Search

International audience; In a super-scalar architecture, the scheduler dynamically assigns micro-operations (µOPs) to execution ports. The port mapping of an architecture describes how an instruction decomposes into µOPs and lists for each µOP the set of ports it can be mapped to. It is used by compilers and performance debugging tools to characterize the performance throughput of a sequence of instructions repeatedly executed as the core component of a loop. This paper introduces a dual equivalent representation: The resource mapping of an architecture is an abstract model where, to be executed, an instruction must use a set of abstract resources, themselves representing combinations of execution ports. For a given architecture, finding a port mapping is an important but difficult problem. Building a resource mapping is a more tractable problem and provides a simpler and equivalent model. This paper describes Palmed, a tool that automatically builds a resource mapping for pipelined, super-scalar, out-of-order CPU architectures. Palmed does not require hardware performance counters, and relies solely on runtime measurements. We evaluate the pertinence of our dual representation for throughput modeling by extracting a representative set of basic-blocks from the compiled binaries of the SPEC CPU 2017 benchmarks. We compared the throughput predicted by existing machine models to that produced by Palmed, and found comparable accuracy to state-of-the art tools, achieving sub-10 % mean square error rate on this workload on Intel's Skylake microarchitecture.

In a super-scalar architecture, the scheduler dynamically assigns micro-operations ($\mu$OPs) to execution ports. The port mapping of an architecture describes how an instruction decomposes into $\mu$OPs and lists for each $\mu$OP the set of ports it can be mapped to. It is used by compilers and performance debugging tools to characterize the performance throughput of a sequence of instructions repeatedly executed as the core component of a loop. This paper introduces a dual equivalent representation: The resource mapping of an architecture is an abstract model where, to be executed, an instruction must use a set of abstract resources, themselves representing combinations of execution ports. For a given architecture, finding a port mapping is an important but difficult problem. Building a resource mapping is a more tractable problem and provides a simpler and equivalent model. This paper describes Palmed, a tool that automatically builds a resource mapping for pipelined, super-scalar, out-of-order CPU architectures. Palmed does not require hardware performance counters, and relies solely on runtime measurements. We evaluate the pertinence of our dual representation for throughput modeling by extracting a representative set of basic-blocks from the compiled binaries of the SPEC CPU 2017 benchmarks. We compared the throughput predicted by existing machine models to that produced by Palmed, and found comparable accuracy to state-of-the art tools, achieving sub-10 % mean square error rate on this workload on Intel's Skylake microarchitecture.

Additional file 1. Details of the synthesis and characterization of the compounds, and structures of all the assayed molecules.

This paper describes Palmed, a tool that automatically builds a resource mapping, a performance model for pipelined, super-scalar, out-of-order CPU architectures. Resource mappings describe the execution of a program by assigning instructions in the program to abstract resources. They can be used to predict the throughput of basic blocks or as a machine model for the backend of an optimizing compiler. Palmed does not require hardware performance counters, and relies solely on runtime measurements to construct resource mappings. This allows it to model not only execution port usage, but also other limiting resources, such as the frontend or the reorder buffer. Also, thanks to a dual representation of resource mappings, our algorithm for constructing mappings scales to large instruction sets, like that of x86. We evaluate the algorithmic contribution of the paper in two ways. First by showing that our approach can reverse engineering an accurate resource mapping from an idealistic performance model produced by an existing port-mapping. We also evaluate the pertinence of our dual representation, as opposed to the standard port-mapping, for throughput modeling by extracting a representative set of basic-blocks from the compiled binaries of the Spec CPU 2017 benchmarks and comparing the throughput predicted by existing machine models to that produced by Palmed.

The polyhedral model has been successfully used in production compilers. Nevertheless, only a very restricted class of applications can benefit from it. Recent proposals investigated how runtime information could be used to apply polyhedral optimization on applications that do not statically fit the model. In this work, we go one step further in that direction. We propose a dynamic analysis that builds a compact polyhedral representation from a program execution. It is able to accurately detect affine dependencies and fixed-stride memory accesses in programs. The analysis scales to real-life applications, which often include some non-affine dependencies and accesses in otherwise affine code. This is enabled by a safe fine-grain polyhedral over-approximation mechanism applied to each analyzed expression. We evaluate our analysis on the entire Rodinia benchmark suite, enabling accurate feedback about potential for complex polyhedral transformations.; Le modèle polyédrique est aujourd'hui utilisé à grande échelle via son intégration dans des compilateurs très largement utilisés. Néanmoins, seule une classe très restreinte de programmes peut en bénéficier. Des travaux récents ont montré comment des informations provenant d'une exécution du programme pouvaient être utilisées afin d'étendre la portée dumodèle polyédrique. Ce travail s'inscrit dans ce contexte d'analyse dynamique de programmes pour appliquer le modèle polyédrique plus largement. Nous proposons une analyse dynamique capable de construire une représentation polyédrique d'un programme à partir d'une éxecution instrumentée. Cette analyse détecte de façon précise les dépendances affines ainsi que les accès mémoire avec incréments constants présents dans le programme. Notre analyse passe à l'échellesur de vraies applications qui contiennent souvent quelques dépendances et accès mémoire non affines. Ce passage à l'échelle est possible grâce à un mécanisme de sur-approximation. Nous évaluons notre analyse sur la suite de benchmarks Rodinia en montrant quel est le retour fourni à l'utilisateur en ce qui concerne de potentielles transformations polyédriques.

The polyhedral model has been successfully used in production compilers. Nevertheless, only a very restricted class of applications can benefit from it. Recent proposals investigated how runtime information could be used to apply polyhedral optimization on applications that do not statically fit the model. In this work, we go one step further in that direction. We propose a dynamic analysis that builds a compact polyhedral representation from a program execution. It is able to accurately detect affine dependencies and fixed-stride memory accesses in programs. The analysis scales to real-life applications, which often include some non-affine dependencies and accesses in otherwise affine code. This is enabled by a safe fine-grain polyhedral over-approximation mechanism applied to each analyzed expression. We evaluate our analysis on the entire Rodinia benchmark suite, enabling accurate feedback about potential for complex polyhedral transformations.; Le modèle polyédrique est aujourd'hui utilisé à grande échelle via son intégration dans des compilateurs très largement utilisés. Néanmoins, seule une classe très restreinte de programmes peut en bénéficier. Des travaux récents ont montré comment des informations provenant d'une exécution du programme pouvaient être utilisées afin d'étendre la portée dumodèle polyédrique. Ce travail s'inscrit dans ce contexte d'analyse dynamique de programmes pour appliquer le modèle polyédrique plus largement. Nous proposons une analyse dynamique capable de construire une représentation polyédrique d'un programme à partir d'une éxecution instrumentée. Cette analyse détecte de façon précise les dépendances affines ainsi que les accès mémoire avec incréments constants présents dans le programme. Notre analyse passe à l'échellesur de vraies applications qui contiennent souvent quelques dépendances et accès mémoire non affines. Ce passage à l'échelle est possible grâce à un mécanisme de sur-approximation. Nous évaluons notre analyse sur la suite de benchmarks Rodinia en montrant quel est le retour fourni à l'utilisateur en ce qui concerne de potentielles transformations polyédriques.

Feline immunodeficiency virus (FIV) is a member of the Retroviridae family. It is the causative agent of an acquired immunodeficiency syndrome (AIDS) in cats and wild felines. Its capsid protein (CA) drives the assembly of the viral particle, which is a critical step in the viral replication cycle. Here, the first atomic structure of full-length FIV CA to 1.67 Å resolution is determined. The crystallized protein exhibits an original tetrameric assembly, composed of dimers which are stabilized by an intermolecular disulfide bridge induced by the crystallogenesis conditions. The FIV CA displays a standard α-helical CA topology with two domains, separated by a linker shorter than other retroviral CAs. The β-hairpin motif at its amino terminal end, which interacts with nucleotides in HIV-1, is unusually long in FIV CA. Interestingly, this functional β-motif is formed in this construct in the absence of the conserved N-terminal proline. The FIV CA exhibits a cis Arg–Pro bond in the CypA-binding loop, which is absent in known structures of lentiviral CAs. This structure represents the first tri-dimensional structure of a functional, full-length FIV CA.

Anti-p17 antibodies are able to neutralize human immunodeficiency virus (HIV) entry in a mouse model. In this study, we identified a region of sequence similarity between the epitopes of anti-p17 neutralizing antibodies and anti-gp41 neutralizing 2F5 antibody and verified cross-reactivity between p17 and 2F5 in vitro. The p17 sequence was modified to increase sequence identity between the p17 and 2F5 epitopes, which resulted in enhanced cross-reactivity in vitro. Immunogenicity of wild-type and modified p17 was characterized in a rabbit model. Both wild-type and mutated p17 induced anti-gp41 responses in rabbits; sera from these animals reacted with gp41 from different HIV clades. Moreover, introduction of the 2F5 sequence in p17 resulted in induction of antibodies with partially neutralizing activity. Based upon these data, we suggest that the natural cross-reactivity between HIV-1 p17 protein and 2F5 antibody can be exploited to induce antibodies with neutralizing activity in an animal model.