Precision in interface with FastJet

[gavinsalam]

Much of the software uses floats. However FastJet uses doubles. Rounding errors can be non-negligible if one sends the float px,py,pz,E directly to FastJet (the problem is relevant mainly for the mass of PseudoJets). Better in communicating with FastJet to take the floats for px,py,pz,m, convert these to double, calculate E in double and then use the new double precision px,py,pz,E values to create the Pseudojets.

The communication to FastJet is illustrated in PR #107, though one might also want to think about how the two-way communication works.

[clementhelsens]

To be checked later, from the raw reconstructedParticle collection, sqrt(px^2+py&2+pz^2+m^2) does not match the energy for part of the collection. Could come from Delphes interface, needs some investigations

root [8] events->Scan("sqrt(RP_px^2+RP_py^2+RP_pz^2+RP_m^2):RP_e:sqrt(RP_px^2+RP_py^2+RP_pz^2+RP_m^2)-RP_e")
***********************************************************
*    Row   * Instance * sqrt(RP_p *      RP_e * sqrt(RP_p *
***********************************************************
*        0 *        0 * 1.3572080 * 1.3500125 * 0.0071955 *
*        0 *        1 * 0.4067147 * 0.3820170 * 0.0246977 *
*        0 *        2 * 0.9756593 * 0.9656247 * 0.0100345 *
*        0 *        3 * 2.3657328 * 2.3616120 * 0.0041207 *
*        0 *        4 * 2.7254972 * 2.7219212 * 0.0035760 *
*        0 *        5 * 0.8448534 * 0.8332451 * 0.0116083 *
*        0 *        6 * 5.3700246 * 5.3682103 * 0.0018143 *
*        0 *        7 * 5.5412582 * 5.5395002 * 0.0017579 *
*        0 *        8 * 1.0097689 * 1.0000767 * 0.0096922 *
*        0 *        9 * 1.4921318 * 1.4855899 * 0.0065418 *
*        0 *       10 * 9.1755610 * 9.1744995 * 0.0010615 *
*        0 *       11 * 8.0808394 * 8.0796337 * 0.0012057 *
*        0 *       12 * 0.3990841 * 0.3738826 * 0.0252014 *
*        0 *       13 * 1.4127064 * 1.4127063 * 4.992e-08 *
*        0 *       14 * 1.1802264 * 1.1802263 * 6.909e-08 *
*        0 *       15 * 1.8873737 * 1.8873739 * -1.72e-07 *
*        0 *       16 * 0.8118204 * 0.8118203 * 2.433e-08 *
*        0 *       17 * 2.7451751 * 2.7451751 * 5.101e-08 *
*        0 *       18 * 1.1429564 * 1.1429565 * -1.19e-08 *
*        0 *       19 * 0.5041071 * 0.5041071 * -8.98e-09 *
*        0 *       20 * 7.8634692 * 7.8634691 * 1.190e-07 *
*        0 *       21 * 1.8220118 * 1.8220118 * 3.545e-09 *
*        0 *       22 * 2.0057414 * 2.0057413 * 4.370e-08 *
*        0 *       23 * 0.8879450 * 0.8879450 * -3.14e-08 *
*        0 *       24 * 0.9413895 * 0.9413895 * -5.87e-09 *

[pjanot]

Salut Clément, When you compute sqrt(RP_px^2+RP_py^2+RP_pz^2+RP_m^2-RP_e^2) instead of sqrt(RP_px^2+RP_py^2+RP_pz^2+RP_m^2)-RP_e, at least for the first two line, you obtain 0.139570 (i.e., the charged pion mass). It seems that RP_e = RP_p in DELPHES for all particles. Patrick On 20 Sep 2021, at 20:42, Clement Helsens ***@***.******@***.***>> wrote: To be checked later, from the raw reconstructedParticle collection, sqrt(px^2+py&2+pz^2+m^2) does not match the energy for part of the collection. Could come from Delphes interface, needs some investigations root [8] events->Scan("sqrt(RP_px^2+RP_py^2+RP_pz^2+RP_m^2):RP_e:sqrt(RP_px^2+RP_py^2+RP_pz^2+RP_m^2)-RP_e") *********************************************************** * Row * Instance * sqrt(RP_p * RP_e * sqrt(RP_p * *********************************************************** * 0 * 0 * 1.3572080 * 1.3500125 * 0.0071955 * * 0 * 1 * 0.4067147 * 0.3820170 * 0.0246977 * * 0 * 2 * 0.9756593 * 0.9656247 * 0.0100345 * * 0 * 3 * 2.3657328 * 2.3616120 * 0.0041207 * * 0 * 4 * 2.7254972 * 2.7219212 * 0.0035760 * * 0 * 5 * 0.8448534 * 0.8332451 * 0.0116083 * * 0 * 6 * 5.3700246 * 5.3682103 * 0.0018143 * * 0 * 7 * 5.5412582 * 5.5395002 * 0.0017579 * * 0 * 8 * 1.0097689 * 1.0000767 * 0.0096922 * * 0 * 9 * 1.4921318 * 1.4855899 * 0.0065418 * * 0 * 10 * 9.1755610 * 9.1744995 * 0.0010615 * * 0 * 11 * 8.0808394 * 8.0796337 * 0.0012057 * * 0 * 12 * 0.3990841 * 0.3738826 * 0.0252014 * * 0 * 13 * 1.4127064 * 1.4127063 * 4.992e-08 * * 0 * 14 * 1.1802264 * 1.1802263 * 6.909e-08 * * 0 * 15 * 1.8873737 * 1.8873739 * -1.72e-07 * * 0 * 16 * 0.8118204 * 0.8118203 * 2.433e-08 * * 0 * 17 * 2.7451751 * 2.7451751 * 5.101e-08 * * 0 * 18 * 1.1429564 * 1.1429565 * -1.19e-08 * * 0 * 19 * 0.5041071 * 0.5041071 * -8.98e-09 * * 0 * 20 * 7.8634692 * 7.8634691 * 1.190e-07 * * 0 * 21 * 1.8220118 * 1.8220118 * 3.545e-09 * * 0 * 22 * 2.0057414 * 2.0057413 * 4.370e-08 * * 0 * 23 * 0.8879450 * 0.8879450 * -3.14e-08 * * 0 * 24 * 0.9413895 * 0.9413895 * -5.87e-09 * — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub<#108 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ABWI4QP45EO26446YL77UQLUC56BZANCNFSM5EMLUUNA>. Triage notifications on the go with GitHub Mobile for iOS<https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android<https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.

[clementhelsens]

salut @pjanot , welcome back on github!
you are right, in delphes all charged particles have the pion mass and neutrals have a null mass, see below, we have some differences in the calculation that might explain why the two pseudo_jet building methods are not fully identical. It's important we keep this in mind and keep track of this issue in case precision analyses are sensitive to it as pointed out by @gavinsalam

root [4] events->Scan("sqrt(RP_px^2+RP_py^2+RP_pz^2+RP_m^2):RP_e:sqrt(RP_px^2+RP_py^2+RP_pz^2+RP_m^2)-RP_e:sqrt(RP_px^2+RP_py^2+RP_pz^2+RP_m^2-RP_e^2):RP_q:RP_m:RP_m-sqrt(RP_px^2+RP_py^2+RP_pz^2+RP_m^2-RP_e^2)")
***********************************************************************************************************
*    Row   * Instance * sqrt(RP_p *      RP_e * sqrt(RP_p * sqrt(RP_p *      RP_q *      RP_m * RP_m-sqrt *
***********************************************************************************************************
*        0 *        0 * 1.3572080 * 1.3500125 * 0.0071955 * 0.1395703 *         1 * 0.1395703 * 5.301e-08 *
*        0 *        1 * 0.4067147 * 0.3820170 * 0.0246977 * 0.1395703 *        -1 * 0.1395703 * 1.645e-09 *
*        0 *        2 * 0.9756593 * 0.9656247 * 0.0100345 * 0.1395707 *         1 * 0.1395703 * -3.61e-07 *
*        0 *        3 * 2.3657328 * 2.3616120 * 0.0041207 * 0.1395715 *         1 * 0.1395703 * -1.20e-06 *
*        0 *        4 * 2.7254972 * 2.7219212 * 0.0035760 * 0.1395709 *         1 * 0.1395703 * -5.73e-07 *
*        0 *        5 * 0.8448534 * 0.8332451 * 0.0116083 * 0.1395705 *        -1 * 0.1395703 * -1.28e-07 *
*        0 *        6 * 5.3700246 * 5.3682103 * 0.0018143 * 0.1395819 *         1 * 0.1395703 * -1.15e-05 *
*        0 *        7 * 5.5412582 * 5.5395002 * 0.0017579 * 0.1395696 *        -1 * 0.1395703 * 7.098e-07 *
*        0 *        8 * 1.0097689 * 1.0000767 * 0.0096922 * 0.1395703 *         1 * 0.1395703 * 6.379e-08 *
*        0 *        9 * 1.4921318 * 1.4855899 * 0.0065418 * 0.1395697 *         1 * 0.1395703 * 6.470e-07 *
*        0 *       10 * 9.1755610 * 9.1744995 * 0.0010615 * 0.1395670 *        -1 * 0.1395703 * 3.322e-06 *
*        0 *       11 * 8.0808394 * 8.0796337 * 0.0012057 * 0.1395885 *         1 * 0.1395703 * -1.81e-05 *
*        0 *       12 * 0.3990841 * 0.3738826 * 0.0252014 * 0.1395703 *        -1 * 0.1395703 * 5.392e-08 *
*        0 *       13 * 1.4127064 * 1.4127063 * 4.992e-08 * 0.0003755 *         0 *         0 * -0.000375 *
*        0 *       14 * 1.1802264 * 1.1802263 * 6.909e-08 * 0.0004038 *         0 *         0 * -0.000403 *
*        0 *       15 * 1.8873737 * 1.8873739 * -1.72e-07 * 0.0008062 *         0 *         0 * -0.000806 *
*        0 *       16 * 0.8118204 * 0.8118203 * 2.433e-08 * 0.0001987 *         0 *         0 * -0.000198 *
*        0 *       17 * 2.7451751 * 2.7451751 * 5.101e-08 * 0.0005292 *         0 *         0 * -0.000529 *
*        0 *       18 * 1.1429564 * 1.1429565 * -1.19e-08 * 0.0001651 *         0 *         0 * -0.000165 *
*        0 *       19 * 0.5041071 * 0.5041071 * -8.98e-09 * 9.515e-05 *         0 *         0 * -9.51e-05 *
*        0 *       20 * 7.8634692 * 7.8634691 * 1.190e-07 * 0.0013681 *         0 *         0 * -0.001368 *
*        0 *       21 * 1.8220118 * 1.8220118 * 3.545e-09 * 0.0001136 *         0 *         0 * -0.000113 *
*        0 *       22 * 2.0057414 * 2.0057413 * 4.370e-08 * 0.0004187 *         0 *         0 * -0.000418 *
*        0 *       23 * 0.8879450 * 0.8879450 * -3.14e-08 * 0.0002364 *         0 *         0 * -0.000236 *
*        0 *       24 * 0.9413895 * 0.9413895 * -5.87e-09 * 0.0001051 *         0 *         0 * -0.000105 *