Search

Experimental monoenergetic proton single-event upset (SEU) cross sections of a 65-nm low core-voltage static random access memory (SRAM) were found to be exceptionally high not only at low energies (< 3 MeV), but also at energies $>$ 3 MeV and extending up to tens of MeV. The SEU cross Section from 20-MeV protons exceeds the 200-MeV proton SEU cross Section by almost a factor of 3. Similarly, monoenergetic neutron cross sections at 14 MeV are about a factor of 3 lower than the 20-MeV proton cross section. Because of Monte Carlo (MC) simulations, it was determined that this strong enhancement is due to the proton direct ionization process as opposed to the elastic and inelastic scattering processes that dominate the SEU response above 3 MeV in other SRAMs. As shown by means of a detailed energy deposition scoring analysis, however, this does not appear to be caused by the critical charge of the SRAM being lower than the charge resulting from the average proton ionization through the linear energy transfer (LET). On the other hand, this is caused by high-energy $\delta $ -rays ( $>$ 1 keV) that can deposit their full kinetic energy within the sensitive volume (SV) of a cell despite their range being theoretically much longer than the characteristic size of the SV. Multiple Coulomb scattering events are responsible for increasing the trajectory path of the $\delta $ -rays within the SV, resulting in a six-fold increase in the probability of upset with respect to the sole electron ionization.

The objective of this thesis was to improve upon the dynamic memory allocator used in U-Blox GNSS receivers. After initial analysis, the main weakness of the currently used Buddy allocator was determined to be high fragmentation, so lowering this was the focus. To understand the problem better and find possible alternatives to Buddy, a literature survey was conducted. The survey examined the fundamental building blocks of dynamic memory allocation and how they could be combined for desired results. Using knowledge gained from the survey, a number of allocators were chosen for further examination. Among them, the most promising were two real-time allocators TLSF and a Half Fit variant called O1Heap due to their focus on integrity and more constrained heaps. After preliminary testing, two more allocators were considered as a response to the observed shortcomings of previously mentioned allocators. First was Half Fit CSTM, which makes slight modifications to the Half Fit algorithm trading constant execution time for lower fragmentation. The second allocator is Half Tree, which is a new allocator created in the scope of this research. It aims to address the weaknesses of the other considered allocators and achieve good results in all required aspects. A testing framework was designed and implemented to assist in evaluating each allocator’s suitability as an alternative for the Buddy allocator. Performance tests measured each allocator’s average and worst execution times while fragmentation tests measured the required heap size and other metrics for fragmentation. The tests were run on a real device to provide more accurate results. Testing focused on real allocation sequences (memory traces) gathered from a production device, but synthetic traces were also included. All allocators went through the same tests, and their results were compared to those of the Buddy allocator. Out of the 4 allocators considered, TLSF and Half Tree were determined to be the most suitable. Half Tree was deemed superior to TLSF because of much better performance while still having low fragmentation. Finally, Half Tree reduces fragmentation to approximately 15% to that of Buddy while only being 0–15% slower in real trace test scenarios. It requires 25% less heap for the same memory traces. Overall, the thesis objectives have been met. Half Tree allocator provides a significant improvement in memory efficiency over Buddy while maintaining comparable performance. Työn tavoitteena oli parantaa U-Bloxin GNSS laitteissa käytettyä muistiallokaattoria. Alustavan analyysin perusteella, nykyisen Buddy-allokaattorin heikkoutena oli korkea fragmentaatio, joten tutkimuksen pääpainona oli tämän alentaminen. Ongelman ja sen taustojen ymmärtämiseksi, sekä parempien allokaattorien löytämiseksi, tehtiin työn alussa kirjallisuusselvitys. Selvityksen avulla määriteltiin dynaamisen muistinhallinan perusteet ja kuinka niiden avulla voidaan rakentaa halutunlaisia allokaattoreita. Selvityksen osana tarkasteltiin useita erilaisia allokaattoreita ja päädyttiin lopulta valitsemaan 2 allokaattoria tarkempaa tutkimusta varten. Nämä olivat reaaliaika käyttöön suunnitellut ja laajalti tunnetut TLSF ja Half Fit variantti O1Heap. Alustavien kokeiden perusteella, toteutettiin kaksi uutta allokaattoria vastaamaan kokeissa havaittuihin TLSF:n ja O1Heapin heikkouksiin. Ensimmäisenä Half Fit CSTM, joka tekee pieniä muutoksia Half Fittiin vaihtaen vakioaikaisen suorituksen matalampaan fragmentaatioon. Half Tree puolestaan on uusi allokaattori suunniteltu tätä tutkimusta varten. Sen tavoitteena oli vastata muiden allokaattereiden heikkouksiin ja suoriutua hyvin kaikilla osa-alueilla. Allokaattoreiden soveltuvuutta Buddyn korvaajaksi arvioitiin kokeellisesti. Kokeiden avulla mitattiin allokaatoreidenn keskimääräistä suorituaikaa, huonointa suoritusaikaa sekä fragmentaatiota ja muistintarvetta. Koejärjestelyt suunniteltiin ja toteutettiin kohdelaitteella tulosten todenmukaisuutta ajatellen. Kokeet keskittyivät laitteelta kerättyihin oikeisiin muistijälkiin, mutta myös synteettisiä muistijälkiä hyödynnettiin. Kaikille 4 allokaattorille suoritettiin samat kokeet ja niiden tuloksia verratiin Buddy -allokaattoriin ja toisiinsa. Valituista allokaattoreista TLSF ja Half Tree osoittautuivat parhaiten soveltuviksi. TLSF on kuitenkin erittäin hidas verrattuna Buddy -allokaattoriin. Half Tree puolestaan on selvästi parempi, tarjoten edelleen matalan fragmentation, mutta ilman merkittävää kompromissia suorituskyvyssä. Half Treen fragmentaatio on noin 15% Buddy-allokaattorin fragmentaatiosta ja suorituskyky heikkenee vain 0–15% muistijäljestä riippuen. Vaaditun muistin määrässä alempi fragmentaatio tarkoittaa noin 25% parannusta. Työ täytti sille asetetut tavoitteet. Half Tree allokaattori tarjoaa merkittävän parannuksen muistitehokkuudessa Buddy-allokaattoriin verrattuna, tekemättä suuria uhrauksia suorituskyvyssä.

International audience; This study analyses the response of synchronous dynamic random access memories to neutron irradiation. Three different generations of the same device with different node sizes (63, 72, and 110 nm) were characterized under an atmospheric-like neutron spectrum at the ChipIr beamline in the Rutherford Appleton Laboratories, UK. The memories were tested with a reduced refresh rate to expose more single-event upsets and under similar conditions provided by a board specifically developed for this type of study in test facilities. The board has also been designed to be used as a nanosatellite payload in order to perform similar tests. The neutron-induced failures were studied and characterized, presenting the occurrence of single-bit upsets and stuck bits. The cross sections for each type of event and technology node show that the 110 nm model is more sensitive to neutron-induced single-event effects than the other models.

Proton direct ionization from low-energy protons has been shown to have a potentially significant impact on the accuracy of prediction methods used to calculate the upset rates of memory devices in space applications for state-of-the-art deep sub-micron technologies. The general approach nowadays is to consider a safety margin to apply over the upset rate computed from high-energy proton and heavy ion experimental data. The data reported here present a challenge to this approach. Different upset rate prediction methods are used and compared in order to establish the impact of proton direct ionization on the total upset rate. No matter the method employed the findings suggest that proton direct ionization can contribute to up to 90% of the total upset rate on average for a general selection of space orbits, with peaks of up to 99%. Such results suggest that an approach based on the characterization of the low-energy portion of the proton spectrum would be more convenient for similar technologies than the application of a general safety margin. Based on data presented here, the previously proposed margin of 5 is exceeded, by large amounts in some cases. peerReviewed

International audience; The R2E project at CERN has tested a few commercial SRAMs and a custom-designed SRAM, whose data are complementary to various scientific publications. The experimental data include low- and high-energy protons, heavy ions, thermal, intermediate- and high-energy neutrons, high-energy electrons and high-energy pions.

International audience; This paper identifies the failure modes of a commercial 130-nm ferroelectric random access memory. The devices were irradiated with heavy-ion and pulsed focused X-ray beams. Various failure modes are observed, which generate characteristic error patterns, affecting isolated bits, words, groups of pages, and sometimes entire regions of the memory array. The underlying mechanisms are discussed.

Reliable memory acquisition is essential to forensic analysis of a cyber-crime. Various methods of memory acquisition have been proposed, ranging from tools based on a dedicated hardware to software only solutions. Recently, a hypervisor-based method for memory acquisition was proposed (Qi et al., 2017; Martignoni et al., 2010). This method obtains a reliable (atomic) memory image of a running system. The method achieves this by making all memory pages non-writable until they are copied to the memory image, thus preventing uncontrolled modification of these pages. Unfortunately, the proposed method has two deficiencies: (1) the method does not support multiprocessing and (2) the method does not support modern operating systems featuring address space layout randomization (ASLR). We describe a hypervisor-based memory acquisition method that solves the two aforementioned deficiencies. We analyze the memory usage and performance of the proposed method. peerReviewed

This paper identifies the failure modes of a commercial 130-nm ferroelectric random access memory. The devices were irradiated with heavy-ion and pulsed focused X-ray beams. Various failure modes are observed, which generate characteristic error patterns, affecting isolated bits, words, groups of pages, and sometimes entire regions of the memory array. The underlying mechanisms are discussed. peerReviewed