#gpu

anonymiss@despora.de

The #AI #supply chain:

"It makes visible the #connection between an #engineer training an #algorithm in the #UK, a miner extracting #tantalum in #Kazakhistan, an engineer in #Mexico working in a #data centre, a #worker in #Taiwan #manufacturing GPUs and a worker in #Kenya dismantling e-waste"

source: https://twitter.com/ana_valdi/status/1747200486392950785

#economy #technology #supplyChain #resources #globalization #internet #software #hardware #gpu #labour #map #news

california@diaspora.permutationsofchaos.com

Tinker board with #AI support and #GPIO connector compatible with #RaspberryPi

URI: https://hub.libre.computer/t/2023-09-25-libre-computer-aml-a311d-cc-alta-ai-sbc-announcement/2905

#AML-A311D-CC features:

  • 4 M Cores
  • 2 E Cores
  • 4 M #GPU Cores
  • 5+ TOPs #NPU Cores (AI)
  • Up to 4GB LPDDR4X
  • #USB Type C Power + 2.0 Data Dual Role
  • #HDMI 2.1
  • 3.5mm Jack with CVBS and Analog Stereo Audio Output
  • Gigabit #Ethernet with WOL
  • 4 USB Type-A 3.0 Hub
  • PoE Connector
  • 40-Pin GPIO Connector
  • IR Receiver Sensor
  • DSI 4-Lane 22-Pin Connector up to 1080P
  • CSI 4-Lane 22-Pin Connector up to 8MP with 2 Cameras
  • 16MB #OpenSource UEFI BIOS
  • #eMMC 5.x Slim Connector
  • #MicroSD Card Slot with UHS SDR104 Support
  • Price $45 (2GB Ram version)

If you want to know how to use the AI module (NPU):
* https://www.cnx-software.com/2020/01/13/getting-started-with-amlogic-npu-on-khadas-vim3-vim3l/
* https://www.cnx-software.com/2023/11/09/libre-computer-aml-a311d-cc-alta-sbc-features-amlogic-a311d-ai-processor/
* https://forum.khadas.com/c/khadas-vim3/30

board

#hardware #iot #technology #ARM

danie10@squeet.me

3 reasons to ditch Nvidia for AMD in 20233 reasons to ditch Nvidia for AMD in 2023

Black coloured Graphics Processor Unit with two fans on the front and the name RADEON on it. Behind it is a black box with words AMD Radeon RX7900XT on it.
Nvidia’s GPUs are considered second to none in the enthusiast PC space, and there are plenty of convincing reasons to go with Team Green for your next build. XDA Developers highlighted a few of those reasons to consider an Nvidia GPU over an AMD one recently, going over things like DLSS and the raw performance of RTX GPUs. AMD graphics cards, however, have also come a long way, and they aren’t trailing too far behind in 2023.

In fact, there are some good reasons to consider them over Nvidia’s options, and you certainly can’t count them out of the race. If you are in the market to buy a new graphics card and are split between AMD and Nvidia, then here are a few reasons why you should consider an AMD GPU for your build.

OK, admittedly two of the reasons are related, so this could be more of two good reasons. I did opt to go with an AMD Ryzen 7 for my last CPU I bought (my first non-Intel in decades), and I’ve been very happy with that choice. I only realised a month later, after I’d bought an Nvidia GPU, that I never really took a serious look at the AMD GPUs.

I’m certainly going to do so next time I buy a GPU (I don’t buy them with every PC upgrade I do). In my case, too, I’m using Linux, so I really don’t get to use some of those additional extra Nvidia Windows-only features. I have way less to lose, actually.

A GPU comparison for Linux users would be quite interesting to see, where it compares head-to-head on open source as well as OEM proprietary drivers.

See https://www.xda-developers.com/reasons-ditch-nvidia-for-amd/
#Blog, #GPU, #technology

faab64@diasp.org

A shocking new discovery about China's successful ability to manufacture 7nm System on Chip in a local factory, based on domestic CPU and GPU architecture.

The technology blogs and "experts" are all in shock, because they had no idea what China was able to produce such advanced SoC like "Huawei's HiSilicon Kirin 9000S" that is powering the company's flagship mobile phone P60 Pro.

Ever since the US started the campaign to sanction and boycott Huawei a few years ago, the company has turned away from using American based products and any technology that could be controlled by the American/Western sanctions.

The impressive (yet not so fast with today's levels) SoC is nothing but a wake up call for the western companies dominating the #CPU, #GPU and #SoC market, from #Intel, to #AMD, #NVidia and of course #Qualcum.

The technology war that the US started, is not going to end well when China manages to create even more efficient and powerful CPUs to power not only expensive mobile phones, but also mid range and low cost ones as well as tablets and ultra portable tablets.

The treatment of #Russia after the War in #Ukraine and how the whole western world tried to isolate and corner Russia was a wake up call for China in so many fronts, and they seems to be making giant leaps rather than baby steps.

Huawei's HiSilicon Kirin 9000S looks to be a quite complex SoC packing four high-performance cores (one at up to 2.62 GHz and two at up to 2,150 MHz) and four energy-efficient cores (up to 1,530 MHz) based on the company's own TaiShan microarchitecture (which still looks to be found on the Armv8a ISA ) as well as the Maleoon 910 graphics processing unit operating at up to 750 MHz, based on screenshots by Huawei Central. CPU and GPU cores run at relatively low clocks compared to frequencies of Arm's cores featured in previous generations of HiSilicon's SoCs.

But low frequencies can be explained by the fact that SMIC makes the new SoC on its unannounced 2nd generation 7nm fabrication process, which could be a breakthrough for #SMIC, Huawei, and China's high-tech industry. Although TechInsights calls this fabrication technology SMIC's 2nd generation production node, state-controlled Global Times claims that China's foundry champion uses its 5nm-class manufacturing technology to make the SoC. But these two names seem to describe the same thing, which was once known as SMIC's N+2.

#Technology #SoC #Huawei #7nmTechnlogy #ChipManufacturing #China #US #Politics #Economy

https://www.tomshardware.com/news/huaweis-new-mystery-7nm-chip-from-chinese-fab-defies-us-sanctions

analysisparalysis@pod.beautifulmathuncensored.de

Petals dropped. Now you can use large models on a single gpu it says.
Reminds me of the “more expensive setup is better” argument that makes people buy new graphics cards and CPUs.

But what we need to define is application, not numbers.

Just like “With iPod, Apple has invented a whole new category of digital music player that lets you put your entire music collection in your pocket and listen to it wherever you go” and not “we invented a X GB mp3 player with Y KB cache”.

Application focus, not specs focus.

My goal is to have a model that gives accurate answers to questions about a document and that has the “decency” to admit that the answer cannot be found in those documents.

There are several options out there with local llms, yet none of them can be configured.

If a reply is bad, all you can do is choose another model. People hope that bigger was better, so they try to stuff huge and even bigger(petals) models into their computer, but what do those models really do?

They contain VAST corpuses on all kinds of topics. I assume, you won’t need 99% of those billions of parameters in your entire lifetime.

THIS is where you need to start: limit models by application. If you only search in English, don’t get a model that also contains Urdu.

If you only talk about computer science, don’t get that model that contains psychology.

Now the problem is that there are no models available that are specific - and good at what they do there.

To summarize, we need models that are specializable, by defining requirements and creating a new model based on that, which is much faster, smaller and gives you the right results.

EDIT: Another issue is what if the answers don’t satisfy you, would a human give better answers? Is it the model, what is it? You only have a few parameters like temperature and that’s it. Besides, a huge issue is model entitlement, the dreaded “This is not morally okay, and I will not”…Shut up, you are living on MY computer, I tell you what you do and don’t do.

Yet I feel like running against windmills, nobody debates that anywhere and people just report about the great, now 10 times more parameter model as if that saved humanity. #AI #MachineLearning #Models #ApplicationFocus #SpecsFocus #ComputerScience #English #Urdu #SpecializableModels #Accuracy #DocumentAnalysis #GPU #Hardware #Technology

frenchhope@diaspora-fr.org

Partages info-culturels du 21/01/2023

🛈📰🎹🎨📚 👨‍💻🌍🔗

En vrac : un rattrapage du plus significatif que j'ai déjà partagé aujourd’hui + divers autres liens susceptibles de vous intéresser mais n'incluant pas volontairement ce que vous avez déjà partagé ! Une #RevueDeWeb illustrée à l’aide de #MidJourney et #WordCloudLayout

Extended SARS-CoV-2 RBD booster vaccination induces humoral and cellular immune tolerance in mice - ScienceDirect

#étude #covid19 #vaccin #sarscov2 #immunité #vaccination #àtraduire


Le rythme du déclin cognitif serait lié à la perte de masse musculaire… - rtflash.fr | tregouet.org

#étude #santé #déclincognitif #massemusculaire #muscle #sarcopénie


Comment Firefox va très bientôt renforcer sa sécurité grâce à une fonction attendue depuis six ans

#cybersécurité #navigateur #web #mozillafirefox #bacàsable #gpu


Les traumatismes psychologiques modifient la structure du cerveau… - rtflash.fr | tregouet.org

#psychologie #traumatisme #cerveau #neuroscience #santé #santémentale


30 % des téléchargements de logiciels malveillants à partir du cloud proviennent de OneDrive, 7,6 % viennent de Github, 7,2 % de Sharepoint et 2,8 % de Google Drive

#cybersécurité #téléchargement #cloud #onedrive #logicielmalveillant


Clément Viktorovitch sur Twitter : "« Le Conseil d'Orientation des Retraites vous attrape en flagrant délit d'hyperbole : une exagération manifeste de la réalité, qui vise à jouer sur les peurs des Français. » Décryptage de la réforme, face au Ministre Gabriel Attal. Pour @Qofficiel ! https://t.co/ LVdhIvR4ak" / Twitter

#médias #politique #réforme #retraite #clémentviktorovitch #débat


I⸱D⸱E 👩‍💻🧐♀ 🌎🌊 aka Isabelle Delseny-Ernest sur Twitter : "Janvier 2023 : # licenciements #tech #GAFAM ➕ Amazon : 18 000 ➕ Microsoft : 10 000 ➕ Google/ Alphabet : 12 000 ➕ Salesforce : 7 000 ➕ META en nov 22 : 11 000 ➕ Twitter en nov 22 : 50 000 🟰1525 annonces de licenciements dans des entreprises technologiques > 230 000 pers 😱" / Twitter

#emploi #économie #licenciement #bigtech #tech #gafam


Le gouvernement américain n’a plus le droit d’emprunter et tente d’éviter un défaut de paiement

#politique #économie #étatsunis #faillite #défautdepaiement #finance #emprunt #dettepublique


Climat : Les démocraties ne sont pas à la hauteur car elles ne sont pas des démocraties - Cyril Dion - YouTube

#écologie #politique #démocratie #cyrildion #oligarchie #ploutocratie #conventioncitoyennepourleclimat


Quelle stratégie ? Combattre la réforme des retraites - Nicolas Framont (Frustration Magazine) - YouTube

#politique #économie #réforme #retraite #entreprise #bourgeoisie #nicolasframont #frustrationmagazine #rapportdeforce


ENNEMIS OU ADVERSAIRES ? BRANCO RÉPOND À BÉGAUDEAU - YouTube

#politique #françoisbegaudeau #juanbranco


FRÉDÉRIC LORDON ET LES FANTASMES DE POLICIERS SUR WHATSAPP - YouTube

#politique #police #violencespolicières #répression #pouvoir


FRIOT FUSTIGE ROUSSEL (et l'effondrement idéologique du PCF) - YouTube

#politique #bernardfriot #fabienroussel #pcf #communisme #socialdémocratie


Ces décideurs qui s'attaquent aux symptômes plutôt qu'aux problèmes réels [Ibrahima Fall] - YouTube

#management #décideur #entreprise #symptôme #problème #problèmeréel #traitantisme #ibrahimafall #procédure #travailréel #travailprescrit #processus #collectifdetravail #collectiondindividus #intérêtgénéral #intéretcollectif #interêtpersonnel #individualisation #psychologisation #sophismedelacauseunique


Retour à la maison : et si chez soi redevenait le centre de la vie [Ghislain Deslandes] - YouTube

#économie #chezsoi #ghislaindeslandes #entreprise #sphèreprivée #affairesprivées #domaineprivé


Les femmes seraient plus éthiques que les hommes, les jeunes moins que les anciens [Jocelyn Husser] - YouTube

#éthique #femme #homme #jeunesse #personnesâgées #seniors #jocelynhusser


Comprendre l’héroïsme hypermoderne au cœur de nos sociétés [Olivier Fournout] - YouTube

#société #héroisme #individualisme #olivierfournout #médias #management #entrepreneuriat


Supprimer (ou baisser) la TVA sur les produits essentiels : la fausse bonne idée [A.Mirlicourtois] - YouTube

#politique #économie #tva #inflation #entreprise #subvention


Le digital vs les biens matériels : la fin des choses [Ghislain Deslandes] - YouTube

#numérique #matière #matériel #chose #objets #ghislaindeslandes #information #décorporalisation #écran #bonheur #philosophie #calcul #gouvernanceparlesnombres #motivation #engagement


Un pouvoir d'achat en chute libre ? La réalité des chiffres [Alexandre Mirlicourtois] - YouTube

#économie #inflation #consommation #pouvoirdachat #alexandremilicourtois


Apaiser le climat social en entreprise en organisant démocratiquement l’activité [Corentin Gombert] - YouTube

#management #entreprise #démocratie #corentingombert #coopérative #scop #scic


Quand le manager produit la norme et devient… juge [Hugo Gaillard] - YouTube

#management #entreprise #juridique #justice #hugogaillard


La Russie brandit de nouveau sa torpille nucléaire Poséidon : faut-il s'inquiéter ?

#russie #arme #nucléaire #bombeatomique #bombeàhydrogène #technologie #tsunami #radioactivité


Six Minutes of Intense Exercise Boosts a Crucial Molecule in Your Brain : ScienceAlert

#santé #cerveau #exercicephysique #hauteintensité #hiit #protéines #àtraduire #bdnf


Best Privacy Focused Search Engines of 2023 [Compared]

#moteurderecherche #vieprivée #confidentialité #comparatif #àtraduire #classement


2 verres d’alcool par semaine suffiraient à développer un cancer, selon un rapport | News | Paris

#santé #cancer #alcool #consommation


Un joggeur attaqué par des rottweilers : le procureur demande l'euthanasie des deux chiens

#jogging #justice #faitdivers #chien #rottweiler #euthanasie


Peut-on séparer le bilan carbone de l'artiste ?

#art #artiste #culture #bilancarbone #industrie #économie #spectacle #surconsommation #empreinteenvironnementale


Droite radicale : les femmes s’y mettent aussi (surtout en France) | CNRS Le journal

Les femmes qui votent pour Marine Le Pen sont tout aussi acquises à ces idées que les hommes, sur la même ligne nationaliste, anti-immigrés, anti-Union européenne. Le charisme de Marine Le Pen et le changement d’image du RN les a encouragées à franchir le pas.

#société #politique #femme #droite #droiteradicale #extrêmedroite #france #religion #féminisme


CodeGPT: The VSCode Extension with ChatGPT-Like Functionalities | by The PyCoach | Geek Culture | Jan, 2023 | Medium

#développement #codegpt #vscode #chatgpt #gpt3 #àtraduire


Craiyon : un générateur d'images par IA open source et gratuit

#craiyon #générationdimages #intelligenceartificielle #opensource


Maître Pandaï sur Twitter : "Ah, ça y est, on a retrouvé l'argent magique ?" / Twitter

en plus de faire des choix douteux sur le plan économique, ces chiens nous mènent tout droit à la guerre ! >🔴 ALERTE INFO >Le budget des armées augmentera d'un tiers sur 2024-2030 à 400 milliards d'euros >https:// l.bfmtv.com/wZW

#politique #économie #armée #défense #france

lorenzoancora@pod.mttv.it

Mozilla Firefox is slow even if hardware acceleration is enabled?

Try enabling gfx.x11-egl.force-enabled in about:config to force the use of EGL1 and then restart the web browser.
If this works, enabling layers.acceleration.force-enabled can boost your performance even more by enabling OpenGL alongside EGL. 2
This also works on Flatpak if the security permissions are correct (use Flatseal3 if needed).

Tags: #linux #gnulinux #sysadmin #gpu #gpu-linux #gpuacceleration #egl #opengl #x11 #xorg #mozilla #firefox #hack #flatpak #flatseal

https://www.khronos.org/egl/
https://flathub.org/apps/details/com.github.tchx84.Flatseal
https://linuxreviews.org/HOWTO_Make_Mozilla_Firefox_Blazing_Fast_On_Linux


anonymiss@despora.de

#RISC-V based #GPU Processor

source: https://www.think-silicon.com/neox-graphics

NEOX™ is a parallel multicore and multithreaded GPU #architecture based on the RISC-V RV64C ISA instruction set with adaptive NoC. The number of cores varies from 4 to 64 organized in 1-16 cluster elements, each configured for cache sizes and thread counts . Depending on cluster / core configuration, NEOX™ compute power is ranging from 12.8 to 409.6 GFLOPS at 800MHz with support for FP16, FP32 and optionally FP64 and SIMD instructions.

Does anyone know how far #LibreRISC is with their GPU?

#news #hardware #graphic #neox #riscv

anonymiss@despora.de

#Ethereum miners spent $15 billion on GPUs in the last two years

source: https://www.techradar.com/in/news/ethereum-miners-spent-dollar15-billion-on-gpus-in-just-over-one-year

But now with its value tanking, many cryptoinvestors have found themselves unable to recoup the losses from spending so much money on GPUs and other parts. One such story, as outlined in the Bloomberg article, has one man who invested $30,000 in #cryptomining #hardware in mid-2021 but has only made about $5,000 so far through crypto.

That’s not even considering the environmental impact of all this mining, as the level of #carbon #emissions has always been a major #problem with cryptocurrencies. And it could be argued that between the waste of the GPUs purchased for #mining coupled with the environmental #damage, the value of Ethereum alone declining so steeply has made this all for nothing.

#news #environment #climate #crypto #cryptoCurrency #co2 #gpu #money #profit #crash #eth