There’s an artist in everyone. Some bring their ideas to life with digital tools, others capture the perfect moment with a camera or love to grab pen and paper to create little doodles or pieces of lettering. And even if you think you’re far from being an artist, well, why not explore it? It might just be hidden somewhere deep inside of you. For more than 14 years already our monthly wallpapers series has been the perfect opportunity to do just that: to break out of your daily routine and get fully immersed in a creative little project. This month is no exception, of course. For this post, artists and designers from across the globe once again put their creative skills to the test and designed beautiful, unique, and inspiring desktop wallpapers to accompany you through the new month. You’ll find their artworks compiled below, along with a selection of June favorites from our wallpapers archives that are just too good to be forgotten. A huge thank-you to everyone who shared their designs with us this time around — you’re smashing! If you, too, would like to get featured in one of our next wallpapers posts, please don’t hesitate to submit your design. We can’t wait to see what you’ll come up with! You can click on every image to see a larger preview. We respect and carefully consider the ideas and motivation behind each and every artist’s work. This is why we give all artists the full freedom to explore their creativity and express emotions and experience through their works. This is also why the themes of the wallpapers weren’t anyhow influenced by us but rather designed from scratch by the artists themselves. June Is For Nature “In this illustration, Earth is planting a little tree — taking care, smiling, doing its part. It’s a reminder that even small acts make a difference. Since World Environment Day falls in June, there’s no better time to give back to the planet.” — Designed by Ginger IT Solutions from Serbia. preview with calendar: 320x480, 640x480, 800x480, 800x600, 1024x768, 1024x1024, 1152x864, 1280x720, 1280x800, 1280x960, 1280x1020, 1400x1050, 1440x900, 1600x1200, 1680x1050, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440 without calendar: 320x480, 640x480, 800x480, 800x600, 1024x768, 1024x1024, 1152x864, 1280x720, 1280x800, 1280x960, 1280x1020, 1400x1050, 1440x900, 1600x1200, 1680x1050, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440 Tastes Of June “A vibrant June wallpaper featuring strawberries and fresh oranges, capturing the essence of early summer with bright colors and seasonal charm.” — Designed by Libra Fire from Serbia. preview with calendar: 320x480, 640x480, 800x480, 800x600, 1024x768, 1024x1024, 1152x864, 1280x720, 1280x800, 1280x960, 1280x1024, 1366x768, 1400x1050, 1440x900, 1600x1200, 1680x1050, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440 without calendar: 320x480, 640x480, 800x480, 800x600, 1024x768, 1024x1024, 1152x864, 1280x720, 1280x800, 1280x960, 1280x1024, 1366x768, 1400x1050, 1440x900, 1600x1200, 1680x1050, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440 A Bibliophile’s Shelf “Some of my favorite things to do are reading and listening to music. I know that there are a lot of people that also enjoy these hobbies, so I thought it would be a perfect thing to represent in my wallpaper.” — Designed by Cecelia Otis from the United States. preview with calendar: 320x480, 640x480, 800x480, 800x600, 1024x768, 1024x1024, 1152x864, 1280x720, 1280x800, 1280x960, 1280x1024, 1400x1050, 1440x900, 1600x1200, 1680x1050, 1680x1200, 1920x1200, 1920x1440, 2560x1440 without calendar: 320x480, 640x480, 800x480, 800x600, 1024x768, 1024x1024, 1152x864, 1280x720, 1280x800, 1280x960, 1280x1024, 1400x1050, 1440x900, 1600x1200, 1680x1050, 1680x1200, 1920x1200, 1920x1440, 2560x1440 Solana “Spanish origin, meaning ‘sunshine’.” — Designed by Bhabna Basak from India. preview with calendar: 1440x900, 1600x1200, 1680x1050, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440 without calendar: 1440x900, 1600x1200, 1680x1050, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440 Here Comes The Sun Designed by Ricardo Gimenes from Spain. preview with calendar: 640x480, 800x480, 800x600, 1024x768, 1024x1024, 1152x864, 1280x720, 1280x800, 1280x960, 1280x1024, 1366x768, 1400x1050, 1440x900, 1600x1200, 1680x1050, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440, 3840x2160 without calendar: 640x480, 800x480, 800x600, 1024x768, 1024x1024, 1152x864, 1280x720, 1280x800, 1280x960, 1280x1024, 1366x768, 1400x1050, 1440x900, 1600x1200, 1680x1050, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440, 3840x2160 Nature’s Melody “With eyes closed and music on, she blends into the rhythm of the earth, where every note breathes nature.” — Designed by Design Studio from India. preview with calendar: 800x600, 1280x1024, 1600x1200, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440 without calendar: 800x600, 1280x1024, 1600x1200, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440 Silent Glimmer “In the hush of shadows, a single amber eye pierces the dark — silent, watchful, eternal.” — Designed by Kasturi Palmal from India. preview with calendar: 800x600, 1280x1024, 1600x1200, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440 without calendar: 800x600, 1280x1024, 1600x1200, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440 Ice Cream “To me, ice cream is one of the most iconic symbols of summer. So, what better way to represent the first month of summer than through an iconic summer snack.” — Designed by Danielle May from Pennsylvania, United States. preview with calendar: 640x480, 800x480, 800x600, 1024x768, 1024x1024, 1152x864, 1280x720, 1280x800, 1280x960, 1280x1024, 1400x1050, 1440x900, 1600x1200, 1680x1050, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440 without calendar: 640x480, 800x480, 800x600, 1024x768, 1024x1024, 1152x864, 1280x720, 1280x800, 1280x960, 1280x1024, 1400x1050, 1440x900, 1600x1200, 1680x1050, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440 Silly Cats “I really loved the fun content aware effect and wanted to play around with it for this wallpaper with some cute cats.” — Designed by Italia Storey from the United States. preview with calendar: 320x480, 640x480, 800x480, 800x600, 1024x768, 1024x1024, 1152x864, 1280x720, 1280x800, 1280x960, 1280x1024, 1366x768, 1400x1050, 1440x900, 1600x1200, 1680x1050, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440 without calendar: 320x480, 640x480, 800x480, 800x600, 1024x768, 1024x1024, 1152x864, 1280x720, 1280x800, 1280x960, 1280x1024, 1366x768, 1400x1050, 1440x900, 1600x1200, 1680x1050, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440 In Case Of Nothing To Do Designed by Ricardo Gimenes from Spain. preview with calendar: 640x480, 800x480, 800x600, 1024x768, 1024x1024, 1152x864, 1280x720, 1280x800, 1280x960, 1280x1024, 1366x768, 1400x1050, 1440x900, 1600x1200, 1680x1050, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440, 3840x2160 without calendar: 640x480, 800x480, 800x600, 1024x768, 1024x1024, 1152x864, 1280x720, 1280x800, 1280x960, 1280x1024, 1366x768, 1400x1050, 1440x900, 1600x1200, 1680x1050, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440, 3840x2160 Pink Hours “With long-lasting days, it is pleasant to spend hours walking at dusk. This photo was taken in an illuminated garden.” — Designed by Philippe Brouard from France. preview with calendar: 1024x768, 1366x768, 1600x1200, 1920x1080, 1920x1200, 2560x1440, 2560x1600, 2880x1800 without calendar: 1024x768, 1366x768, 1600x1200, 1920x1080, 1920x1200, 2560x1440, 2560x1600, 2880x1800 What’s The Best That Could Happen? Designed by Grace DiNella from Doylestown, PA, United States. preview with calendar: 320x480, 640x480, 800x480, 1024x768, 1024x1024, 1152x864, 1280x720, 1280x800, 1280x960, 1280x1024, 1400x900, 1400x1050, 1680x1050, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440 without calendar: 320x480, 640x480, 800x480, 1024x768, 1024x1024, 1152x864, 1280x720, 1280x800, 1280x960, 1280x1024, 1400x900, 1400x1050, 1680x1050, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440 Purrsuit “Recently I have been indulging in fishing as a means of a hobby, and the combined peace and thrill of the activity inspires me. I also love cats, so I thought combining the two subjects would make a stellar wallpaper, especially considering that these two topics already fall hand in hand with each other!” — Designed by Lilianna Damian from Scranton, PA, United States. preview with calendar: 640x480, 800x480, 800x600, 1024x768, 1024x1024, 1152x864, 1280x720, 1280x800, 1280x960, 1280x1024, 1400x1050, 1440x900, 1600x1200, 1680x1050, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440 without calendar: 640x480, 800x480, 800x600, 1024x768, 1024x1024, 1152x864, 1280x720, 1280x800, 1280x960, 1280x1024, 1400x1050, 1440x900, 1600x1200, 1680x1050, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440 Happy Best Friends Day! “Today’s all about celebrating the ones who laugh with us, cry with us, and always have our backs — our best friends. Whether it’s been years or just a few months, every moment with them means something special. Tag your ride-or-die, your soul sibling, your partner in crime - and let them know just how much they mean to you.” — Designed by PopArt Studio from Serbia. preview with calendar: 320x480, 640x480, 800x480, 800x600, 1024x768, 1024x1024, 1152x864, 1280x720, 1280x800, 1280x960, 1280x1024, 1400x1050, 1440x900, 1600x1200, 1680x1050, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440 without calendar: 320x480, 640x480, 800x480, 800x600, 1024x768, 1024x1024, 1152x864, 1280x720, 1280x800, 1280x960, 1280x1024, 1400x1050, 1440x900, 1600x1200, 1680x1050, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440 Travel Time “June is our favorite time of the year because the keenly anticipated sunny weather inspires us to travel. Stuck at the airport, waiting for our flight but still excited about wayfaring, we often start dreaming about the new places we are going to visit. Where will you travel to this summer? Wherever you go, we wish you a pleasant journey!” — Designed by PopArt Studio from Serbia. preview without calendar: 320x480, 640x480, 800x480, 800x600, 1024x768, 1024x1024, 1152x864, 1280x720, 1280x800, 1280x960, 1280x1024, 1400x1050, 1440x900, 1600x1200, 1680x1050, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440 Summer Coziness “I’ve waited for this summer more than I waited for any other summer since I was a kid. I dream of watermelon, strawberries, and lots of colors.” — Designed by Kate Jameson from the United States. preview without calendar: 320x480, 1024x1024, 1280x720, 1680x1200, 1920x1080, 2560x1440 Deep Dive “Summer rains, sunny days, and a whole month to enjoy. Dive deep inside your passions and let them guide you.” — Designed by Ana Masnikosa from Belgrade, Serbia. preview without calendar: 320x480, 640x480, 800x480, 800x600, 1024x768, 1024x1024, 1152x864, 1280x720, 1280x800, 1280x960, 1280x1024, 1400x1050, 1440x900, 1600x1200, 1680x1050, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440 All-Seeing Eye Designed by Ricardo Gimenes from Spain. preview without calendar: 640x480, 800x480, 800x600, 1024x768, 1024x1024, 1152x864, 1280x720, 1280x800, 1280x960, 1280x1024, 1366x768, 1400x1050, 1440x900, 1600x1200, 1680x1050, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440, 3840x2160 Join The Wave “The month of warmth and nice weather is finally here. We found inspiration in the World Oceans Day which occurs on June 8th and celebrates the wave of change worldwide. Join the wave and dive in!” — Designed by PopArt Studio from Serbia. preview without calendar: 320x480, 640x480, 800x480, 800x600, 1024x768, 1024x1024, 1152x864, 1280x720, 1280x800, 1280x960, 1280x1024, 1366x768, 1400x1050, 1440x900, 1600x1200, 1680x1050, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440 Create Your Own Path “Nice weather has arrived! Clean the dust off your bike and explore your hometown from a different angle! Invite a friend or loved one and share the joy of cycling. Whether you decide to go for a city ride or a ride in nature, the time spent on a bicycle will make you feel free and happy. So don’t wait, take your bike and call your loved one because happiness is greater only when it is shared. Happy World Bike Day!” — Designed by PopArt Studio from Serbia. preview without calendar: 320x480, 640x480, 800x480, 800x600, 1024x768, 1024x1024, 1152x864, 1280x720, 1280x800, 1280x960, 1280x1024, 1366x768, 1400x1050, 1440x900, 1600x1200, 1680x1050, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440 Oh, The Places You Will Go! “In celebration of high school and college graduates ready to make their way in the world!” — Designed by Bri Loesch from the United States. preview without calendar: 320x480, 1024x768, 1280x1024, 1440x900, 1680x1050, 1680x1200, 1920x1440, 2560x1440 Merry-Go-Round Designed by Xenia Latii from Germany. preview without calendar: 320x480, 640x480, 800x480, 800x600, 1024x768, 1152x864, 1280x720, 1280x800, 1280x960, 1280x1024, 1366x768, 1400x1050, 1440x900, 1600x1200, 1680x1050, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440 Summer Surf “Summer vibes…” — Designed by Antun Hirsman from Croatia. preview without calendar: 640x480, 1152x864, 1280x1024, 1440x900, 1680x1050, 1920x1080, 1920x1440, 2650x1440 Expand Your Horizons “It’s summer! Go out, explore, expand your horizons!” — Designed by Dorvan Davoudi from Canada. preview without calendar: 800x480, 800x600, 1024x1024, 1152x864, 1280x720, 1280x800, 1280x960, 1280x1024, 1366x768, 1400x1050, 1440x900, 1600x1200, 1680x1050, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440 Gravity Designed by Elise Vanoorbeek from Belgium. preview without calendar: 1024x768, 1024x1024, 1152x864, 1280x720, 1280x800, 1280x960, 1280x1024, 1366x768, 1400x1050, 1440x900, 1600x1200, 1680x1050, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440 Yoga Is A Light, Which Once Lit, Will Never Dim “You cannot always control what goes on outside. You can always control what goes on inside. Breathe free, live and let your body feel the vibrations and positiveness that you possess inside you. Yoga can rejuvenate and refresh you and ensure that you are on the journey from self to the self. Happy International Yoga Day!” — Designed by Acodez IT Solutions from India. preview without calendar: 320x480, 640x480, 800x480, 800x600, 1024x768, 1024x1024, 1152x864, 1280x720, 1280x960, 1280x1024, 1366x768, 1400x1050, 1440x900, 1600x1200, 1680x1050, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440 Evolution “We’ve all grown to know the month of June through different life stages. From toddlers to adults with children, we’ve enjoyed the weather with rides on our bikes. As we evolve, so do our wheels!” — Designed by Jason Keist from the United States. preview without calendar: 320x480, 800x600, 768x1024, 1280x800, 1280x1024, 1440x900, 1920x1080, 2560x1440 Summer Party Designed by Ricardo Gimenes from Spain. preview without calendar: 640x480, 800x480, 800x600, 1024x768, 1024x1024, 1152x864, 1280x720, 1280x800, 1280x960, 1280x1024, 1366x768, 1400x1050, 1440x900, 1600x1200, 1680x1050, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440, 3840x2160 Splash Designed by Ricardo Gimenes from Spain. preview without calendar: 640x480, 800x480, 800x600, 1024x768, 1024x1024, 1152x864, 1280x720, 1280x800, 1280x960, 1280x1024, 1366x768, 1400x1050, 1440x900, 1600x1200, 1680x1050, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440, 3840x2160 Reef Days “June brings the start of summer full of bright colors, happy memories, and traveling. What better way to portray the goodness of summer than through an ocean folk art themed wallpaper. This statement wallpaper gives me feelings of summer and I hope to share that same feeling with others.” — Designed by Taylor Davidson from Kentucky. preview without calendar: 480x800, 1024x1024, 1242x2208, 1280x1024 Solstice Sunset “June 21 marks the longest day of the year for the Northern Hemisphere — and sunsets like these will be getting earlier and earlier after that!” — Designed by James Mitchell from the United Kingdom. preview without calendar: 1280x720, 1280x800, 1366x768, 1440x900, 1680x1050, 1920x1080, 1920x1200, 2560x1440, 2880x1800 Wildlife Revival “This planet is the home that we share with all other forms of life and it is our obligation and sacred duty to protect it.” — Designed by LibraFire from Serbia. preview without calendar: 320x480, 640x480, 800x480, 800x600, 1024x768, 1024x1024, 1152x864, 1280x720, 1280x800, 1280x960, 1280x1024, 1366x768, 1400x1050, 1440x900, 1600x1200, 1680x1050, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440 Pineapple Summer Pop “I love creating fun and feminine illustrations and designs. I was inspired by juicy tropical pineapples to celebrate the start of summer.” — Designed by Brooke Glaser from Honolulu, Hawaii. preview without calendar: 640x480, 800x600, 1024x768, 1152x720, 1280x720, 1280x800, 1280x960, 1366x768, 1440x900, 1680x1050, 1920x1080, 1920x1200, 1920x1440, 2560x1440 Handmade Pony Gone Wild “This piece was inspired by the My Little Pony cartoon series. Because those ponies irritated me so much as a kid, I always wanted to create a bad-ass pony.” — Designed by Zaheed Manuel from South Africa. preview without calendar: 800x600, 1024x768, 1280x960, 1280x1024, 1680x1050, 1920x1200, 2560x1440, 2880x1800 Window Of Opportunity “‘Look deep into nature and then you will understand everything better,’ A.E.” — Designed by Antun Hiršman from Croatia. preview without calendar: 1024x768, 1280x960, 1366x768, 1440x900, 1600x1200, 1680x1200, 1920x1080, 1920x1440, 2560x1440 Viking Meat War Designed by Ricardo Gimenes from Spain. preview without calendar: 320x480, 1024x768, 1024x1024, 1280x800, 1280x960, 1280x1024, 1366x768, 1400x1050, 1440x900, 1600x1050, 1600x1200, 1680x1050, 1680x1200, 1920x1080, 1920x1200, 1920x1440, 2560x1440, 2880x1800 #prelude #summer #june #wallpapers #edition

Gironda Residence | © Simone Bossi Located just steps from Piazza del Popolo in Ravenna, the Renaissance-era Casa Guaccimanni holds centuries of architectural and historical weight. Constructed in the fifteenth century for the Venetian podestà Nicolò Giustinian, the building evolved through noble ownership and later became home to Vittorio and Alessandro Guaccimanni, sons of Risorgimento figure Luigi Guaccimanni. Architecturally, the structure is characterized by a tripartite plan with a central corridor flanked by large rooms, an interior courtyard with a double loggia, and decorative elements spanning Renaissance to Neoclassical periods. Once concealed beneath plaster, its frescoed veranda and exposed wooden ceilings speak to a layered history of intervention, concealment, and rediscovery. Gironda Residence in Casa Guaccimanni Technical Information Architects1-13: Giovanni Mecozzi Architetti Location: Casa Guaccimanni, Via Armando Diaz, Ravenna, Italy Client: Emanuela Docimo Project Years: 2022 – 2024 Original Structure: 15th Century Photographs: © Andrea Sestito, © Simone Bossi, © Omar Sartor The new and the old never touch, but gently brush against each other, maintaining a distance capable of generating tension. – Giovanni Mecozzi Gironda Residence in Casa Guaccimanni Photographs © Omar Sartor © Andrea Sestito © Andrea Sestito © Andrea Sestito © Omar Sartor © Simone Bossi © Simone Bossi © Simone Bossi © Omar Sartor © Omar Sartor © Omar Sartor © Andrea Sestito © Omar Sartor Design Intent: Reversibility and Temporal Tension The recent architectural project by Giovanni Mecozzi Architetti centers on the noble floor of the palazzo, reinterpreted as a contemporary residence named Gironda. Rather than imposing a new visual regime onto the historic shell, the intervention operates with restraint, foregrounding the building’s original character while establishing new spatial and material conditions. At the core of the project lies a design philosophy rooted in reversibility. Mecozzi’s intervention resists permanence. The furnishings and spatial devices introduced into the historic rooms are self-supporting and detached from the structure. No new element makes physical contact with the floors, ceilings, or walls, preserving the integrity of the original surfaces. This strategy avoids irreversible alterations and allows the architecture to remain temporally flexible. Architect Giovanni Mecozzi articulates this approach succinctly: “The new and the old never touch, but gently brush against each other, maintaining a distance capable of generating tension.” This spatial tension is not decorative but conceptual, prompting occupants to consider the relationship between historical continuity and contemporary transformation. The design does not attempt to erase time but rather exposes its layers through careful juxtaposition. The project draws conceptual and chromatic inspiration from Ravenna’s early Christian and Byzantine mosaics. Rather than replicate ornamental motifs, Mecozzi extracts abstract qualities such as color, luminosity, and surface texture, integrating them as subtle spatial references throughout the residence. Gironda Residence Material Strategy Access to the residence is organized through a longitudinal hallway that bisects the plan, connecting a balcony on the north façade with a loggia overlooking the garden to the south. This corridor becomes a spine for circulation and orientation, punctuated by entries into five main rooms: the kitchen, veranda, and three independent suites. Each suite functions as a self-contained spatial environment. The original large rooms have been reimagined with integrated volumes housing diverse domestic functions: bathrooms, saunas, walk-in closets, reading nooks, and home cinemas. These new programmatic layers are embedded within freestanding furniture structures, which operate more as inhabitable objects than architectural partitions. Color becomes an operative tool for spatial differentiation. The three principal suites, the Gold Room, the Blue Room, and the Green Room, are introduced chromatically through thresholds that face the main corridor. This prelude of color sets the tone for each room’s unique interior experience. Within, glossy glass tiles, gilded surfaces, and a reduced palette of materials establish a scenographic yet restrained environment. The flooring, a Venetian terrazzo installed during earlier restoration work in the 2000s, has been retained. Its beveled borders and rounded corners respond to the proportions of each room, reinforcing a visual continuity that binds the new interventions with the inherited context. In contrast to the historical envelope, the furniture and spatial devices employ a language of monochromatic forms and minimal detailing, occasionally verging on neoplastic abstraction. This tension between old ornament and new abstraction is one of the project’s defining features. Furnishings curated by Atelier Biagetti, known for their theatrical and ironic sensibility, further enrich the atmosphere. These pieces do not mimic the historical setting but create moments of visual friction and playful ambiguity, enhancing the multi-temporal character of the interiors. Architectural Significance and Cultural Dialogue The Gironda residence exemplifies a growing discourse in contemporary architecture around adaptive reuse that neither mimics nor erases the past. Rather than treating heritage as a constraint or an aesthetic to be curated, Mecozzi engages it as an active agent in spatial transformation. The project is a case study in reversible architecture, where temporality is embedded in the design, not just its historical references. This intervention prompts broader questions about the role of preservation in contemporary practice. Can architectural interventions occupy historic contexts without becoming parasitic or nostalgic? Mecozzi’s project suggests that they can adopt a posture of critical distance and conceptual clarity. Gironda does not attempt to restore Casa Guaccimanni to a previous state or impose a singular vision of modernity. Instead, it crafts a dialogue between past and present, structured through spatial strategies, material choices, and chromatic cues. In doing so, it opens a new chapter in the building’s ongoing life, one that is fully contemporary yet deeply rooted in architectural memory. Gironda Residence in Casa Guaccimanni Plans Floor Plan | © Giovanni Mecozzi Architetti Golden Room Layout | © Giovanni Mecozzi Architetti Door Detail | © Giovanni Mecozzi Architetti Gironda Residence in Casa Guaccimanni Image Gallery About Giovanni Mecozzi Giovanni Mecozzi is an Italian architect based in Ravenna, Italy, and the founder of Giovanni Mecozzi Architetti, a multidisciplinary studio specializing in architecture, interior design, and landscape projects. After graduating from the University of Ferrara with an architecture degree, Mecozzi gained international experience working in Spain, including collaborating with Mendaro Arquitectos in Madrid. Upon returning to Italy, he co-founded GMA, focusing on projects emphasizing the relationship between architecture, the client, and the context, with a particular interest in renovating and transforming historical buildings.  Credits and Additional Notes Design Team: Giovanni Mecozzi, Cecilia Verdini, Filippo Minghetti Construction: EdilcostruzioniElectrical Systems: Elektra ServiceMechanical and Hydraulic Systems: Nuova OLP Structural Alterations: Not applicableCustom Furniture: Idea LegnoCurtains and Fabrics: Selezione Arredamenti, Ravenna Lighting: ViabizzunoResin Coatings and Flooring: Kerakoll Rugs and Carpeting: Centro Moquette, Rimini Bathroom Furnishings: Salaroli, Ravenna Furniture, Artwork, and Design Objects Selected by: Atelier BiagettiFurniture Designers: Alberto Biagetti and Laura Baldassarri #gironda #residence #giovanni #mecozzi #renovation

Samsung and other companies would be affected, but perhaps not as much as Apple. Credit: Li Hongbo/VCG via Getty Images In a Friday morning Truth social post, President Donald Trump threatened to levy a 25 percent tariff on iPhones if Tim Cook didn't move manufacturing to the United States. It turns out the situation is a little more complicated than that.While speaking to the media later that day, Trump clarified that the tariff would apply to any company selling foreign-made phones in the U.S., not just Apple. The president said the new 25% smartphone tariff could arrive by the end of June, per Bloomberg. He also made sure to single out Samsung, the second-most popular smartphone brand in the U.S. market. This broader approach makes a bit more sense than the initial threat against Apple, as it was unclear how the Trump administration planned to place tariffs on one single company's products.Even so, it wasn't the first time Trump made a tariff threat against a specific company, and it may not be the last. Besides Apple, President Trump previously threatened to target toymaker Mattel with tariffs. Milan Miric, PhD, Associate Professor of Data Sciences and Operations at the University of Southern California's Marshall School of Business, explained to Mashable how President Trump could effectively target a single company with tariffs. You May Also Like SEE ALSO: Apple smart glasses could come as soon as 2026 "For Apple, hardware products are their most important business line. All of the other competitors in the U.S. that can compete on hardware would be foreign companies manufacturing abroad," Miric told Mashable via email. "Therefore, if you wanted to target consumer electronics coming from China, the U.S. company that would be the most directly affected is Apple."For context, Apple relies far more heavily on hardware sales to bolster its business than U.S.-based competitors like Google and Microsoft, which are primarily service companies that happen to sell some hardware too. Mashable Light Speed Want more out-of-this world tech, space and science stories? Sign up for Mashable's weekly Light Speed newsletter. By clicking Sign Me Up, you confirm you are 16+ and agree to our Terms of Use and Privacy Policy. Thanks for signing up! According to Miric, all of this could be a prelude to something resembling a trade deal with Apple, as Trump has negotiated with foreign governments."You could imagine a scenario where large American companies strike a compromise with the government, where some of their own products that are very popular with American consumers and important to American business get exceptions, but then tariffs apply broadly to foreign companies, effectively providing additional protection to these American companies," Miric said. Related Stories Earlier this year, Apple promised to spend billion in the U.S. over the next four years and build a new factory in Texas, but iPhone manufacturing specifically is unlikely to return to the United States. As Mashable's Stan Schroeder previously reported, a U.S.-made iPhone would likely cost at least While new tariffs on smartphones could be arriving as soon as June, the president's tariff policy has included a few surprising reversals. Wall Street is paying attention, however.Samsung and Apple stock both fell on Friday after the president's remarks. Topics Apple Samsung Tariffs Alex Perry Tech Reporter Alex Perry is a tech reporter at Mashable who primarily covers video games and consumer tech. Alex has spent most of the last decade reviewing games, smartphones, headphones, and laptops, and he doesn’t plan on stopping anytime soon. He is also a Pisces, a cat lover, and a Kansas City sports fan. Alex can be found on Bluesky at yelix.bsky.social. #trump #warns #apple #samsung #that

ZEUS is expected to produce even higher levels of power later this year. Credit: Marcin Szczepanski / Michigan Engineering Get the Popular Science daily newsletter💡 Breakthroughs, discoveries, and DIY tips sent every weekday. A laser named after the most powerful god in ancient Greece is living up to its name. According to the University of Michigan, the first official experiment from the Zettawatt-Equivalent Ultrashort pulse laser Systemyielded 2 petawattsof power. For reference, that’s more than 100 times the electricity output across the entire planet—but you’ll likely miss it if you blink. The school gym-sized ZEUS is the successor to the university’s HERCULES system, which topped out at 300 terawatts. ZEUS relied on a reconfigured target for its debut demonstration, which required firing a laser pulse at a cell containing helium. The subsequent interaction produced plasma as it tore electrons away from the helium atoms, resulting in a mixture of both positive ions and free electrons. The electrons started gaining speed behind the laser pulse in a phenomenon known as wakefield acceleration. Because light moves slower in plasma, the electrons can catch up to the laser beam. Those free electrons also get more time to speed up given the size of the target chamber and thereby hit higher speeds. The recent demonstration is a prelude to the signature ZEUS experiment scheduled for later this year. In that test, the accelerating electrons will also smack into laser pulses coming from the opposite direction. This is where things getmore complicated—but to condense it down, the effect makes a 3-petawatt laser appear one million times more powerful, hence the “zettawatt-equivalent” in ZEUS’ name. Accomplishing this and other experimental feats does require some safeguards. ZEUS includes optical devices known as diffraction gratings that stretch out the initial infrared pulse over time. This ensures the initial power doesn’t get so intense that it begins tearing apart the air around it. Another goal is to ultimately create beams of electrons with energies similar to those found in particle accelerators hundreds of feet longer than ZEUS at a fraction of both its size and operating costs. At only million to construct, the University of Michigan previously described the machine as a “bargain.” Years of construction, calibration, and expertise is showcased in an astoundingly short amount of time. ZEUS’ 2 petawatt firing lasted just 25 quintillionths of a second. But future experiments will make the most of these moments.  John Neesand laser engineer Paul Campbellwork in Target Area 1, where the first 2 petawatt user experiment will take place. ZEUS is now the most powerful laser in the US. Credit: Marcin Szczepanski / Michigan Engineering “The fundamental research done at the NSF ZEUS facility has many possible applications, including better imaging methods for soft tissues and advancing the technology used to treat cancer and other diseases,” explained Vyacheslav Lukin, program director in the National Science d Division of Physics, which is responsible for the ZEUS project. Meanwhile, ZEUS experiments could also help researchers explore the dynamics of positron jets that shoot out of black holes, or how gamma ray bursts operate.  “One of the great things about ZEUS is it’s not just one big laser hammer, but you can split the light into multiple beams,” said Franklin Dollar, a University of California professor of physics and astronomy who oversaw the 2 petawatt experiment. “Having a national resource like this, which awards time to users whose experimental concepts are most promising for advancing scientific priorities, is really bringing high-intensity laser science back to the U.S.” #test #fired #its #most #powerful

Control is still the GOAT when it comes to uncanny beauty in video games Unhomely alonely. Image credit: Remedy Entertainment Feature by Christian Donlan Contributing Editor Published on May 18, 2025 Stop me if you've heard this before. We say "uncanny", but in Germany it's "unheimlich". The unheimlich was a big deal for people like Freud, and it's hard not to love the term, just a little bit. Unheimlich means, well, it means uncanny - weird, eerie, unsettling. But more specifically, it translates as "unhomely." Unhomely. Now that is a word that carries a chill, a creep of the flesh, a word that registers an arachnid skittering in the corner of your vision. When something is familiar and unfamiliar all at once! You should feel like you're at home, but... Testify! Rotary telephones with no dials. Bodies suspended in the air with a kind of ballerina poise and elegance. Staff portraits, but they're, like, full-blown oil paintings, dark eyes and unknowable aspects. When it comes to the uncanny, there's one big budget game that really delivers on it for me. It's Control. It's a shooter, I guess, a third-person action game inspired by everything from The X-Files to House of Leaves to tropical Brutalism. There's a splinter of Psi-Ops: The Mindgate Conspiracy and even Stranglehold in there too. But it's also pure, delicious, slow-dripping uncanniness. It's familiar and unfamiliar. It's... Well... Control is set within a building known as the Oldest House, and here, already, things are getting weird. Lots of games are perfectly house-sized. Edith Finch. Maniac Mansion. But a special few are set within houses that feel much bigger than the game they contain - much bigger than a single imagination could ever understand. Jet Set Willy. Impossible Mission. Control. The Oldest House contains Control, then, but it also feels like it contains so many other things, so many other implausible, improbable, impossible things. This week, for example, Remedy gave us a taste of FBC: Firebreak, a hectic multiplayer action game. Yes, it's set in the world of Control, but more specifically it's set within the Oldest House. Why not? There's plenty of room. Here's Aoife's take on Control from back in the day.Watch on YouTube In Control's fiction, the Oldest House is the headquarters of the Federal Bureau of Control. This is a government agency that deals with extremely bizarre stuff. But the Bureau's surfaces are beautiful and often quietly elegant. The face the Oldest House presents to the world is one of tastefully curated mid-century modernism. Wood looks like wood. Tile looks like tile. Concrete looks like concrete, and there's lots of it, along with indoor bays for tropical plants and wide staircases and bright overhead lighting that's perfect for giving ghostly shape to cigarette smoke. Players come to the Oldest House as Jesse, a woman searching for her kidnapped brother. That's a straightforward set up, and a game as unmoored as Control needs it, just as it needs the surprisingly strict confines of its third-person combat. Shoot stuff with a gun that can change form. Throw stuff around with your mind. Explore and clear out monsters as you dig for the truth. There's this unexpected core of extreme orderliness at the center of Control, in terms of what you do most of the time and why you've been told you're doing it. In retrospect, I guess the name of the game was a hint. This orderliness allows the designers to then bolt on all kinds of antic zaniness with no danger of the whole thing collapsing into fripperies, and it reminds me a little of the way that even the most extravagant skyscraper begins with the fixed core of its elevator shafts.Control. | Image credit: Remedy Entertainment The plot is really just there to direct you back and forth through the Oldest House itself, though, and like all uncanny things, transformation is always in the air. Sometimes the Oldest House is doing the transforming. Sometimes you're more directly in charge. Both these approaches are brilliant. When it comes to the Oldest House's own tricks, it tends to start with subtlety. Places will take slightly too long to reach. The signage looks weird and announces unusual departments. The technology is of the wrong era - green-screen computers? Bakelite phones? Radio equipment that looked like it may have been of use during the space race? This is all just the prelude, the aura before the migraine. But soon the Old House is revealing itself in full. Rooms telescope or twist, like the architect behind the place was willing to bust out a Spirograph. Smooth surfaces give way to jumbled blocks. In one area, chill infects the corridors and the glass freezes over, and are those tree trunks in the distance? In another, a furnace radiates the full heat of the sun. This is where that splinter of internal order really helps. It helps to orient. Because you know what you're doing and why you're doing it, the building can transform around you in quite wild ways without becoming frustrating. But the Oldest House also has a lovely sense of pace. It will go wild for a few moments and completely transform itself, but then it turns quiet again. Safe? Ah, but now you're left questioning the simplest things - the background hum, the volume of a room. Did I walk this way before? Was this hallway pointed in the same direction? Control. | Image credit: Remedy Entertainment Further on things get really, really weird, but in a way the Oldest House is at its best before it drops too far into sci-fi. It's at its best when it's got something tricksy, something of the funfair to it. There's an incredible sequence towards the end of the adventure when the Oldest House flings everything at you and you're racing through an environment that warps and twists and strobes, an environment that truly contains multitudes. This is fantastic, and feels like the middle-eight in a song, changing things up and leaving them elevated. But it's not the moment people mention the most in my experience when they talk about Control. That moment is an office that is simply overrun with Post-its. Post-its belong in offices, of course, but here they cover absolutely everything. Familiar and unfamiliar. And kind of funny, too. And here's a thing I really admire about Control. It's willing to be very funny, but not in the way that games about hellish offices normally try to be funny. It doesn't take the top-down approach, so often lifted wholesale from Portal's precision wit and cruelty, in which humans are sane lab-rats in an insane hierarchy of faux camaraderie and corporate double-speak, a world in which there are a million euphemisms for dying at work and there's a slice of cake promised if you can survive. Control. | Image credit: Remedy Entertainment No. Control is funny because it looks at the way all offices affect the people who work within them - how the creeping of weird rules and weird ways of being is slow and measured and often hard to spot. The comedy here comes from how that quietly insane hierarchy in turn warps ordinary people. The fact that it relies on the supernatural, in a way, only makes it more relatable. It's the comedy - and horror - of a room in which someone's really busted out the Post-its. It's the comedy of lamping someone with a photocopier. Here's a trailer for FBC: Firebreak. Are you ready to go back in?Watch on YouTube And with that image, we've the other kind of transformation in Control. It's the transformation brought about by the player, as they encounter the Oldest House's many enemies and use gunfire and numerous supernatural abilities to fight back. The gun you're given is pretty great, but Control's really about telekinesis - about picking stuff up with your brain, watching it sway woozily in the air, and then lobbing it somewhere in order to really hurt something. It's the transformation of impact, of violence and splatter. All that concrete is so eager to come apart in gritty chunks. All those office supplies are ready for a second life as a projectile or cudgel. You can take the Oldest House apart just as quickly as it tries to build itself into something monstrous. When I first played Control I would often end action sequences realising I was just firing wildly into the walls. They seemed, I guess, like a viable target. And while Control's mid-century vibes are very pleasant to engage with, it's the violence - this is a very weird thing to say - that's truly beautiful. When I first played the game, I remember chucking stuff at a filing cabinet for hours, just to enjoy the rippling of its drawers caused by the impact. In the moment of connection, it was briefly alive. And after a particularly fraught battle I remember staring at a crater on the wall, made by the impact of some kind of white foam or powder - maybe the aftermath of a fire extinguisher? This crater, this moon landing on a wall, struck me as being one of the most beautiful things I'd ever seen in a game. Not least because I suspect if I'd played that encounter slightly differently, I would not have seen it. There's so much more to Control than this, but the uncanny takes strange shapes when it lives in the mind, and this is the shape Control has taken for me. It's not the shape of the Oldest House, because who ever could map such a thing? But it's a shape that clearly fits inside the Oldest House. Deep inside, where it works its strange, dark, endlessly charming magic. #control #still #goat #when #comes

May 13, 20256 min readDeep Math from String Theory Appears in Clashing Black HolesResearchers have shown that abstract mathematical functions from the frontiers of theoretical physics have a real-world use in modeling gravitational wavesBy Ramin Skibba edited by Lee BillingsAn illustration of two black holes about to merge and emitting copious gravitational waves. Chris Henze/NASA/SPL/Getty ImagesA decade ago astrophysicists at the Laser Interferometer Gravitational-Wave Observatory, operated by the California Institute of Technology and the Massachusetts Institute of Technology, managed to detect subtle ripples in spacetime called gravitational waves, released by a pair of black holes spiraling into each other, for the first time. That impressive discovery—which earned the 2017 Nobel Prize in Physics—has since become commonplace, with researchers regularly detecting gravitational waves from myriad far-distant celestial sources.And as the numbers of gravitational-wave observations have increased, physicists’ careful modeling is revealing new details about their mysterious origins. Some of the most intriguing gravitational-wave events, it turns out, could arise not from catastrophic collisions but rather from near misses. Furthermore, these cosmic close calls might be best understood using concepts derived from string theory—a notional theory of everything that posits that all of nature is fundamentally composed of countless, wriggling subatomic strings. This arguably marks the first linkage to date between a core mathematical aspect of the arcane theory and real-world astrophysics.At least, that’s the conclusion of an international team of researchers that applied geometric structures inspired by particle physics and string theory to the behavior of black holes when the colossal objects closely pass and deflect each other. Such interactions between black holes or neutron starscan be studied through the deflection angle, the energy released through the near miss and the momentum of the objects’ recoil—all of which may be discerned in gravitational waves. The team’s results were published in the journal Nature on Wednesday.On supporting science journalismIf you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.Black Holes as ParticlesIn their study, the researchers used an obscure class of abstract mathematical functions to solve the formidable equations involved in determining the radiated energy from a near miss. “You need these new functions, which, in math and mathematical physics, have been studied intensively but, to date, have not appeared in any real physical observable. That’s what makes it quite interesting,” says Jan Plefka, a theoretical physicist at the Humboldt University of Berlin and a co-author of the new study. Those obscure functions, known as six-dimensional Calabi-Yau manifolds, had never been shown as directly relevant to descriptions of real astrophysical phenomena before.In the years since LIGO’s initial detection, two additional major gravitational-wave observatories, Europe’s Virgo and Japan’s Kamioka Gravitational-Wave Detector, have come online. Together, they form the international LIGO-Virgo-KAGRA collaboration and have amassed detections of nearly 300 gravitational-wave events over the past decade, mostly from colliding pairs of black holes. Also called black hole “mergers,” such events are the cacophonous moment when these dense gravitational behemoths smash together to form a single, larger beast. Plefka and his colleagues are studying different interactions known as “scattering” events, which occur when paired black holes slip by each other, usually in prelude to their eventual coalescence. During these close encounters, the clashing gravity of the black holes causes each to accelerate past the other, generating a significant gravitational-wave signal, but the objects are sufficiently separated to avoid merging.It’s no coincidence that this resembles elementary particles deflecting each other. “You can use the techniques developed for the scattering of microscopic objects to describe this scattering of macroscopic ones,” Plefka says. Considered from far enough away, well beyond the event horizon—that pivotal region within which neither matter nor light can escape—a black hole can be modeled as a particlelike point with mass and spin, albeit one that generates gravitational rather than electromagnetic waves.On that basis, Plefka and his colleagues applied techniques from quantum field theory that are more typically used to analyze the behavior of elementary particles. “We’re building on decades of work that has been done to make predictions for collider experiments,” says Gustav Mogull, a particle physicist at Queen Mary University of London and one of Plefka’s co-authors.Closer to Complex RealitiesThe team’s goal was to bring its numerical approximations as close as possible to mirroring reality—which, of course, tends to be more messy. To do that, Mogull, Plefka and their team toiled to crank up the complexity of their calculations. In this work, the researchers incorporated five levels of that complexity—to what is known as the fifth post-Minkowskian order of precision—for describing the scattering angles of black hole pairs, their radiated energies and their recoils.This is where the Calabi-Yau geometric structures, normally associated with string theory, come in. In string theory, Calabi-Yau geometries involve the compactification of higher dimensions. Here, they are not merely abstractions but instead emerge from the researchers’ calculations of black hole scattering. It’s perhaps ironic that string theory, notoriously derided as untestable, has given rise to mathematical structures of relevance to measurable physics far from the rarefied realm of strings.A visualization of two black holes scattering off each other and emitting gravitational waves, which are rendered in shades of blue. This visualization was computed with the aid of advanced mathematical functions known as Calabi-Yau manifolds.Mathias Driesse/Humboldt Universtität zu BerlinAny mathematical function is associated with some kind of geometry, Mogull explains—and as the function increases in complexity, so, too, does its geometry. In the case of something basic, such as the sine or cosine functions used in trigonometry, that geometry is a simple circle. Elliptic functions, on the other hand, imply a doughnut-shaped geometry called a torus, which is also a Calabi-Yau onefold. It turns out that the functions Mogull, Plefka and their team developed for black hole scattering are associated with Calabi-Yau threefold structures, which involve six-dimensional surfaces. “I don’t think the appearance of Calabi-Yaus was that unexpected within our community. I would say this represents confirmation of something that people had suspected but was yet to be verified,” Mogull says.To demonstrate the utility of their approach in their study, Plefka, Mogull and their colleagues compare their approximations of black hole scattering angles to other, presumably more precise ones that were derived from numerical simulations. Such simulations can be time-consuming to run, even on state-of-the-art supercomputers—hence the search for accurate approximations. The team’s highest-order approximation closely matches the results of supercomputer number crunching for cases of black holes that gently deflect each other across great distances. But when the black holes come closer to a head-on collision, the team’s calculations begin diverging from the numerical simulations.Such work may seem to be a purely academic exercise, but in fact, the research could prove vital for making new discoveries. Signals from scattering black holes and neutron stars should be within reach of the next generation of gravitational-wave detectors that are set to come online in the late 2030s. These detectors, which will also need a new generation of models called waveform templates to discern true gravitational-wave signals from a sea of cosmic and terrestrial noise, include the proposed Einstein Telescope in Europe and Cosmic Explorer in the U.S. The latter, like LIGO, is supported by the National Science Foundation, and so far these kinds of gravitational-wave projects have avoided being directly targeted by the Trump administration’s aggressive proposed cuts to federally funded science.The work’s prospect for enhancing our understanding of gravitational-wave sources excites scientists who are gearing up for this new wave of detectors. They include Jocelyn Read, a physicist at California State University, Fullerton, who works with the Cosmic Explorer project. “Next-generation facilities can measure nearby signals with exquisite fidelity,” she says.“So having very accurate and precise predictions from our current theories is definitely needed to test them against those kinds of future observations,” Read adds.Yet she also urges caution when assessing the significance of Plefka, Mogull and their colleagues’ work. “If they’re talking about implications for gravitational-wave astronomy, there are a few more steps that are needed,” she says. And their team has competitors, too, including some who have been deploying numerical simulations of their own.These kinds of approximate methods could eventually inform the waveform templates that are so crucial for filtering out noise in upcoming gravitational-wave detectors, Plefka says. Geraint Pratten, a LIGO physicist at the University of Birmingham in England, agrees. “I think it’s a heroic calculation by the group. It will provide a lot of insight into how we can structure next-generation waveform models,” he says. Pratten adds that more work will need to be done to move past the new study’s limitations. For example, the paper focuses on black holes without spin and those that undergo “unbound” scattering, which means that they deflect each other and never meet again. In reality, most, if not all, black holes are thought to spin, and usually scattering events precede an eventual merger.But in any case, he believes some gravitational waves from black hole and neutron star deflections will eventually be detectable, such as through observations of globular clusters, where these dense objects are packed together in a small space, cosmically speaking.For Plefka, Mogull and their peers, this macroscopic version of quantum field theory is still a young field, and there are many new types of astrophysically relevant calculations that they and others can do. These esoteric Calabi-Yau structures, formerly on the frontier of theoretical physics, could be just the beginning. “You’ve had this whole new class of mathematical functions—these theoretical things that had appeared in string theory,” Mogull says. “And we’re saying, ‘Look, this is tangible. Thisis something you can try to detect, try to measure. This is real physics now.’” #black #hole #mergers #show #strange

GPU Architecture & Working intuitively explained Author(s): Allohvk Originally published on Towards AI. GPU Origins The image displayed on a computer screen is made up of millions of tiny pixels. In early days, “graphics controllers” were given instructions by the CPU on how to calculate the individual pixel values so that the appropriate image could be displayed. These were ok for conventional displays but for a really good gaming experience, images need to be built dozens of times per second. The CPU was not really designed to handle these kind of loads. The whole process of creating the image could be parallelized big-time simply by (a) dividing the image into smaller blocks (b) carrying out computations for each block in parallel & (c) grouping them back again. The results of one block don’t influence the results of the other blocks. CPU’s multi-threading capabilities was not really conceived for such massive parallelization. Enter the GPU! Sony first used the term GPU in 1994, in its PlayStation consoles. The technology was perfected by NVIDIA which soon became a leader. GPUs have numerous computation cores (much more than a CPU) and gaming programmers could write Shaders — programs to run graphics computations on the GPU in a massively parallelized way to create the screen images in super-fast time. The GPU is inspired by the CPU but was specifically designed to enable massive multi-threaded operations on its numerous computation cores seamlessly. Creating threads, switching between threads etc is much faster on a GPU. Some smart developers also realized that these parallel processing capabilities could be used for other computationally intensive tasks as well! 2005: Steinkrau implements a simple 2-layer Neural Net on a GPU 2006: Kumar et. al. trains a CNN model for document processing 2007: NVIDIA released Compute Unified Device Architecture (CUDA) — a custom language extending C to exploit data parallelism on GPUs. Now developers had much more granular control over the image rendering. 2008 a landmark paper by Raina et al was released. This paper pretty much showed everyone how to train deep layers on a GPU 2014: NVIDIA released CuDNN — a dedicated CUDA library for Deep Learning. Very soon PyTorch, TensorFlow etc incorporated CuDNN, setting the stage for modern GPU usage for AI! A GPU is an ASIC or Application-Specific Integrated Circuit having a processor (hosting numerous computational cores), a memory soldered onto it (we want to avoid going to the CPU RAM for everything), a cooling system (well, they heat up pretty fast) and a BIOS chip (same role as a CPU — to store settings, run startup diagnostics etc). This card is then plugged into the motherboard slot using the PCI Express interface. The terms GPU and graphics card are often used interchangeably. Some GPUs like the one in Apple M3 do not have a dedicated memory but instead use the system RAM itself which is possible due to its unique design. Google has the TPU (Tensor Processing Unit) which is its own ASIC. We discuss the GPU memory, the processing cores, the LLM workflows happening inside them & common topologies for clustering. Photo by Thomas Foster on Unsplash 1. GPU Memory module — The VRAM Instead of having the GPU talk to the regular RAM, it made sense to create another RAM physically closer to the GPU die so that data retrieval is faster. So a graphics card has a memory called VRAM — Video Random Access Memory in addition to the computation engines . VRAM is connected to the computation engine cores via a Bus called the memory interface. 1.1 What is DRAM? Let us talk first of RAM technology in general. All memory whether it is the CPU RAM or the GPU VRAM are mostly based on DRAM technology which consists of a capacitor and a transistor. The capacitor’s charge represents the data stored. Due to its very nature, this charge gradually leaks. To prevent data loss, a refresh circuit periodically rewrites the data back, restoring its charge. Hence the name — Dynamic RAM due to these preiodic refreshes. Most computers use Synchronous DDR5 DRAM’s as their CPU RAMs. Synchronous because it utilizes the system clock for better performance. In other words the action (of retrieving & storing data) is operationally coordinated by an external clock signal. Tying the operations to the clock makes it faster. The processor knows the exact timing & number of cycles in which the data will be available from the RAM to the bus & can plan better. We have DDR1 (1st Gen Double Data Rate Synchronous Dynamic RAM released in 2000) to DDR5 which is the choice of CPU RAM as of today. 1.2 What is SGRAM? Let us now talk about the VRAMs in GPUs. The VRAM is a type of SGRAM — Synchronous Graphics RAM. The current generation of VRAMs being used is GDDR6. Yes, this is 6th generation GDDR, the G standing for “Graphics”. While DDR & GDDR share common origins and early couple of generations were similar, the branches separated after DDR3. So as of 2025, DDR5 rules in CPU RAM and GDDR6 rules for consumer-grade GPU RAMs. Conceptually DDRs and GDDRs are similar but note that DDRs are used by CPUs which need low latency whereas GDDRs are used by GPUs which are OK to compromise latency for extremely high throughput. Crudely, the former has more frequent smaller calculations & the latter deals with much higher volume of data & some delays are forgiven considering the vast volumes of data being processed. Even more crudely, the former is a bullet train with 6–8 coaches while the latter a 3 Kilometre long goods train. 1.3 GDDR VRAMs explained in detail GDDR memory are individual chips soldered to the PCB (Printed Circuit Board) very close to the GPU die. The physical proximity improves the speed of data transfer from the VRAM to the GPU processor. There are pins in a GDDR which can be thought of as individual wires that connect it to the processor. Bus width is literally the number of such connections. GDDR6 has 32 pins spread across 2 channels with roughly 16 Gbits.p.s bandwidth per pin. Bandwidth is total amount of data being moved & if you had one single metric at your disposal to take a decision, it would be this. Before we go further, let us try to understand this metric intuitively. 1.4 Calculating GPU Memory Bandwidth intuitively Memory Bandwidth is the max rate at which data can be transferred between the GPU and the VRAM. We discussed that data transmission is synchronized with the clock. The clock cycle is measured in hertz & represents the number of cycles per second. Let us say we have a clock operating at 1000 MHz. This literally means 1 billion clock ticks per second. How long does a tick last? Literally 1/(1 billion) i.e. 1 nano second. Data is sent to and fro every clock cycle. So every nano-second, a bus-full of data is sent from the VRAM to the processor & vice versa. How many seats on the bus? Well, we discussed this earlier… This is the memory interface or the bus width… literally the physical count of bits that fit into the bus. A 128-bit bus would ferry 128 bits every nano-second. The D in G’D’DR6 stands for Double. Basically, data is transmitted on both the rising and falling edges of the clock cycle, so 256 bits every nano-second. How many bytes in 1 sec? 256/8 i.e. 32 billion bytes per second or better still 32 GB/s as Giga is the preferred term when measuring data. The capital B denotes bytes whereas the small b denotes bits… a source of confusion. A more practical formula is: Bandwidth = Clock * Bus Width x Data Rate, where the Data Rate is the number of data transfers per cycle. GDDR6 is Double Data Rate (as just discussed) and Quad pumped, which quadruples the (doubled) speed. So effectively the Data Rate is 8. Sometimes, you may encounter the same information crouched in different semantics. E.g., if frequency of command clock (CK#) is N, then the write command clock (WK#) is 2N. GDDR6 rates then are QDR (quad data rate) in reference to WK# and ODR (Octal Data Rate) in reference to the CK#. Some OEMs multiply the clock speed & data rate & call it a clock rate or something. In that case, the bandwidth is simply that number multiplied by the bus width. In general, this raw formula can be used: num_of_transfers per second * num_of_bits per transfer / 8. “Boost clock” mechanism allows the GPU and GDDR memory to operate at even higher speeds than the default clock when conditions allow it. Boost clock metric refers to the max such operating clock speed. A 1750 MHz clock means: 1.75GHz is the frequency of command clock(CK#). The frequency of the write clock (WK#) is 3.5GHz due to the G”D”DR The Quad pumping takes it to 3.5*4=14 G bits moved in 1 second from each pin on the bus. We could have bus widths of up to 384 bits! So we get a bandwidth of 14*384 Giga bits per second. Divide by 8 to get 672 GB/s. GDDR6 bandwidth can go upto 1 TB/s. Wow! 1.5 What is HBM VRAM in a GPU? When reading or writing data, contention is created when the VRAM has occupied memory channels & is busy receiving or delivering other data. This contention creates latency & this affects bandwidth. Increasing the number of memory channels is a great option. A type of memory called HBM (High-Bandwidth Memory) has lower access latency than GDDR6, since it has 8-memory channels versus 2 channels in GDDR6. HBM also has a wider bus. HBM has 1024 pins spread across 8 channels of 128 pins with roughly 2 Gbits.p.s bandwidth per pin. Compare this with (an equivalent) GDDR which has 32 pins spread across 2 channels with roughly 16 Gbits. p.s bandwidth per pin. Notice how HBM keeps the Gbit/sec per pin much lower than GDDR. This saves power (which is important as we shall see). In spite of this, it has higher bandwidth than GDDR6 due to the wider bus & higher channels. As we discussed, a pin is literally a wire connecting the VRAM to the processor. Having 1024 wires connected from the processor to the VRAM is not possible on a standard PCB. Therefore, an “interposer” is used as an intermediary to connect the VRAM & the processor. Just like a regular IC, wires (connections) are etched in this silicon “interposer” in the desired quantity. After this, the HBM device(s) & the processor are mounted atop this “interposer”. The slightly twisted workaround is called a 2.5D architecture.Another difference is that while GDDR chips are soldered to the PCB surrounding the GPU die, an HBM structure is a vertical stack of DRAMs like a high rise building. The stacked memory dies are linked using microscopic wires with TSV (Through-Silicon Vias) which are vertical electrical connections giving super fast connectivity between the DRAMs. There are huge challenges to stacking items vertically especially around designing heat sinks & managing thermal safety but somehow HBM manufacturers have made this happen. HBM has become a gold standard today for AI data centers. It was introduced to the Market by SK Hynix in 2013. Today, we have the 3rd generation HBM3 and their main client is Nvidia. Due to investments made way back, SK Hynix is leading the pack along with Samsung and a relatively recent entrant named Micron. We hear a lot about chips and TSMC but HBM is a key technology to watch out for in the coming years. We typically have more than one HBM devices inside the GPU die. GDDR6 co-exists with HBM3. The markets are complementary. The former addresses PCs & other consumer GPUs whereas the latter addresses data center GPUs. Ultra large scale AI deployments like ChatGPT likely leverage the use of a cluster of NVIDIA GPUs working in tandem. Connecting such GPU’s involves the use of NVIDIA NVLink technology which requires fast GPU memory bandwidth speeds and it’s the reason why HBM is prevalent in such systems. If not for the wide bus width and fast data transfer rates offered by HBM, these kind of clusters would be very difficult to design. Besides the VRAM, GPUs also include high-speed memory caches that are even closer to the GPU’s processing cores. There is a physical limit to the sizes of these caches. An L1 cache is usually in KB and an L2 cache is usually a few MB. Different hardware & software strategies exist to keep the most useful, and most reused data present in caches. 2. Cooling Mechanisms in a GPU Higher clock speeds generally result in increased heat generation necessitating the need for cooling solutions to maintain optimal operating temperatures. Usual cooling methods are: Passive Cooling: These do not have any powered moving components. They take advantage of optimized airflow to take heat away. Fans are used to dissipate heat by blowing cool air across the heat sinks, which are metal components designed to absorb & disperse heat In water cooling, water is circulated through the GPU surface using pipes & a radiator. The hot liquid running through the pipes is in turn cooled down by the radiator fan. Hybrid cooling — which uses a combination of the above 3. GPU Computation cores — Processors Let us now talk about the processors on the GPU. Unlike CPUs which contain only a few cores, the GPU literally has 1000’s of cores & specializes in running tasks in parallel across these cores using SIMD (Single Instruction, Multiple Data) units. Let us stick to NVIDIA terminology. There are multiple processing units called Streaming Multiprocessor (SM) on a NVIDIA GPU. For e.g. an H100 has upto 144 SMs. What is inside an SM? Well there are mainly 2 type of execution units — CUDA cores & Tensor cores. There is also a small memory SRAM which is Shared between all threads running in that SM. More specifically, every SM has a few KB memory that is partitioned between L1 cache & Shared Memory usage. 3.1 CUDA core versus Tensor core in a GPU — The difference Tensor cores are a pretty recent innovation (from V100 onwards) and are specifically designed for faster matrix multiplication. Let us discuss CUDA cores first. These are the computation engines for regular math operations. Each CUDA core can execute one operation per clock cycle. But their strength lies in parallel processing. Many CUDA cores working together can accelerate computation by executing processes in parallel. Tensor Cores are specialized hardware units designed to accelerate “mixed precision” training. The earliest version allowed 4×4 FP16 matrices to be multiplied & added to an FP32 output matrix. By using lower-precision FP16 inputs in the computations, the calculations are vastly accelarated & by retaining FP32 outputs for the rest of the procedure, accuracy is not compromised too much. Modern tensor cores use even lower precision formats in DL computations. See this for more details. There may also specialized units like the transformer engine designed to accelerate models built with the Transformer blocks. A single GPU can be partitioned into multiple fully contained and isolated instances, with their own memory, cache & cores via MIG or Multi Instance GPU technology. 3.2 GPU operations — A FLOP show Let us now talk about actual operations. A FLOP (Floating Point Operation) is a single floating-point calculation like an addition. Performance of a GPU is usually measured in TeraFLOP/s. Tera is a trillion, FLOP stands for floating-point operations and the ‘s’ stands for per second. Most matrix ops involve a multiply and an add. It makes sense to fuse these ops together to get an Fused Multiply-Add (FMA) op. If we know the FMA speed, we can simply double it to get the FLOP counts per clock. To get the peak FLOP/s rate, we multiply this by the clock rate & the number of SMs. Note that we have FP16, FP32, FP64 & Int8 cores with varying speeds. For e.g.: Say there are 4 tensor cores in each SM & 114 SMs in an H100 Say each tensor core delivers 512 FP16 FMA ops per clock. Careful here: Read the specs clearly to check whether the FMA ops per clock metric is per SM or per individual core. For e.g., this link of A100 is per coreper SM Let the Clock speed = 1620 MHz So TFLOP/s = 1620 * (2*512) * 4 * 114= 756 TFLOP/s of performance! 756 Trillion operations per second. Wow! What would Babbage say to that? 4. Putting everything together — LLM Operations in a GPU Given this immense compute-power, we can now make a reasonable guess that LLM inference is memory-I0 bound, not compute bound. In other words, it takes more time to load data to the GPU’s compute cores than it does for those cores to perform LLM computations on that data. The processing itself is super-fast & there is enough & more compute power available. To start with, the training data needs to be downloaded from a remote source to the CPU memory From there, it needs to be transferred to the GPU via the system bus and PCIe bus. The host(CPU)-to-device(GPU) bandwidth is limited by the CPU frequency, PCIe bus, GPU devices & the number of PCIe lanes available. Once the data & weights are in the GPU VRAM, they are then ferried across to the SRAM where the processors perform operations on it. After the operation the data is moved back to the VRAM & from there it is moved back to the CPU RAM. This is a rather simplistic view. Inside the GPU, the tensors are repeatedly moved back and forth between VRAM & SRAM (the memory allocated to an SM). Can you guess why? We saw that SRAM size is in KB so large matrices are not going to fit in there … which explains why there is a constant movement between VRAM which holds all the tensors and SRAM which holds the data on which compute operations are performed. So there is typically a memory-op where tensors are moved from VRAM to SRAM, then a compute-op SRAM and memory-op to move tensors back from SRAM to VRAM. Computations like a matrix multiplication involving 2 large matrices need several such memory + compute ops before the action is completed. During the training of GPT-3, the tensor cores on the GPUs used were found to be idle ~50% of the time. So, to extract the best from the infrastructure, data movement needs to be fast enough to ensure the computation cores are kept reasonably occupied. Surely, there is scope for some smart person to come up with shortcuts. Enter Flash attention & other such hacks. But that is a story for another day! 5. Linking GPUs for LLM training — Topologies While LLM inferencing is manegable with a readymade collection of GPUs such as a DGX server (contains 8 H100s), LLM training needs far more GPUs. Before we discuss how to connect GPUs for larger workloads, it makes sense to see how CPU servers are connected in a datacentre. I am not an expert in this area, so please feel free to point out any incorrect interpretations I may have made from the references I quote. 5.1 Generic concepts on linking processors Each server has a card attached to it called the Network Interface Card (NIC). RDMA technology enables direct memory access to a remote server via the NIC hardware. RoCE (RDMA over Converged Ethernet) protocol uses the RDMA technology & adapts it to Ethernet networks. So now, a server can talk to a remote server over a network. A network switch is a device connecting multiple servers in a network, enabling them to communicate with each other. This is the basic technology. Now let us come to the topology. So we assemble all the servers physically in one place and pile them up vertically them in neat racks.A very basic topology is to connect each server in a rack to a switch that usually sits on Top of the Rack, aptly named the ToR switch. The ToR switches of different racks are connected to a Spine switch. This topology is a basic implementation of Clos topology — named after Charles Clos who invented this scheme to originally arrange telephone nodes in a “leaf-n-spine” arrangement. The leaf switches are nothing but the ToR switches in modern data centers. Source: Fig 1–1 from https://www.oreilly.com/library/view/bgp-in-the/9781491983416/ch01.html Fat tree is a variant of Clos. Like before, we have servers arranged into racks connecting to Top-of-the-Rack (ToR) switches. ToR switches are connected to the aggregation switches to provide connectivity across racks, forming a pod. The pods are interconnected with spine switches, allowing any-to-any communication across servers. To be noted is the fact that there are multiple paths connecting servers. So there is lot of redundancy built-in. In a typical App deployment running hundreds of microservices on dozens of servers, it is useful to have such fully connected, high bandwidth networks. You never know who is going to talk to whom so it never hurts to overprovision on bandwidth and connectivity. However, network loads during AI training do not follow these patterns. They are more predictable & this allows us to build optimized, cheaper & less power-hungry networks. 5.2 Linking GPUs via proprietary technology like NVLink We can strap together H100’s by leveraging the proprietary NVLink & NVSwitch technologies. NVLink provides the high-speed connection between individual GPUs, while NVSwitch is a chip that enables multiple GPUs to communicate through NVLink, forming a high-bandwidth network. See this nice article for details. NVIDIA’s P100 GPU introduced the NVLink1. At that time there was no NVSwitch chip, and the GPUs were connected in a ring-like configuration, which resulted in a lack of direct point-to-point communication between GPUs. The NVSwitch1 chip was introduced with the V100, followed by the NVSwitch2 chip with the A100 GPU. We are in the third-generation NVSwitch3 which can support a cluster of up to 256 H100 GPUs. Each H100 GPU in such a cluster is connected to the internal NVSwitch3 chip through 18 NVLink4.0 connections. This is how trillion parameter LLMs are inferenced. 5.3 Linking GPUs via RoCE in a rail-optimized topology But as they say, ye dil mange more… Meta reportedly trains its newer models on a cluster that’s over 100K H100’s. Phew! How to they manage to link it all up? The standard NVLink tricks can only scale to a limited number of GPUs. Beyond that, we have to use the network topologies discussed earlier & fall back on technologies like RoCE, which allows data to be directly transferred from one GPU’s memory to another without involving the CPU. So you have 8 GPUs in one DGX server. You have several such DGX servers in the data centre. Each GPU is assigned a NIC (yes!) & connected via RDMA to all other GPUs thru’ a variant of Clos network called “rail-optimized network”. The idea here is to set up dedicated connections between groups of GPUs with rail switches. If a GPU wants to communicate with a GPU which is in a different group, then it has to go thru’ the spine switch (which takes a lil more time). To implement this, each GPU in a DGX server is indexed serially. A rail is the set of GPUs with the same index on different servers & these are interconnected with a rail switch via RDMA. These rail switches are subsequently connected to spine switches forming any-to-any GPU network. Source: Fig 1 from https://arxiv.org/pdf/2307.12169 This topology streamlines traffic flow. It is like having dedicated lanes for high speed vehicles instead of generally mixing all traffic together. Rail paths are direct connections between a bunch of GPUs with same index. Spine switches serve as the connecting points for differently-indexed GPUs. For e.g., communication between GPU1 of server 1 and GPU1 of server 2 happens via their dedicated rail switch 1. If GPU1 of server 1 needs to reach GPU5 of another server, it has to go thru’ a spine switch. The workloads are designed so as to minimize data transfers across rails (since it has to go thru’ the extra spine switch). The good news is that this can be neatly done for AI training ensuring that most of the traffic stays within the rails, and does not cut across. In fact, there is a recent paper which suggests that you can consider removing costly spine switches altogether as inter-rail communication is minimal. Can you guess how? 5.4 Linking GPUs via RoCE in a rail-only topology Well, we have the superfast connectivity using NVLink to communicate between a limited set of GPUs (upto 256). So you create these High Bandwith (HB) domains which use NVLink for communication. You have several such HB domains. We then have the same indexing system and rail connections to interconnect the HB domains. But there are no spine switches! Can you guess how GPU1 of HB domain 1 can talk to GPU5 of another HB domain? Yes! Transfer data via superfast NVLink to GPU5 of HB domain 1 first. Then use the dedicated rail of GPU5 to talk to the GPU5 in another HB domain! This is a rail-only topology as oppsed to rail-optimized topology! Given these topologies, we can now plan the training pipeline to have pipeline parallelism, tensor parallelism &/or data parallelism but that is a story for another day. See this, this & this for more details. 100K H100’s consume a LOT of power. Tech companies are exploring nuclear power options to generate clean energy needed for long term sustenance. Else, a 100K GPU cluster may have to be broken down to smaller clusters and connected using optical transceivers across the buildings in a campus. This (unplanned) article is a prelude to — Optimizing LLM inference: Key Faultlines & workarounds. To deeply understand how we can optimize LLM operations, we need to understand more about the silicon on which they are executed. Though there are lots of manuals/guides on individual aspects like memory, processors, networking etc, I couldn’t find a concise and reader-friendly thread linking together these various aspects & hence took a shot. This is the 9th of a 15-series article titled My LLM diaries. LLM Quantization — From concepts to implementation LoRA & its newer variants explained like never before In-Context learning: The greatest magic show in the kingdom of LLMs RAG in plain English — Summary of 100+ papers HNSW — Story of the world’s most popular Vector search algorithm VectorDB origins, Vamana & on-disk Vector search algorithms Taming LLMs — A study of few popular techniques Understanding LLM Agents: Concepts, Patterns & Frameworks Anatomy of a GPU — A peek into the hardware fuelling LLM operations Optimizing LLM Inference — Key Faultlines & workarounds LLM Serving — Architecture considerations LLM evaluation & other odds and ends Look Ma, LLMs without Prompt Engineering LLMs on the laptop — A peek into the Silicon Taking a step back — On model sentience, conscientiousness & other philosophical aspects Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor. Published via Towards AI المصدر: https://towardsai.net/p/machine-learning/gpu-architecture-working-intuitively-explained