most significant 16 bits of this longword are the least significant 16 bits of the TAS Number for this event. The Lower 16 bits are all set high (FFFF in hex). The longword #3 is the Controller Word. The most significant 8 bits are the Crate ID, which is 81 or 91 for a Level 1.5 Calorimeter Trigger Crate. The other 24 bits are currently set to 0. The longword #4 is the Version Number. As described in D0 Note 968, the most significant 3 bits have a special meaning standard to the whole experiment. Bit 31 (MSB) = Sign Bit = 0 Bit 30 = 0 For D0 = 1 For NWA Bit 29 = 0 For Data = 1 For MonteCarlo Bit 0-28 = Version Number (Integer) The 29 bit Version Number is used to describe the format of the data provided by the Level 1.5 Calorimeter Trigger. This is the key to decode the meaning, position, order and format of each word of a Level 1.5 Calorimeter Trigger Crate Data Block section of the "TRGR" Zebra Bank. This document describes the format at D0. The data blocks built according to this document will have the version number 2. Subsequent revisions and updates made to this document will result in an increase in this Version Number. The header longwords #5 thru #7 are "Detector Dependent" words as described in D0 note 968. The longword #5 is a Revision Number. The 4 bytes of this longword are each assigned an individual meaning: Bit 24-31 = L1.5 CT Hardware Revision Number (0-255) Bit 16-23 = L1.5 CT Global DSP Software revision number (0-255) Bit 8-15 = L1.5 CT Local DSP Software revision number (0-255) Bit 0-7 = L1.5 CT Engine/Readout Control Rev number (0-255) These Revision Numbers complement the Version Number appearing in longword #4 of the header. These Revision Numbers characterize objects that contributed to the content or gathering of the L1.5 Calorimeter Trigger Data Block. The Version Number (longword #4) is expected to be changed in very rare occasions. These Revision Numbers (longword #6) might be updated more often, without impact on the format of the data block, thus without changing the Version Number. The format of the 6th longword, the Data Validity Word, is shown in the following table: 0:7 Data Valid Flag 8:15 M103 L1.5 FW needed a L1.5 Decision Cycle for this event 16:23 M103 L1.5 FW Cycle terminated early 24:31 Processing Error Global Flag "Data Valid Flag" This is a very important location because it indicates whether or not ANY of the rest of THIS L15 Cal Trig Data Block is VALID. For example in an event where a crate of L15 Cal Trig was not required to evaluate any of its Terms then ALL of the rest of the L15 CT Data Block from this crate will be marked as NOT VALID. This will happen when either no L1.5 Specific Triggers fired for this event, or none of the L1.5 Specific Triggers that fired for this event required any term from this L1.5 CT Crate. A value of Zero means that the data is meaningful. A Non-Zero value marks the data as not valid. "M103 L1.5 FW needed a L1.5 Decicion Cycle for this event" means that at least one Specific Trigger requiring L1.5 confirmation, and no pure L1 Specific Triggers, have fired for this event. Recall that only if no pure L1 Specific Triggers fired does the L1.5 Framework require L1.5 Trigger subsystems to return Answers and Dones. A value of Zero in this flag means that an L1.5 Framework Decision Cycle was performed for this event. A non-Zero value means that no L1.5 Framework Decision Cycle was performed and thus no L1.5 Cal Trig Answers were returned to the L1.5 Trigger Framework. "M103 L1.5 FW Cycle terminated early" This flag shows that the L1.5 FW Decision cycle completed before this crate could send its L1.5 Terms back to the M103 L1.5 FW. This L1.5 CT Crate Returned nothing to M103 because there was, but no longer is, a L15 FW Decision Cycle that requires Terms from this L15CT crate (e.g. muon L15 confirmation was also asked for and it finished first). A value of Zero means that the L1.5 FW Cycle did not terminate early. A non-Zero value shows that the Readout Control hardware and software detected that the L1.5 Cycle was terminated before this crate could return its terms to the L1.5 Trigger Framework. "Processing Error Global Flag" This is a Global Flag showing that a processing error happened during the Local DSP, Global DSP, Engine Control or Readout Control event processing. Whenever a Processing error occurs, all Terms that need being evaluated are returned as passed to the L1.5 Framework and the error is also described in greater detail in the Frame Section below. A value of Zero means that no Error was detected. A Non-Zero value marks that an error was detected, and that the Frame Section should be examined for more details. The format of the 7th longword, the Mark and Force Pass Mask, is shown in the following table: 0:7 Mask of Terms causing this event to be marked and passed 8:31 Unused, forced to all Zeroes "Mark and Force Pass Mask" is an 8 bit mask which shows if this Event is a Mark And Force Pass Event. A bit is set for every Term that flagged this event as Mark And Pass. A Term forces the Mark and Pass decision if its 1 of N counter has expired and if it needs to be evaluated for this event. For a Mark and Pass event the answer returned to the M103 L1.5 Hardware Framework for ALL the terms being evaluated is forced TRUE. The intent is to maximize the chances the event will be accepted at the completion of the L1.5 Cycle and transferred to Level 2. If this event is not flagged as a Mark and Force Pass Event then the DeBug Section at the end of the Data Block might not contain any information that is specific to this event. The length of the Crate Header section is fixed from event to event and from run to run. Frame Code Section ------------------ The L15 Cal Trig Frame code has its own section in the L15 Cal Trig Data Block. The principal information in this section is the status of each of the L15 Cal Trig Terms in this crate. 1st Longword Frame Code Section "Longwords to Follow" Word Count (5 ?) 2nd Longword L15 CT Engine Control Starting Status 3rd Longword L15 CT Engine Control Finishing Status 4th Longword L15 CT Readout Control Finishing Status 5th Longword List of L1 Sp Trigger Fired for this Event (in bits 0:15) (upper 16 bits are zeroes) 6th Longword 0:8 Mask of Terms that were Evaluated in this L15 CT Cycle 9:15 Mask of Terms that Passed 16:23 Mask of Terms for which their Evaluation is Incomplete 24:31 Mask of Terms returned to the L1.5 Hardware Framework "List of L1 Sp Trigger Fired for this Event" This is the List of L1 Spec Trig's Fired (Trigs' #0:15 only). This list controls which set of l1.5 CalTrig Terms this crate needs to evaluate for this event. "Mask of Terms that were Evaluated in this L15 CT Cycle" This is exactly the list of Terms that needed to be evaluated during this L15 CT Cycle. "Mask of Terms that Passed" These are initialized to all zero and a bit is set to 1 only if the evaluation of the Term is complete and the term passed. "Mask of Terms for which their Evaluation is Incomplete" These are initialized to all zero and a bit is set to 1 only if the Term was to be evaluated and that the evaluation was Incomplete (e.g. saturation of Local DSP list). "Mask of Terms returned to the L1.5 Hardware Framework" This is the mask of this crate's actual Level 1.5 CalTrig Answers returned to the M103 L1.5 FW. This will match the "Mask of Terms that Passed" except for one of the following reasons. Reasons that this crate of L15 Cal Trig did not Return to the M103 L15 FW the Mask of Terms that Passed: 1. Returned nothing to M103 because there is not a L15 FW Decision Cycle under way for this event because a pure L1 Spec Trig fired for this event. 2. Returned nothing to M103 because there was but no longer is a L15 FW Decision Cycle that requires Terms from this L15CT crate (e.g. muon L15 confirmation was also asked for and it finished first) 3. If the processing of a Term that needed to be evaluated for this event is incomplete, then this will force its bit to be returned high to the M103 L15 FW. 4. The L15 CT processing cycle is a Mark and Force Pass cycle and all Terms are returned as passed to the M103 L15 FW. A L15 CT processing cycle is a Mark and Force Pass cycle if the 1 of N counter had expired for any of the Terms that needed to be evaluated for this event. The length of the Frame Code Section is fixed from event to event and from run to run. Frame Parameter Section ----------------------- The Frame Parameter Section of the L15 Cal Trig Data Block contains information that may be needed for further processing of the event (e.g. L2 Filtering or offline processing). This information includes parameter data that describes the programming of the L15 Cal Trig Frame. This parameter data has a component which is not Term-specific (called the Universal Parameters), and a component which is Term-specific (called the Frame Parameters). This section is both fixed format and fixed data from one event to the next in a given run (i.e. until the programming of the L1.5 CalTrig is changed). The length of this section might change from run to run, as fewer or more terms are defined. 1st Longword Frame Parameter Section "Longwords to Follow" Word Count Parameter Data (up to 9*16 + 1 longwords) 2nd Longword List Length 0:7 Number of Universal Blocks (always 1) 8:15 Number of Longwords per Universal Block (16) 16:23 Number of Frame Term Blocks (i.e. number of terms defined) 24:31 Number of Longwords per Frame Parameter Term Block (16) 3rd thru 18th Universal Parameter Data for this Crate 3rd Universal Parameter Block Header 0:7 Memory Map Revision Number 8:15 Memory Map Version Number 16:23 Reserved 24:31 Crate ID 4th Number of Terms defined for this Crate 5th Number of Universal Parameters defined for this Crate 6th 1st Universal Parameter for this Crate . . . 18th 13th Universal Parameter for this Crate 19th thru 34th Frame Parameter Data for this Crate's Term #0 19th Frame Parameter Data Block Header 0:7 Term Number for this Term's Frame Parameter Block (i.e. 0) 8:15 Parameter Block Type Flag (0 = Frame Parameters) 16:31 Reserved 20th Pass_1_of_N Value for this Term 21st Mask of Specific Triggers which are mapped to this Term 22nd Number of Frame Parameters for this Term 23rd 1st Frame Parameter for this Term . . . 34th 12th Frame Parameter for this Term q thru q+15 Frame Parameter Data for this Crate's next Term By definition, this exactly is the list of parameters (as obtained and derived from COOR's messages) that TCC has computed and downloaded to all the DSP's. A Level 1.5 Tool called for a particular Term receives the pointer to the first word in the parameter Term Block. The Tool will be required to check all these parameters for proper match, range and consistency. Tool Parameter Section ---------------------- The Tool Parameter Section of the L15 Cal Trig Data Block contains information that may be needed for further processing of the event (e.g. L2 Filtering or offline processing). This information includes parameter data that describes the programming of the L15 Cal Trig Tools. This parameter data has a Term- specific component dedicated to the Local Tools (called the Local Parameters), and a Term-specific component dedicated to the Global Tools (called the Global Parameters). This section is both fixed format and fixed data from one event to the next in a given run (i.e. until the programming of the L1.5 CalTrig is changed). The length of this section might change from run to run, as fewer or more terms are defined. 1st Longword Tool Parameter Section "Longwords to Follow" Word Count Parameter Data (up to 2*8*16 + 1 longwords) 2nd Longword List Length 0:7 Number of Local Term Blocks (i.e. number of terms defined) 8:15 Number of Longwords per Local Parameter Term Block (16) 16:23 Number of Global Term Blocks (i.e. number of terms defined) 24:31 Number of Longwords per Global Parameter Term Block (16) 3rd thru 18th Parameter Data for this Crate's Term #0 Local Tool 3rd Longword 0:7 Term Number for this Term's Local Parameter block (i.e. 0) 8:15 Flag indicating Parameter Block Type (1 = Local Parameters) 16:23 Reference Set Type that was used for this Term 0 : EM Ref Set FF (hex) : Tot Ref Set 24:31 Special byte indicating whether TCC detected that the Reference Set used for this Term was identical to one of the Level 1 Reference Sets. If it was, this word also shows which L1 Ref Set was used. 0,1,2 or 3 : Number of the Matching Level 1 Reference Set FF (hex) : Doesn't match any Level 1 Reference Set. 4th Tool Number (unique tool ID number) used by this Term 5th Number of Tool Parameters, i.e. # of valid longwords following 6th First Tool Dependent Parameter . 18th 13th Tool Dependent Parameter n thru n+15 Parameter Data for this crate's next Term Local DSP . . p thru p+15 Parameter Data for this Crate's Term #0 Global Tool pth Longword 0:7 Term Number for this Term's Global Parameter Block (i.e. 0) 8:15 Flag indicating Parameter Block Type (2 = Global Parameters) 16:31 Reserved p+1 Tool Number (unique tool ID number) used by this Term p+2 Number of Tool Parameters, i.e. # of valid longwords following p+3 First Tool Dependent Parameter . p+15 13th Tool Dependent Parameter q thru q+15 Parameter Data for this crate's next Term Global DSP . . By definition, this exactly is the list of parameters (as obtained and derived from COOR's messages) that TCC has computed and downloaded to all the DSP's. A Level 1.5 Tool called for a particular Term receives the pointer to the first word in the parameter Term Block. The Tool will be required to check all these parameters for proper match, range and consistency. Local DSP Section ----------------- This section contains the information generated by each of the 11 Local DSP's. This information is exactly the same as the information that the Local DSP's send to the Global DSP. Each Local DSP makes a "List of Identified Objects". Each List of Identified Objects from each Local DSP has a fixed number of entries in it. For now we plan on 8 entries per Local DSP List of Identified Objects. This is the maximum number of entries that one Local DSP can pass to the Global DSP. On a given event, the Local DSP may have filled none or only a fraction of these reserved entry blocks which are then called "valid entries". But the length of this section remains constant from one event to the next. The lists of Identified Objects in this section come from the DSP's in increasing eta order. Each of the 11 Lists of Identified Objects (one list from each of the Local DSP's) is preceded by a "DSP Header" longword which indicates the following: 0:7 Which DSP is this list coming from. (e.g. A1, A2, B3, C4 coded in hex) 8:15 How many valid Entries are there in this list 0,1,.. 8 : Number of Valid Entries FF (hex) : this Local DSP list has overflowed 16:23 How many Entries are in this list (Currently defined as 8) 24:31 How many longwords long is each entry (Currently defined as 3) Each entry is made from a fixed number of longwords. For now we plan 3 longwords per entry. It is important to note that the information in an entry in a List of Identified Objects comes from both the Local DSP Frame Code and the Local DSP Tool. The information in the first longword comes from the Frame Code and the information in the second and third longwords comes from the tool. The first longword in an entry will contain the following 4 pieces of information (one per byte): 0:7 Term Number that generated this entry 8:15 Local DSP Tool Number that generated this entry 16:23 Eta coordinate of the Trigger Tower, i.e. -20..-1 or +1..+20 (signed integer) 24:31 Phi coordinate of the Trigger Tower, i.e. 1..32 The second longword in an entry will contain the following information: 0:7 Object Type Code 8:15 Real or Mark and Pass Data 0 : real entry that passed the Local Term Algorithm FF (hex) : entry that failed the Local Term Algorithm, but was saved because this is a mark and pass event 16:31 Object Energy (units 1/4 GeV per count, this is also a signed integer quantity) The third longword in an entry will contain the following information: 0:31 Tool specific information, probably two 16 bit signed integers The format of the Local DSP Section of the L1.5 Cal Trig Data Block will be as follows (recall that this list is composed of the results of all 11 Local DSP's): 1st Longword Local_DSP Section "Longwords to Follow" Word Count 275 = (11 lists x (1 header/list + (8 entries/list x 3 LW/entry)) 2nd Header from DSP A2 eta -20:-19 3rd List Entry #1 Longword #1 4th List Entry #1 Longword #2 5th List Entry #1 Longword #3 . . 24th List Entry #8 Longword #1 25th List Entry #8 Longword #2 26th List Entry #8 Longword #3 27th Header from DSP A3 eta -18:-15 List Entries 1:8 52nd Header from DSP A4 eta -14:-11 List Entries 1:8 77th Header from DSP A1 eta -10:-7 List Entries 1:8 102nd Header from DSP B3 eta -6:-3 List Entries 1:8 127th Header from DSP B4 eta -2:+2 List Entries 1:8 152nd Header from DSP B1 eta +3:+6 List Entries 1:8 177th Header from DSP C3 eta +7:+10 List Entries 1:8 202nd Header from DSP C4 eta +11:+14 List Entries 1:8 227th Header from DSP C1 eta +15:+18 List Entries 1:8 252nd Header from DSP C2 eta +19:+20 List Entries 1:8 The length of the Local DSP Section is fixed from event to event and from run to run. Global DSP Section ------------------ This section contains information from the Global DSP. There is only one Global DSP. The principal purpose of this data is to give details about the results from the Global algorithm work done on each of the L15 Cal Trig Terms that were evaluated for this event. The following are two presentations of the details of the Global DSP Section of the Level 1.5 Calorimeter Trigger Data Block. The first presentation shows how this section of the Data Block appears when Global DSP Tool Numbers 1 through 3 are in use. Following this is a description of this section of the Data Block when Global DSP Tool Number 4 is in use. ----- Details about the Global DSP Section as it ----- ----- Appears with Global DSP Tools Numbers 1:3 ----- The Global DSP Section is composed of 32 slots for Entries. Each Entry is 2 longwords long. This format is described in the Header longword below. Note that this is a change from the Data Block Version 1. The first longword (after the longword word count longword) in the Global DSP Section is a DSP Header longword. It contains the following 4 pieces of information: 0:7 Which DSP is this list coming from (This will always be B2 in hex) 8:15 How many valid Entries are there in this list (11 for this revision of the Data Block) 16:23 How many Entries are in this list (Currently defined as 32) 24:31 How many longwords long is each entry (Currently defined as 2) Note that this is exactly the same format as the DSP Header longword in the Local DSP Section. Note also that the Number of Valid Entries is the same as the number of Local DSP's. Each of the 11 Entries contains the Counts of Identified Objects for a single Local DSP. These Entries are formatted as below: Entry, 1st longword: bits ---- 0:15 Count of Objects found for Term #0 in a Local DSP (this is a 16-bit UNSIGNED integer, but will always be in the range 0:128) 16:31 Count of Objects found for Term #1 in a Local DSP (this is a 16-bit UNSIGNED integer, but will always be in the range 0:128) Entry, 2nd longword: bits ---- 0:15 Count of Objects found for Term #2 in a Local DSP (this is a 16-bit UNSIGNED integer, but will always be in the range 0:128) 16:31 Count of Objects found for Term #3 in a Local DSP (this is a 16-bit UNSIGNED integer, but will always be in the range 0:128) The Entries will be arranged in the standard ascending eta order. The entire Global DSP Section will then be as follows: 1st Longword Global_DSP Section "Longwords to Follow" Word Count 65 = (32 entries x 2 LW/entry) + 1 header 2nd Header from DSP B2 3rd Term 0, 1 Object Count LDSP A2 eta -20:-19 4th Term 2, 3 Object Count LDSP A2 eta -20:-19 5th Term 0, 1 Object Count LDSP A3 eta -18:-15 6th Term 2, 3 Object Count LDSP A3 eta -18:-15 7th Term 0, 1 Object Count LDSP A4 eta -14:-11 8th Term 2, 3 Object Count LDSP A4 eta -14:-11 9th Term 0, 1 Object Count LDSP A1 eta -10:-7 10th Term 2, 3 Object Count LDSP A1 eta -10:-7 11th Term 0, 1 Object Count LDSP B3 eta -6:-3 12th Term 2, 3 Object Count LDSP B3 eta -6:-3 13th Term 0, 1 Object Count LDSP B4 eta -2:+2 14th Term 2, 3 Object Count LDSP B4 eta -2:+2 15th Term 0, 1 Object Count LDSP B1 eta +3:+6 16th Term 2, 3 Object Count LDSP B1 eta +3:+6 17th Term 0, 1 Object Count LDSP C3 eta +7:+10 18th Term 2, 3 Object Count LDSP C3 eta +7:+10 19th Term 0, 1 Object Count LDSP C4 eta +11:+14 20th Term 2, 3 Object Count LDSP C4 eta +11:+14 21st Term 0, 1 Object Count LDSP C1 eta +15:+18 22nd Term 2, 3 Object Count LDSP C1 eta +15:+18 23rd Term 0, 1 Object Count LDSP C2 eta +19:+20 24th Term 2, 3 Object Count LDSP C2 eta +19:+20 25th (zeros) . . 66th (zeros) When using Global DSP Tool Numbers 1:3 the length of the Global DSP Section is fixed from event to event and from run to run. ----- Details about the Global DSP Section as it ----- ----- Appears with Global DSP Tools Number 4 ----- When Global DSP Tool Number 4 is in operation the Global DSP Section will appear are follows: 1st Longword Global_DSP Section "Longwords to Follow" Word Count 97 = (16 entries x 6 Longwords/entry) + 1 header 2nd Header from DSP B2 0:7 Which DSP is this list coming from (This will always be B2 in hex) 8:15 How many valid Entries are there in this list (11 for this revision of the Data Block) 16:23 How many Entries are in this list (Currently defined as 16) 24:31 How many longwords long is each entry (Currently defined as 6) 3rd Electron Count for Terms 0 and 1 \ 4th Electron Count for Terms 2 and 3 | 5th Jet_Phi_Mask for Term 0 | 6th Jet_Phi_Mask for Term 1 | from LDSP A2 eta -20:-19 7th Jet_Phi_Mask for Term 2 | 8th Jet_Phi_Mask for Term 3 / 9rd Electron Count for Terms 0 and 1 \ 10th Electron Count for Terms 2 and 3 | 11th Jet_Phi_Mask for Term 0 | 12th Jet_Phi_Mask for Term 1 | from LDSP A3 eta -18:-15 13th Jet_Phi_Mask for Term 2 | 14th Jet_Phi_Mask for Term 3 / 15rd Electron Count for Terms 0 and 1 \ 16th Electron Count for Terms 2 and 3 | 17th Jet_Phi_Mask for Term 0 | 18th Jet_Phi_Mask for Term 1 | from LDSP A4 eta -14:-11 19th Jet_Phi_Mask for Term 2 | 20th Jet_Phi_Mask for Term 3 / 21rd Electron Count for Terms 0 and 1 \ 22th Electron Count for Terms 2 and 3 | 23th Jet_Phi_Mask for Term 0 | 24th Jet_Phi_Mask for Term 1 | from LDSP A1 eta -10:-7 25th Jet_Phi_Mask for Term 2 | 26th Jet_Phi_Mask for Term 3 / 27rd Electron Count for Terms 0 and 1 \ 28th Electron Count for Terms 2 and 3 | 29th Jet_Phi_Mask for Term 0 | 30th Jet_Phi_Mask for Term 1 | from LDSP B3 eta -6:-3 31th Jet_Phi_Mask for Term 2 | 32th Jet_Phi_Mask for Term 3 / 33rd Electron Count for Terms 0 and 1 \ 34th Electron Count for Terms 2 and 3 | 35th Jet_Phi_Mask for Term 0 | 36th Jet_Phi_Mask for Term 1 | from LDSP B4 eta -2:+2 37th Jet_Phi_Mask for Term 2 | 38th Jet_Phi_Mask for Term 3 / 39rd Electron Count for Terms 0 and 1 \ 40th Electron Count for Terms 2 and 3 | 41th Jet_Phi_Mask for Term 0 | 42th Jet_Phi_Mask for Term 1 | from LDSP B1 eta +3:+6 43th Jet_Phi_Mask for Term 2 | 44th Jet_Phi_Mask for Term 3 / 45rd Electron Count for Terms 0 and 1 \ 46th Electron Count for Terms 2 and 3 | 47th Jet_Phi_Mask for Term 0 | 48th Jet_Phi_Mask for Term 1 | from LDSP C3 eta +7:+10 49th Jet_Phi_Mask for Term 2 | 50th Jet_Phi_Mask for Term 3 / 51rd Electron Count for Terms 0 and 1 \ 52th Electron Count for Terms 2 and 3 | 53th Jet_Phi_Mask for Term 0 | 54th Jet_Phi_Mask for Term 1 | from LDSP C4 eta +11:+14 55th Jet_Phi_Mask for Term 2 | 56th Jet_Phi_Mask for Term 3 / 57rd Electron Count for Terms 0 and 1 \ 58th Electron Count for Terms 2 and 3 | 59th Jet_Phi_Mask for Term 0 | 60th Jet_Phi_Mask for Term 1 | from LDSP C1 eta +15:+18 61th Jet_Phi_Mask for Term 2 | 62th Jet_Phi_Mask for Term 3 / 63rd Electron Count for Terms 0 and 1 \ 64th Electron Count for Terms 2 and 3 | 65th Jet_Phi_Mask for Term 0 | 66th Jet_Phi_Mask for Term 1 | from LDSP C2 eta +19:+20 67th Jet_Phi_Mask for Term 2 | 68th Jet_Phi_Mask for Term 3 / 69th (zeros) . . 98th (zeros) End of the Global DSP Object List Note that the Huge Tool is implemented on the L15CT DSP's in such a way that for an Electron Type Term the Jet_Phi_Mask for that Term will be all zeros and for a Jet Type Term the Electron_Count for that Term will be all zeros. When using Global DSP Tool Number 4 the length of the Global DSP Section is fixed from event to event and from run to run. DeBug Section ------------- The DeBug Section is the only section in the L15 Cal Trig that is of variable length. It does have a fixed maximum length. This section will be minimized for "normal" (non-Mark and Force Pass) events. It will only be expanded for Mark and Pass events. That is to say, for a given run, this section will toggle between two lengths: a "minimal" length for "normal" (non-Mark and Force Pass events), and a "fully expanded" length for Mark and Force Pass events. The "fully expanded" length of this section may change from run to run as different numbers of Terms are defined or as different Tools are used. This section will also be made up from Entries. Different software components (e.g. Local DSP Frame Code, Local DSP Tool Code, etc.) in the L15 Cal Trig can generate these Entries. The general format of the DeBug Section is as follows: 1st Longword DeBug Section "Longwords to Follow" Word Count 2nd Longword Entry #1 . . pth Longword Entry #n Each Entry in the list has the following format: 1st Longword of entry: Entry Header Longword 0:7 Identify which DSP (or 68K) generated this entry (for DSPs: A1, A2, etc in hex) (for 68K: 68 in hex) 8:15 Identify entry type 16:31 Indicate how many additional longwords are part of this entry 2nd Longword of entry: 1st Element in Entry List . . n+1st Longword of entry: nth Element in Entry List Different software components of a single hardware component can generate elements in a single Entry in the DeBug Section. Note that the first longword in an Entry is in a fixed format. The additional longwords in an Entry are format-specific for a given Entry type from a given L15 Cal Trig software component. Each Entry type has a fixed length. The Entry types that have already been identified are: Present in DeBug Present Section in DeBug for Mark Section and Force for Pass "normal" Type Source Entry Type events? events? ---- ------ ---------- -------- -------- 0 68K Service CPU Service Section YES YES 1 Each Local DSP Frame Code EM and Total TT Et data YES NO 2 Each Local DSP Frame Code Reference Set Data YES NO 3 Each Local DSP Tool Code Derived Constants YES NO 4 The Global DSP Frame Code Synchronization Check YES YES Note that only the Type 0 and Type 4 Entries are present in the DeBug Section for "normal" events. The Type 0 Entry will always be the first Entry in the DeBug Section, and contains a flag (described below) which indicates which "flavor" of DeBug Section was generated for this event. The DeBug Section for Mark and Force Pass Events contains Entries of all defined Types (0 through 4). Again the Type 0 Entry will be the first Entry in the DeBug Section, and can be used to determine which "flavor" of DeBug Section is present. The Entry Type Formats for these entry types are (recall that Longword #1 of each Entry is the Entry Header and is not illustrated here): Type 0 (Service Section) ------ Longword #2: Mark and Force Pass Flag: ($FFFFFFFF: This DeBug Section corresponds to a Mark and Force Pass event [i.e. this is a "fully expanded" DeBug Section] $00000000: This DeBug Section does not correspond to a Mark and Force Pass event [i.e. this is a "minimized" DeBug Section]) This Entry will always be the FIRST Entry in the DeBug Section. Type 1 (Trigger Tower Et Data) ------ Longword #2: Trigger Tower EM Energies for relative etas n+0, n+1, n+2, n+3 at phi = 1 0:7 EM Et for TT at eta n+0 phi = 1 8:15 EM Et for TT at eta n+1 phi = 1 16:23 EM Et for TT at eta n+2 phi = 1 24:31 EM Et for TT at eta n+3 phi = 1 Longword #3: Trigger Tower EM Energies for relative etas . n+0, n+1, n+2, n+3 at phi = 2 . . Longword #33: Trigger Tower EM Energies for relative etas n+0, n+1, n+2, n+3 at phi = 32 Longword #34: Trigger Tower Total Energies for relative etas . n+0, n+1, n+2, n+3 at phi = 1 . . Longword #65: Trigger Tower Total Energies for relative etas n+0, n+1, n+2, n+3 at phi = 32 Longword #66: Trigger Tower EM Energies for relative etas . n+4, n+5, n+6, n+7 at phi = 1 . . Longword #97: Trigger Tower EM Energies for relative etas n+4, n+5, n+6, n+7 at phi = 32 Longword #98: Trigger Tower Total Energies for relative etas . n+4, n+5, n+6, n+7 at phi = 1 . . Longword #129: Trigger Tower Total Energies for relative etas n+4, n+5, n+6, n+7 at phi = 32 The energy scale of the Trigger Tower Et data is 1/4 GeV per count. Note that this Trigger Tower Et data will include pedestals. Recall also that the pedestals are eta-dependent (but uniform in phi for a given eta). The EM Et pedestals are also different from the Total Et pedestals. Recall that each Local DSP receives data corresponding to 8 Trigger Tower eta "rings." These rings are designated within a single Local DSP Node as relative eta n+0, n+1, ..., n+7. Relative eta n+0 is always mapped to the least positive (or most negative) Trigger Tower eta index for which the Local DSP Node receives Trigger Tower data. A single Local DSP Node is responsible for finding objects centered in relative eta rings n+2, n+3, n+4, and n+5 (i.e. the central 4 relative etas of the 8 relative etas for which the Local DSP Node has data). Also recall that the Local DSP Nodes "overlap" in eta. As a result, each Trigger Tower Et will appear in the Type 1 Entry from TWO Local DSP Nodes. As an example, the Et data for Trigger Towers with eta indices in the range {+1, +2, +3, +4} will appear in the Type 1 Entries from both: Local DSP B4 (which receives data from eta indices {-4, -3, -2, -1, +1, +2, +3, +4}, has relative eta n+0 mapped to Trigger Tower eta index of -4, and is responsible for finding objects centered at eta indices {-2, -1, +1, +2}) as the Trigger Tower Et data for relative etas n+4, n+5, n+6, and n+7, and Local DSP B1 (which receives data from eta indices {+1, +2, +3, +4, +5, +6, +7, +8}, has relative eta n+0 mapped to Trigger Tower eta index of +1, and is responsible for finding objects centered at eta indices {+3, +4, +5, +6}) as the Trigger Tower Et data for relative etas n+0, n+1, n+2, and n+3. Note that in the case of the two "end" Local DSP Nodes, which are each responsible for 2 eta indices (rather than 4), and have Trigger Tower data for 4 eta rings (rather than 8), "phantom" Trigger Tower data will be included in the Type 1 Entries. Thus, the Type 1 Entries from the "end" Local DSP Nodes will be the same length as Type 1 Entries from the nine other Local DSP Nodes. As an example, consider Local DSP Node C2, which receives data from eta indices {+17, +18, +19, +20}, has relative eta n+0 mapped to Trigger Tower eta index of +17, and is responsible for finding objects centered at eta indices {+19, +20}. In this Local DSP, relative etas n+4, n+5, n+6, and n+7 do not correspond to any Trigger Tower eta indices, so the Trigger Tower Et for these relative etas will be set to a fixed zero energy response of 16 [decimal]. Furthermore, consider Local DSP Node A2, which receives data from eta indices {-20, -19, -18, -17}, has relative eta n+4 mapped to Trigger Tower index -20, and is responsible for finding objects centered at eta indices {-20, -19}. In this Local DSP, relative etas n+0, n+1, n+2, and n+3 do not correspond to any Trigger Tower eta indices, so the Trigger Tower Et for these relative etas will be set to a fixed zero energy response of 16 [decimal]. Type 2 (Reference Set data) ------ Longword #2: Term Number for which this Reference Set applies Longword #3: Reference Set value for relative eta n+2, phi = 1 Longword #4: Reference Set value for relative eta n+3, phi = 1 Longword #5: Reference Set value for relative eta n+4, phi = 1 Longword #6: Reference Set value for relative eta n+5, phi = 1 Longword #7: Reference Set value for relative eta n+2, phi = 2 . . . Longword #130: Reference Set value for relative eta n+5, phi = 32 The energy scale of the Reference Set data is 1/4 GeV per count. BUT, note that the Reference Set data will include (eta-dependent) pedestals. The Reference Set data pedestal is identical to the Trigger Tower Et p pedestal. Also note that, although the Reference Set data is provided as l longwords, the maximum allowable value for Reference Set data is 255. This Reference Set data is in a format which allows it to be compared d directly to the Trigger Tower data (they have the same scale and pedestal). The L1.5 Cal Trig is defined to use a "greater than or equal to" comparison when comparing the Trigger Tower Et's to Reference Set energy. The actual implementation is performed using "strictly greater than" comparisons. This allows a Reference Set value of 255 to exclude a Trigger Tower. The TCC, when converting COOR's Reference Set specification to Reference Set values for the Local DSP's, adjusts the Reference Set value down by 1 count to mask this technical difference. Although the sense of the L1.5 Cal Trig's Reference Set data compariso Note here that each Local DSP Node only includes Reference Set data for the relative etas for which that Node is responsible. Therefore, the Reference Set data for a given Trigger Tower appears only ONCE in the DeBug Section (unlike the Trigger Tower Et which appears TWICE). As an example, the Reference Set Data for Trigger Towers with eta indices in the range {-2, -1, +1, +2} will appear in the Type 2 Entry for Local DSP B4, which is responsible for Trigger Tower eta indices in the range {-2, -1, +1, +2}. Note that in the case of the two "end" Local DSP Nodes, which are each responsible for 2 eta indices (rather than 4), "phantom" Reference Set data will be included in the Type 2 Entries. Thus, the Type 2 Entries from the "end" Local DSP Nodes will be the same length as Type 2 Entries from the nine other Local DSP Nodes. As an example, recall that Local DSP Node C2 is responsible for Trigger Tower eta indices {+19, +20}. The Type 2 Entry from Local DSP Node C2 will contain the normal Reference Set data for Trigger Tower eta indices {+19, +20} as the Reference Set data for relative etas n+2 and n+3. Relative etas n+4 and n+5 do not correspond to any Trigger Tower eta indices, so the Reference Set data for relative etas n+4 and n+5 will be set to 255 in the Type 2 Entry for Local DSP Node C2. Note that a Reference Set element of 255 results in the exclusion of the Trigger Tower (analogous to the Trigger Tower exclusion in the Level 1 Calorimeter Trigger). Note that in Local DSP Node A2, relative etas n+2 and n+3 do not correspond to any Trigger Tower eta indices. The Reference Set data for relative etas n+2 and n+3 will be set to 255 in the Type 2 Entry for Local DSP Node A2. Type 3 (Derived Constants) ------ Longword #2: Term Number for which these Derived Constants are used Longword #3: Tool Number which has generated these Derived Constants Longword #4: Derived Constant #1 . . . Longword #n+3: Derived Constant #n Recall that the number of derived constants is variable across Tools. To decode the Type 3 Entry from a given Tool, it is necessary to extract the Tool Number (longword #3 of the Entry) and from the Tool Number deduce the number and ordering of the derived constants. It is the responsibility of a Tool's author to provide a "key" to decoding the Type 3 Entry which the Tool will generate. Type 4 (Synchronization Check) ------ Longword #2: Wakeup Word given to Global DSP 0:7 Terms to Evaluate Mask 8:15 Flags Byte 16:23 Reserved 24:31 TAS Byte This entry type can be used to check that this Data Block is self- consistent. The TAS Byte should match the 8 LSBits of the TAS Number found in the Crate Header and Crate Trailer. This Entry will always be the LAST Entry in the DeBug Section. Again, the DeBug section will always contain one Type 0 Entry (from the 68K Service CPU) and one Type 4 Entry. This is true even for events that are not Mark and Force Pass events. For Mark and Force Pass Events, the DeBug Section will contain, in addition to the Type 0 and Type 4 Entries: Eleven Type 1 Entries (one from each Local DSP Node) Eleven Type 2 Entries (one from each Local DSP Node) for each Reference Set defined Eleven Type 3 Entries (one from each Local DSP Node) for each Local Tool defined For example, if only one Reference Set and one Local Tool are defined, the DeBug Section for Mark and Force Pass Events will contain: 11 Type 1 Entries, 11 Type 2 Entries, 11 Type 3 Entries. If ten Reference Sets and 6 Local Tools are defined, then the DeBug Section for Mark and Force Pass Events will contain: 11 Type 1 Entries, 11 * 10 = 110 Type 2 Entries, 11 * 6 = 66 Type 3 Entries.