RTSS'15 Results Overview

For the paper: M. Yang, A. Wieder, and B. Brandenburg, "Global Real-Time Semaphore Protocols: A Survey, Unified Analysis, and Comparison", Proceedings of the 36th IEEE Real-Time Systems Symposium, December 2015.

All results, 1440 configurations in total (with PDF figures)

Protocol and Analysis Overviews

The following reports show for each protocol and analysis where it performs best.

Overview for PIP

By pointwise comparison:

Scenarios in which PIP is at least as good as all other tested protocols/analyses (with PDF figures) [214 scenarios]
Scenarios in which PIP is better than all other tested protocols/analyses (with PDF figures) [171 scenarios]

By total area under each curve:

Scenarios in which PIP is at least as good as all other tested protocols/analyses (with PDF figures) [428 scenarios]
Scenarios in which PIP is better than all other tested protocols/analyses (with PDF figures) [383 scenarios]

Overview for PPCP

By pointwise comparison:

Scenarios in which PPCP is at least as good as all other tested protocols/analyses (with PDF figures) [41 scenarios]
Scenarios in which PPCP is better than all other tested protocols/analyses (with PDF figures) [0 scenarios]

By total area under each curve:

Scenarios in which PPCP is at least as good as all other tested protocols/analyses (with PDF figures) [41 scenarios]
Scenarios in which PPCP is better than all other tested protocols/analyses (with PDF figures) [0 scenarios]

Overview for FMLP

By pointwise comparison:

Scenarios in which FMLP is at least as good as all other tested protocols/analyses (with PDF figures) [238 scenarios]
Scenarios in which FMLP is better than all other tested protocols/analyses (with PDF figures) [227 scenarios]

By total area under each curve:

Scenarios in which FMLP is at least as good as all other tested protocols/analyses (with PDF figures) [830 scenarios]
Scenarios in which FMLP is better than all other tested protocols/analyses (with PDF figures) [815 scenarios]

Overview for FMLP+

By pointwise comparison:

Scenarios in which FMLP+ is at least as good as all other tested protocols/analyses (with PDF figures) [5 scenarios]
Scenarios in which FMLP+ is better than all other tested protocols/analyses (with PDF figures) [0 scenarios]

By total area under each curve:

Scenarios in which FMLP+ is at least as good as all other tested protocols/analyses (with PDF figures) [5 scenarios]
Scenarios in which FMLP+ is better than all other tested protocols/analyses (with PDF figures) [0 scenarios]

Overview for PRSB

By pointwise comparison:

Scenarios in which PRSB is at least as good as all other tested protocols/analyses (with PDF figures) [4 scenarios]
Scenarios in which PRSB is better than all other tested protocols/analyses (with PDF figures) [0 scenarios]

By total area under each curve:

Scenarios in which PRSB is at least as good as all other tested protocols/analyses (with PDF figures) [4 scenarios]
Scenarios in which PRSB is better than all other tested protocols/analyses (with PDF figures) [0 scenarios]

Overview for PIP-prior

By pointwise comparison:

Scenarios in which PIP-prior is at least as good as all other tested protocols/analyses (with PDF figures) [3 scenarios]
Scenarios in which PIP-prior is better than all other tested protocols/analyses (with PDF figures) [0 scenarios]

By total area under each curve:

Scenarios in which PIP-prior is at least as good as all other tested protocols/analyses (with PDF figures) [3 scenarios]
Scenarios in which PIP-prior is better than all other tested protocols/analyses (with PDF figures) [0 scenarios]

Overview for PPCP-prior

By pointwise comparison:

Scenarios in which PPCP-prior is at least as good as all other tested protocols/analyses (with PDF figures) [3 scenarios]
Scenarios in which PPCP-prior is better than all other tested protocols/analyses (with PDF figures) [0 scenarios]

By total area under each curve:

Scenarios in which PPCP-prior is at least as good as all other tested protocols/analyses (with PDF figures) [3 scenarios]
Scenarios in which PPCP-prior is better than all other tested protocols/analyses (with PDF figures) [0 scenarios]

Overview for s-ob FMLP

By pointwise comparison:

Scenarios in which s-ob FMLP is at least as good as all other tested protocols/analyses (with PDF figures) [27 scenarios]
Scenarios in which s-ob FMLP is better than all other tested protocols/analyses (with PDF figures) [16 scenarios]

By total area under each curve:

Scenarios in which s-ob FMLP is at least as good as all other tested protocols/analyses (with PDF figures) [46 scenarios]
Scenarios in which s-ob FMLP is better than all other tested protocols/analyses (with PDF figures) [33 scenarios]

Overview for s-ob OMLP

By pointwise comparison:

Scenarios in which s-ob OMLP is at least as good as all other tested protocols/analyses (with PDF figures) [11 scenarios]
Scenarios in which s-ob OMLP is better than all other tested protocols/analyses (with PDF figures) [0 scenarios]

By total area under each curve:

Scenarios in which s-ob OMLP is at least as good as all other tested protocols/analyses (with PDF figures) [12 scenarios]
Scenarios in which s-ob OMLP is better than all other tested protocols/analyses (with PDF figures) [1 scenarios]

Overview for NP-FIFO

By pointwise comparison:

Scenarios in which NP-FIFO is at least as good as all other tested protocols/analyses (with PDF figures) [5 scenarios]
Scenarios in which NP-FIFO is better than all other tested protocols/analyses (with PDF figures) [0 scenarios]

By total area under each curve:

Scenarios in which NP-FIFO is at least as good as all other tested protocols/analyses (with PDF figures) [5 scenarios]
Scenarios in which NP-FIFO is better than all other tested protocols/analyses (with PDF figures) [0 scenarios]

Overview for NP-priority

By pointwise comparison:

Scenarios in which NP-priority is at least as good as all other tested protocols/analyses (with PDF figures) [11 scenarios]
Scenarios in which NP-priority is better than all other tested protocols/analyses (with PDF figures) [0 scenarios]

By total area under each curve:

Scenarios in which NP-priority is at least as good as all other tested protocols/analyses (with PDF figures) [11 scenarios]
Scenarios in which NP-priority is better than all other tested protocols/analyses (with PDF figures) [0 scenarios]

Individual Comparisons

The following reports compare individual (groups of) analyses against other (groups of) analyses.

Comparing any s-oblivious analysis with any LP-based analysis

any s-oblivious analysis vs. any LP-based analysis:

By pointwise comparison:

At least as good: sob ≥ LP (with PDF figures) [30 scenarios]
Better than: sob > LP (with PDF figures) [25 scenarios]

By total area under each curve:

At least as good: sob ≥ LP (with PDF figures) [68 scenarios]
Better than: sob > LP (with PDF figures) [63 scenarios]

any LP-based analysis vs. any s-oblivious analysis:

By pointwise comparison:

At least as good: LP ≥ sob (with PDF figures) [1315 scenarios]
Better than: LP > sob (with PDF figures) [1310 scenarios]

By total area under each curve:

At least as good: LP ≥ sob (with PDF figures) [1378 scenarios]
Better than: LP > sob (with PDF figures) [1372 scenarios]

Pointwise incomparable:

Neither any s-oblivious analysis nor any LP-based analysis on top of each other (with PDF figures) [100 scenarios]

Comparing any s-oblivious analysis with LP-based analysis of the FMLP

any s-oblivious analysis vs. LP-based analysis of the FMLP:

By pointwise comparison:

At least as good: sob ≥ FMLP (with PDF figures) [80 scenarios]
Better than: sob > FMLP (with PDF figures) [26 scenarios]

By total area under each curve:

At least as good: sob ≥ FMLP (with PDF figures) [117 scenarios]
Better than: sob > FMLP (with PDF figures) [63 scenarios]

LP-based analysis of the FMLP vs. any s-oblivious analysis:

By pointwise comparison:

At least as good: FMLP ≥ sob (with PDF figures) [1307 scenarios]
Better than: FMLP > sob (with PDF figures) [1253 scenarios]

By total area under each curve:

At least as good: FMLP ≥ sob (with PDF figures) [1378 scenarios]
Better than: FMLP > sob (with PDF figures) [1323 scenarios]

Pointwise incomparable:

Neither any s-oblivious analysis nor LP-based analysis of the FMLP on top of each other (with PDF figures) [107 scenarios]

Comparing classic analysis of the PIP with LP-based analysis of the PIP

classic analysis of the PIP vs. LP-based analysis of the PIP:

By pointwise comparison:

At least as good: PIP-prior ≥ PIP (with PDF figures) [3 scenarios]
Better than: PIP-prior > PIP (with PDF figures) [0 scenarios]

By total area under each curve:

At least as good: PIP-prior ≥ PIP (with PDF figures) [3 scenarios]
Better than: PIP-prior > PIP (with PDF figures) [0 scenarios]

LP-based analysis of the PIP vs. classic analysis of the PIP:

By pointwise comparison:

At least as good: PIP ≥ PIP-prior (with PDF figures) [1440 scenarios]
Better than: PIP > PIP-prior (with PDF figures) [1437 scenarios]

By total area under each curve:

At least as good: PIP ≥ PIP-prior (with PDF figures) [1440 scenarios]
Better than: PIP > PIP-prior (with PDF figures) [1437 scenarios]

Pointwise incomparable:

Neither classic analysis of the PIP nor LP-based analysis of the PIP on top of each other (with PDF figures) [0 scenarios]

Comparing classic analysis of the PPCP with LP-based analysis of the PPCP

classic analysis of the PPCP vs. LP-based analysis of the PPCP:

By pointwise comparison:

At least as good: PPCP-prior ≥ PPCP (with PDF figures) [3 scenarios]
Better than: PPCP-prior > PPCP (with PDF figures) [0 scenarios]

By total area under each curve:

At least as good: PPCP-prior ≥ PPCP (with PDF figures) [3 scenarios]
Better than: PPCP-prior > PPCP (with PDF figures) [0 scenarios]

LP-based analysis of the PPCP vs. classic analysis of the PPCP:

By pointwise comparison:

At least as good: PPCP ≥ PPCP-prior (with PDF figures) [1440 scenarios]
Better than: PPCP > PPCP-prior (with PDF figures) [1437 scenarios]

By total area under each curve:

At least as good: PPCP ≥ PPCP-prior (with PDF figures) [1440 scenarios]
Better than: PPCP > PPCP-prior (with PDF figures) [1437 scenarios]

Pointwise incomparable:

Neither classic analysis of the PPCP nor LP-based analysis of the PPCP on top of each other (with PDF figures) [0 scenarios]

Comparing classic analysis of the PPCP with classic analysis of the PIP

classic analysis of the PPCP vs. classic analysis of the PIP:

By pointwise comparison:

At least as good: PPCP-prior ≥ PIP-prior (with PDF figures) [243 scenarios]
Better than: PPCP-prior > PIP-prior (with PDF figures) [151 scenarios]

By total area under each curve:

At least as good: PPCP-prior ≥ PIP-prior (with PDF figures) [254 scenarios]
Better than: PPCP-prior > PIP-prior (with PDF figures) [157 scenarios]

classic analysis of the PIP vs. classic analysis of the PPCP:

By pointwise comparison:

At least as good: PIP-prior ≥ PPCP-prior (with PDF figures) [1148 scenarios]
Better than: PIP-prior > PPCP-prior (with PDF figures) [1056 scenarios]

By total area under each curve:

At least as good: PIP-prior ≥ PPCP-prior (with PDF figures) [1287 scenarios]
Better than: PIP-prior > PPCP-prior (with PDF figures) [1190 scenarios]

Pointwise incomparable:

Neither classic analysis of the PPCP nor classic analysis of the PIP on top of each other (with PDF figures) [141 scenarios]

Comparing LP-based analysis of the PPCP with LP-based analysis of the PIP

LP-based analysis of the PPCP vs. LP-based analysis of the PIP:

By pointwise comparison:

At least as good: PPCP ≥ PIP (with PDF figures) [245 scenarios]
Better than: PPCP > PIP (with PDF figures) [0 scenarios]

By total area under each curve:

At least as good: PPCP ≥ PIP (with PDF figures) [245 scenarios]
Better than: PPCP > PIP (with PDF figures) [0 scenarios]

LP-based analysis of the PIP vs. LP-based analysis of the PPCP:

By pointwise comparison:

At least as good: PIP ≥ PPCP (with PDF figures) [1440 scenarios]
Better than: PIP > PPCP (with PDF figures) [1195 scenarios]

By total area under each curve:

At least as good: PIP ≥ PPCP (with PDF figures) [1440 scenarios]
Better than: PIP > PPCP (with PDF figures) [1195 scenarios]

Pointwise incomparable:

Neither LP-based analysis of the PPCP nor LP-based analysis of the PIP on top of each other (with PDF figures) [0 scenarios]

Comparing any FIFO-queue-based protocol with any priority-queue-based protocol

any FIFO-queue-based protocol vs. any priority-queue-based protocol:

By pointwise comparison:

At least as good: FIFO ≥ PRIO (with PDF figures) [326 scenarios]
Better than: FIFO > PRIO (with PDF figures) [313 scenarios]

By total area under each curve:

At least as good: FIFO ≥ PRIO (with PDF figures) [904 scenarios]
Better than: FIFO > PRIO (with PDF figures) [888 scenarios]

any priority-queue-based protocol vs. any FIFO-queue-based protocol:

By pointwise comparison:

At least as good: PRIO ≥ FIFO (with PDF figures) [228 scenarios]
Better than: PRIO > FIFO (with PDF figures) [215 scenarios]

By total area under each curve:

At least as good: PRIO ≥ FIFO (with PDF figures) [558 scenarios]
Better than: PRIO > FIFO (with PDF figures) [539 scenarios]

Pointwise incomparable:

Neither any FIFO-queue-based protocol nor any priority-queue-based protocol on top of each other (with PDF figures) [899 scenarios]

Comparing any FIFO-queue-based protocol with any priority-queue-based protocol

Restriction: considering only configurations with homogeneous periods.

any FIFO-queue-based protocol vs. any priority-queue-based protocol:

By pointwise comparison:

At least as good: homogen-only-FIFO ≥ PRIO (with PDF figures) [266 scenarios]
Better than: homogen-only-FIFO > PRIO (with PDF figures) [257 scenarios]

By total area under each curve:

At least as good: homogen-only-FIFO ≥ PRIO (with PDF figures) [653 scenarios]
Better than: homogen-only-FIFO > PRIO (with PDF figures) [643 scenarios]

any priority-queue-based protocol vs. any FIFO-queue-based protocol:

By pointwise comparison:

At least as good: homogen-only-PRIO ≥ FIFO (with PDF figures) [24 scenarios]
Better than: homogen-only-PRIO > FIFO (with PDF figures) [15 scenarios]

By total area under each curve:

At least as good: homogen-only-PRIO ≥ FIFO (with PDF figures) [80 scenarios]
Better than: homogen-only-PRIO > FIFO (with PDF figures) [68 scenarios]

Pointwise incomparable:

Neither any FIFO-queue-based protocol nor any priority-queue-based protocol on top of each other (with PDF figures) [439 scenarios]

Comparing any FIFO-queue-based protocol with any priority-queue-based protocol

Restriction: considering only configurations with heterogeneous periods.

any FIFO-queue-based protocol vs. any priority-queue-based protocol:

By pointwise comparison:

At least as good: hetero-only-FIFO ≥ PRIO (with PDF figures) [60 scenarios]
Better than: hetero-only-FIFO > PRIO (with PDF figures) [56 scenarios]

By total area under each curve:

At least as good: hetero-only-FIFO ≥ PRIO (with PDF figures) [251 scenarios]
Better than: hetero-only-FIFO > PRIO (with PDF figures) [245 scenarios]

any priority-queue-based protocol vs. any FIFO-queue-based protocol:

By pointwise comparison:

At least as good: hetero-only-PRIO ≥ FIFO (with PDF figures) [204 scenarios]
Better than: hetero-only-PRIO > FIFO (with PDF figures) [200 scenarios]

By total area under each curve:

At least as good: hetero-only-PRIO ≥ FIFO (with PDF figures) [478 scenarios]
Better than: hetero-only-PRIO > FIFO (with PDF figures) [471 scenarios]

Pointwise incomparable:

Neither any FIFO-queue-based protocol nor any priority-queue-based protocol on top of each other (with PDF figures) [460 scenarios]

Comparing LP-based analysis of the FMLP+ with LP-based analysis of the PRSB

LP-based analysis of the FMLP+ vs. LP-based analysis of the PRSB:

By pointwise comparison:

At least as good: FMLP+ ≥ PRSB (with PDF figures) [741 scenarios]
Better than: FMLP+ > PRSB (with PDF figures) [735 scenarios]

By total area under each curve:

At least as good: FMLP+ ≥ PRSB (with PDF figures) [1171 scenarios]
Better than: FMLP+ > PRSB (with PDF figures) [1165 scenarios]

LP-based analysis of the PRSB vs. LP-based analysis of the FMLP+:

By pointwise comparison:

At least as good: PRSB ≥ FMLP+ (with PDF figures) [143 scenarios]
Better than: PRSB > FMLP+ (with PDF figures) [137 scenarios]

By total area under each curve:

At least as good: PRSB ≥ FMLP+ (with PDF figures) [275 scenarios]
Better than: PRSB > FMLP+ (with PDF figures) [269 scenarios]

Pointwise incomparable:

Neither LP-based analysis of the FMLP+ nor LP-based analysis of the PRSB on top of each other (with PDF figures) [562 scenarios]

Comparing any PI-based protocol with any RSB-based protocol

any PI-based protocol vs. any RSB-based protocol:

By pointwise comparison:

At least as good: PI ≥ RSB (with PDF figures) [1434 scenarios]
Better than: PI > RSB (with PDF figures) [1427 scenarios]

By total area under each curve:

At least as good: PI ≥ RSB (with PDF figures) [1440 scenarios]
Better than: PI > RSB (with PDF figures) [1433 scenarios]

any RSB-based protocol vs. any PI-based protocol:

By pointwise comparison:

At least as good: RSB ≥ PI (with PDF figures) [7 scenarios]
Better than: RSB > PI (with PDF figures) [0 scenarios]

By total area under each curve:

At least as good: RSB ≥ PI (with PDF figures) [7 scenarios]
Better than: RSB > PI (with PDF figures) [0 scenarios]

Pointwise incomparable:

Neither any PI-based protocol nor any RSB-based protocol on top of each other (with PDF figures) [6 scenarios]

Comparing LP-based analysis of the FMLP with LP-based analysis of the FMLP+

LP-based analysis of the FMLP vs. LP-based analysis of the FMLP+:

By pointwise comparison:

At least as good: FMLP ≥ FMLP+ (with PDF figures) [1435 scenarios]
Better than: FMLP > FMLP+ (with PDF figures) [1419 scenarios]

By total area under each curve:

At least as good: FMLP ≥ FMLP+ (with PDF figures) [1440 scenarios]
Better than: FMLP > FMLP+ (with PDF figures) [1424 scenarios]

LP-based analysis of the FMLP+ vs. LP-based analysis of the FMLP:

By pointwise comparison:

At least as good: FMLP+ ≥ FMLP (with PDF figures) [16 scenarios]
Better than: FMLP+ > FMLP (with PDF figures) [0 scenarios]

By total area under each curve:

At least as good: FMLP+ ≥ FMLP (with PDF figures) [16 scenarios]
Better than: FMLP+ > FMLP (with PDF figures) [0 scenarios]

Pointwise incomparable:

Neither LP-based analysis of the FMLP nor LP-based analysis of the FMLP+ on top of each other (with PDF figures) [5 scenarios]

Comparing any s-oblivious analysis with any LP-based analysis of uncontrolled priority inversion

any s-oblivious analysis vs. any LP-based analysis of uncontrolled priority inversion:

By pointwise comparison:

At least as good: sob ≥ NP-any (with PDF figures) [737 scenarios]
Better than: sob > NP-any (with PDF figures) [731 scenarios]

By total area under each curve:

At least as good: sob ≥ NP-any (with PDF figures) [1308 scenarios]
Better than: sob > NP-any (with PDF figures) [1302 scenarios]

any LP-based analysis of uncontrolled priority inversion vs. any s-oblivious analysis:

By pointwise comparison:

At least as good: NP-any ≥ sob (with PDF figures) [48 scenarios]
Better than: NP-any > sob (with PDF figures) [42 scenarios]

By total area under each curve:

At least as good: NP-any ≥ sob (with PDF figures) [139 scenarios]
Better than: NP-any > sob (with PDF figures) [132 scenarios]

Pointwise incomparable:

Neither any s-oblivious analysis nor any LP-based analysis of uncontrolled priority inversion on top of each other (with PDF figures) [661 scenarios]

Comparing any LP-based analysis of uncontrolled priority inversion with classic analysis of the PIP

any LP-based analysis of uncontrolled priority inversion vs. classic analysis of the PIP:

By pointwise comparison:

At least as good: NP-any ≥ PIP-prior (with PDF figures) [133 scenarios]
Better than: NP-any > PIP-prior (with PDF figures) [130 scenarios]

By total area under each curve:

At least as good: NP-any ≥ PIP-prior (with PDF figures) [334 scenarios]
Better than: NP-any > PIP-prior (with PDF figures) [329 scenarios]

classic analysis of the PIP vs. any LP-based analysis of uncontrolled priority inversion:

By pointwise comparison:

At least as good: PIP-prior ≥ NP-any (with PDF figures) [481 scenarios]
Better than: PIP-prior > NP-any (with PDF figures) [478 scenarios]

By total area under each curve:

At least as good: PIP-prior ≥ NP-any (with PDF figures) [1112 scenarios]
Better than: PIP-prior > NP-any (with PDF figures) [1108 scenarios]

Pointwise incomparable:

Neither any LP-based analysis of uncontrolled priority inversion nor classic analysis of the PIP on top of each other (with PDF figures) [829 scenarios]

Comparing any LP-based analysis of uncontrolled priority inversion with classic analysis of the PPCP

any LP-based analysis of uncontrolled priority inversion vs. classic analysis of the PPCP:

By pointwise comparison:

At least as good: NP-any ≥ PPCP-prior (with PDF figures) [195 scenarios]
Better than: NP-any > PPCP-prior (with PDF figures) [192 scenarios]

By total area under each curve:

At least as good: NP-any ≥ PPCP-prior (with PDF figures) [393 scenarios]
Better than: NP-any > PPCP-prior (with PDF figures) [385 scenarios]

classic analysis of the PPCP vs. any LP-based analysis of uncontrolled priority inversion:

By pointwise comparison:

At least as good: PPCP-prior ≥ NP-any (with PDF figures) [464 scenarios]
Better than: PPCP-prior > NP-any (with PDF figures) [461 scenarios]

By total area under each curve:

At least as good: PPCP-prior ≥ NP-any (with PDF figures) [1057 scenarios]
Better than: PPCP-prior > NP-any (with PDF figures) [1052 scenarios]

Pointwise incomparable:

Neither any LP-based analysis of uncontrolled priority inversion nor classic analysis of the PPCP on top of each other (with PDF figures) [784 scenarios]

Comparing LP-based analysis of the FMLP with LP-based analysis of the PIP

LP-based analysis of the FMLP vs. LP-based analysis of the PIP:

By pointwise comparison:

At least as good: FMLP ≥ PIP (with PDF figures) [326 scenarios]
Better than: FMLP > PIP (with PDF figures) [312 scenarios]

By total area under each curve:

At least as good: FMLP ≥ PIP (with PDF figures) [904 scenarios]
Better than: FMLP > PIP (with PDF figures) [887 scenarios]

LP-based analysis of the PIP vs. LP-based analysis of the FMLP:

By pointwise comparison:

At least as good: PIP ≥ FMLP (with PDF figures) [229 scenarios]
Better than: PIP > FMLP (with PDF figures) [215 scenarios]

By total area under each curve:

At least as good: PIP ≥ FMLP (with PDF figures) [558 scenarios]
Better than: PIP > FMLP (with PDF figures) [539 scenarios]

Pointwise incomparable:

Neither LP-based analysis of the FMLP nor LP-based analysis of the PIP on top of each other (with PDF figures) [899 scenarios]

Comparing s-oblivious analysis of the FMLP with classic analysis of the PIP

s-oblivious analysis of the FMLP vs. classic analysis of the PIP:

By pointwise comparison:

At least as good: FMLP-sob ≥ PIP-prior (with PDF figures) [560 scenarios]
Better than: FMLP-sob > PIP-prior (with PDF figures) [557 scenarios]

By total area under each curve:

At least as good: FMLP-sob ≥ PIP-prior (with PDF figures) [978 scenarios]
Better than: FMLP-sob > PIP-prior (with PDF figures) [975 scenarios]

classic analysis of the PIP vs. s-oblivious analysis of the FMLP:

By pointwise comparison:

At least as good: PIP-prior ≥ FMLP-sob (with PDF figures) [210 scenarios]
Better than: PIP-prior > FMLP-sob (with PDF figures) [207 scenarios]

By total area under each curve:

At least as good: PIP-prior ≥ FMLP-sob (with PDF figures) [466 scenarios]
Better than: PIP-prior > FMLP-sob (with PDF figures) [462 scenarios]

Pointwise incomparable:

Neither s-oblivious analysis of the FMLP nor classic analysis of the PIP on top of each other (with PDF figures) [673 scenarios]

Comparing LP-based analysis of the FMLP with LP-based analysis of the PIP

Restriction: considering only configurations with heterogeneous periods.

LP-based analysis of the FMLP vs. LP-based analysis of the PIP:

By pointwise comparison:

At least as good: hetero-only-FMLP ≥ PIP (with PDF figures) [60 scenarios]
Better than: hetero-only-FMLP > PIP (with PDF figures) [56 scenarios]

By total area under each curve:

At least as good: hetero-only-FMLP ≥ PIP (with PDF figures) [251 scenarios]
Better than: hetero-only-FMLP > PIP (with PDF figures) [245 scenarios]

LP-based analysis of the PIP vs. LP-based analysis of the FMLP:

By pointwise comparison:

At least as good: hetero-only-PIP ≥ FMLP (with PDF figures) [204 scenarios]
Better than: hetero-only-PIP > FMLP (with PDF figures) [200 scenarios]

By total area under each curve:

At least as good: hetero-only-PIP ≥ FMLP (with PDF figures) [477 scenarios]
Better than: hetero-only-PIP > FMLP (with PDF figures) [471 scenarios]

Pointwise incomparable:

Neither LP-based analysis of the FMLP nor LP-based analysis of the PIP on top of each other (with PDF figures) [460 scenarios]

Comparing LP-based analysis of the FMLP with LP-based analysis of the PIP

Restriction: considering only configurations with homogeneous periods.

LP-based analysis of the FMLP vs. LP-based analysis of the PIP:

By pointwise comparison:

At least as good: homogen-only-FMLP ≥ PIP (with PDF figures) [266 scenarios]
Better than: homogen-only-FMLP > PIP (with PDF figures) [256 scenarios]

By total area under each curve:

At least as good: homogen-only-FMLP ≥ PIP (with PDF figures) [653 scenarios]
Better than: homogen-only-FMLP > PIP (with PDF figures) [642 scenarios]

LP-based analysis of the PIP vs. LP-based analysis of the FMLP:

By pointwise comparison:

At least as good: homogen-only-PIP ≥ FMLP (with PDF figures) [25 scenarios]
Better than: homogen-only-PIP > FMLP (with PDF figures) [15 scenarios]

By total area under each curve:

At least as good: homogen-only-PIP ≥ FMLP (with PDF figures) [81 scenarios]
Better than: homogen-only-PIP > FMLP (with PDF figures) [68 scenarios]

Pointwise incomparable:

Neither LP-based analysis of the FMLP nor LP-based analysis of the PIP on top of each other (with PDF figures) [439 scenarios]

Comparing LP-based analysis of the FMLP and PIP with LP-based analyis of any other protocol

LP-based analysis of the FMLP and PIP vs. LP-based analyis of any other protocol:

By pointwise comparison:

At least as good: FMLP-or-PIP ≥ REST (with PDF figures) [1433 scenarios]
Better than: FMLP-or-PIP > REST (with PDF figures) [1379 scenarios]

By total area under each curve:

At least as good: FMLP-or-PIP ≥ REST (with PDF figures) [1439 scenarios]
Better than: FMLP-or-PIP > REST (with PDF figures) [1385 scenarios]

LP-based analyis of any other protocol vs. LP-based analysis of the FMLP and PIP:

By pointwise comparison:

At least as good: REST ≥ FMLP-or-PIP (with PDF figures) [55 scenarios]
Better than: REST > FMLP-or-PIP (with PDF figures) [1 scenarios]

By total area under each curve:

At least as good: REST ≥ FMLP-or-PIP (with PDF figures) [55 scenarios]
Better than: REST > FMLP-or-PIP (with PDF figures) [1 scenarios]

Pointwise incomparable:

Neither LP-based analysis of the FMLP and PIP nor LP-based analyis of any other protocol on top of each other (with PDF figures) [6 scenarios]

Comparing LP-based analysis of the FMLP and PIP with LP-based analyis of any other protocol (other than the PPCP)

LP-based analysis of the FMLP and PIP vs. LP-based analyis of any other protocol (other than the PPCP):

By pointwise comparison:

At least as good: FMLP-or-PIP ≥ REST2 (with PDF figures) [1433 scenarios]
Better than: FMLP-or-PIP > REST2 (with PDF figures) [1420 scenarios]

By total area under each curve:

At least as good: FMLP-or-PIP ≥ REST2 (with PDF figures) [1440 scenarios]
Better than: FMLP-or-PIP > REST2 (with PDF figures) [1427 scenarios]

LP-based analyis of any other protocol (other than the PPCP) vs. LP-based analysis of the FMLP and PIP:

By pointwise comparison:

At least as good: REST2 ≥ FMLP-or-PIP (with PDF figures) [13 scenarios]
Better than: REST2 > FMLP-or-PIP (with PDF figures) [0 scenarios]

By total area under each curve:

At least as good: REST2 ≥ FMLP-or-PIP (with PDF figures) [13 scenarios]
Better than: REST2 > FMLP-or-PIP (with PDF figures) [0 scenarios]

Pointwise incomparable:

Neither LP-based analysis of the FMLP and PIP nor LP-based analyis of any other protocol (other than the PPCP) on top of each other (with PDF figures) [7 scenarios]

Comparing any LP-based analysis of uncontrolled priority inversion with any other protocol/analysis

any LP-based analysis of uncontrolled priority inversion vs. any other protocol/analysis:

By pointwise comparison:

At least as good: NP-any ≥ REST (with PDF figures) [13 scenarios]
Better than: NP-any > REST (with PDF figures) [0 scenarios]

By total area under each curve:

At least as good: NP-any ≥ REST (with PDF figures) [13 scenarios]
Better than: NP-any > REST (with PDF figures) [0 scenarios]

any other protocol/analysis vs. any LP-based analysis of uncontrolled priority inversion:

By pointwise comparison:

At least as good: REST ≥ NP-any (with PDF figures) [1439 scenarios]
Better than: REST > NP-any (with PDF figures) [1426 scenarios]

By total area under each curve:

At least as good: REST ≥ NP-any (with PDF figures) [1440 scenarios]
Better than: REST > NP-any (with PDF figures) [1427 scenarios]

Pointwise incomparable:

Neither any LP-based analysis of uncontrolled priority inversion nor any other protocol/analysis on top of each other (with PDF figures) [1 scenarios]