Meta tags:
Headings (most frequently used words):
in, and, popular, nvidia, the, blackwell, ai, to, with, how, gives, gpu, gets, store, compute, at, gpus, clouds, new, for, newsletter, boost, search, aws, it, pushing, platforms, system, operating, server, sparks, linux, outgrows, cloud, meta, chiplet, bandwidth, sign, buildout, up, our, massive, inside, supercomputers, waferscale, hexamesh, topologies, more, database, edges, muscles, fpga, spartan, flexing, amd, fray, analysis, third, inspired, pci, vector, express, joins, switches, couchbase, nature, retimers, by, gen, parameter, hyperscale, spend, research, switching, ethernet, over, take, odms, until, long, xpus, on, can, images, so, billion, 75, anthropic, amazon, did, than, better, build, generative, 3d, goes, supercomputer, cerebras, harder, competing, easier, cheaper, models, trillion, attack, systems, ceiba, speed, reboot, their, superchip, grace, bringing, hyperscalers, materials, of, range, expanding, about,
Text of the page (most frequently used words):
the (62), and (53), #march (28), 2024 (25), #nvidia (18), for (17), #compute (17), 2023 (16), with (12), that (12), #blackwell (11), new (11), #cloud (10), #memory (10), gpu (10), can (9), #more (9), how (8), 2021 (8), are (8), #gpus (7), connect (7), better (7), why (7), #platform (6), 2022 (6), what (6), datacenter (6), when (6), amd (6), you (6), amazon (6), arm (6), systems (6), popular (6), than (6), engines (5), next (5), its (5), money (5), needs (5), harder (5), have (5), now (5), their (5), may (5), millimeters (5), system (4), store (4), december (4), bring (4), crazy (4), july (4), models (4), way (4), risc (4), database (4), october (4), getting (4), world (4), has (4), search (4), would (4), but (4), anthropic (4), switching (4), gives (4), generative (4), materials (4), hpc (4), aws (4), about (4), then (4), from (4), your (4), supercomputers (4), clouds (4), hyperscale (4), september (4), data (4), push (3), ethernet (3), where (3), boost (3), other (3), silicon (3), performance (3), google (3), images (3), year (3), down (3), did (3), comes (3), supercomputer (3), grace (3), there (3), over (3), half (3), problem (3), software (3), chiplet (3), which (3), operating (3), ceiba (3), lot (3), long (3), chip (3), take (3), november (3), build (3), being (3), platforms (3), all (3), not (3), future (3), centers (3), large (3), storage (3), good (3), easy (3), analysis (3), designs (3), week (3), newsletter (3), cheaper (3), bringing (2), enterprises (2), databases (2), reboot (2), vector (2), really (2), wafers (2), subscribe (2), research (2), highlights (2), speed (2), diameter (2), servers (2), were (2), inbox (2), massive (2), featuring (2), this (2), our (2), cerebras (2), been (2), need (2), between (2), odms (2), parameter (2), openai (2), trillion (2), leave (2), edge (2), attack (2), people (2), web (2), superchip (2), lots (2), control (2), fabric (2), model (2), day (2), going (2), nvlink (2), high (2), technologies (2), hyperscalers (2), interconnect (2), goes (2), extreme (2), post (2), few (2), means (2), big (2), directly (2), services (2), back (2), put (2), time (2), meta (2), range (2), expanding (2), stories (2), lithography (2), ultraviolet (2), gets (2), exascale (2), until (2), nothing (2), eights (2), cxl (2), mostly (2), instinct (2), business (2), coiled (2), spring (2), april (2), language (2), invests (2), picks (2), haven (2), roadmaps (2), changed (2), stop (2), january (2), cost (2), llms (2), heavily (2), move (2), like (2), doe (2), peeks (2), first (2), undisputed (2), buy (2), explains (2), architecture (2), champ (2), shove (2), vmware (2), unix (2), buys (2), intel (2), 1486 (2), profitable (2), pace (2), pervasive (2), hierarchies (2), processing (2), applications (2), subsystems (2), home (2), xpus (2), spend (2), get (2), billion (2), technology (2), networking (2), game (2), hyperconverged (2), two (2), had (2), server (2), gtc, sign, historical, some, books, chief, events, founder, conference, technical, his, efa2, company, create, email, editor, jointly, built, here, contact, based, warranted, called, name, implies, contributors, gh200, peak, adapter, hopper, cpu, complexes, brought, sneak, got, switch, together, sales, last, elastic, context, properties, executive, used, invest, they, sell, campus, pulling, seeing, corporate, support, budgets, didn, damper, investments, talking, video, personal, generating, training, microsoft, division, anything, architectures, enterprise, latest, movement, because, looking, spending, kinds, various, integer, meaning, point, floating, focus, while, invested, also, pony, copyright, such, officer, organizations, cookies, functional, finding, increasing, rapidly, meet, size, catalysts, privacy, providers, promising, accompanying, maker, unveiled, huang, jensen, batteries, uses, takes, fastest, longer, most, case, information, quantum, determine, specialized, using, running, major, simulations, computer, requires, often, discovery, design, around, researchers, goal, competing, apparently, easier, edges, them, talk, though, even, aren, before, noted, muscles, trace, fpga, spartan, flexing, armed, analyzed, processed, stored, generated, fact, origins, supercomputing, along, reticle, fray, content, joins, couchbase, past, learning, machine, ramp, decades, began, popularity, end, ask, computing, who, depends, exactly, increasingly, strategies, manufacturing, faster, catch, let, could, machines, top, adding, those, moratorium, four, still, publication, register, whole, things, utilization, partnership, public, place, whom, makers, read, includes, published, vendors, infrastructure, stackhouse, become, pushing, continuing, publishing, bandwidth, inc, retimers, switches, express, pci, limit, pretty, linkedin, pity, promised, millimeter, 450, 300, larger, any, make, buildout, inside, out, does, mainstream, genai, magnitude, orders, one, ago, rss, per, vaults, too, help, bottle, genie, maybe, luck, bags, facebook, several, able, turf, twitter, want, waferscale, gen, third, 25x, change, executions, much, techniques, nature, inspired, topologies, hexamesh, needed, depth, cut, sizes, every, transistor, nanometer, coverage, intended, pun, stone, set, ultimately, offers, trillions, handful, hundreds, handle, sparks, linux, outgrows, viable, commercially, only, constantly, rankings, datastores, different, 418, becomes, stack, evolving, code,
Text of the page (random words):
the next platform the next platform home compute store connect control code ai hpc enterprise hyperscale cloud edge latest march 28 2024 how to build a better blackwell gpu than nvidia did connect march 27 2024 amazon gives anthropic 2 75 billion so it can spend it on aws xpus ai march 25 2024 how long until the odms take over ethernet switching connect march 21 2024 nvidia research gives generative ai images and 3d a speed boost ai march 21 2024 expanding the search for a range of new materials hpc march 21 2024 hyperscalers bringing nvidia s grace blackwell superchip to their clouds cloud march 20 2024 aws and nvidia reboot ceiba supercomputer for blackwell gpus cloud march 19 2024 how nvidia blackwell systems attack 1 trillion parameter ai models ai search for connect how to build a better blackwell gpu than nvidia did march 28 2024 0 while a lot of people focus on the floating point and integer processing architectures of various kinds of compute engines we are spending more and more of our time looking at memory hierarchies and interconnect hierarchies and that is because compute is easy and data movement and memory are getting ai amazon gives anthropic 2 75 billion so it can spend it on aws xpus march 27 2024 0 if microsoft has the half of openai that didn t leave then amazon and its amazon web services cloud division needs the half of openai that did leave meaning anthropic and that means amazon needs to pony up a lot more money than google which has also invested in anthropic connect how long until the odms take over ethernet switching march 25 2024 5 for more than a year now we have been talking about how investments in ai servers have put a damper on budgets for servers used to support other corporate applications and we may be seeing enterprises pulling back on campus switching so they can invest more in datacenter switching to ai nvidia research gives generative ai images and 3d a speed boost march 21 2024 3 time is money when it comes to generative ai as is the case with most technologies but with generative ai it is big money and the longer it takes for an ai model to do anything such as training on data or generating images and video the more money is hpc expanding the search for a range of new materials march 21 2024 2 finding new functional materials for batteries and catalysts and lots of other uses is a major goal of researchers around the world and the design and discovery of new materials often requires computer simulations running on the world s fastest supercomputers using specialized software that can determine properties at the quantum cloud hyperscalers bringing nvidia s grace blackwell superchip to their clouds march 21 2024 5 at his company s gtc 2024 technical conference this week nvidia co founder and chief executive officer jensen huang unveiled the chip maker s massive blackwell gpus and accompanying nvlink networking systems promising a future where hyperscale cloud providers hpc centers and other organizations of size and means can meet the rapidly increasing cloud aws and nvidia reboot ceiba supercomputer for blackwell gpus march 20 2024 1 last november we got a sneak peak at a supercomputer called ceiba that would be built jointly by nvidia and amazon web services based on its then new gh200 grace hopper cpu gpu compute complexes its nvlink switch 3 memory fabric for gpus and the elastic fabric adapter efa2 ethernet interconnect that is sign up to our newsletter featuring highlights analysis and stories from the week directly from us to your inbox with nothing in between subscribe now more analysis ai how nvidia blackwell systems attack 1 trillion parameter ai models march 19 2024 1 we like datacenter compute engines here at the next platform but as the name implies what we really like are platforms how compute storage networking and systems software are brought together to create a platform on which to build applications some historical context is warranted to put the blackwell ai with blackwell gpus ai gets cheaper and easier competing with nvidia gets harder march 18 2024 12 if you want to take on nvidia on its home turf of ai processing then you had better bring more than your a game you better bring your a game several vaults of money and a few bags of good luck maybe a genie in a bottle would help too ai cerebras goes hyperscale with third gen waferscale supercomputers march 14 2024 7 it is a pity that we can t make silicon wafers any larger than 300 millimeters in diameter if there were 450 millimeter diameter silicon wafers as we were promised so long ago this being pi day and all then cerebras systems would have been able to get a 2 25x performance ai inside the massive gpu buildout at meta platforms march 13 2024 2 if you handle hundreds of trillions of ai model executions per day and are going to change that by one or two orders of magnitude as genai goes mainstream you are going to need gpus lots of gpus and apparently meta platforms does and it is getting out its big control the cloud outgrows linux and sparks a new operating system march 12 2024 42 ultimately every problem in the constantly evolving it software stack becomes a database problem which is why there are 418 different databases and datastores in the db engines rankings and there are really only a handful of commercially viable operating systems but what if the operating system is the problem connect hexamesh chiplet topologies inspired by nature march 11 2024 1 with the reticle limit for chip manufacturing pretty much set in stone pun intended at 26 millimeters by 33 millimeters down to 2 nanometer transistor sizes with extreme ultraviolet lithography techniques and being cut in half to 26 millimeters by 16 5 millimeters for the high na extreme ultraviolet lithography needed to store couchbase joins the vector search in database fray march 8 2024 1 as we have noted before vector databases aren t new even though people talk about them that way and in fact can trace their origins back a few decades exactly when depends on who you ask but their popularity began to ramp along with machine learning technologies over the past 10 compute amd flexing spartan fpga muscles in clouds and at edges march 8 2024 1 the edge is continuing to become a place where it infrastructure vendors need to be and that includes chip makers all of whom have strategies to push their silicon to where the data is increasingly being generated and needs to be stored processed and analyzed amd armed with the technology connect pushing pci express switches and retimers to boost server bandwidth march 6 2024 5 things would go a whole lot better for server designs if we had a two year or better still a four year moratorium on adding faster compute engines to machines that way we could let memory subsystems and i o subsystems catch up and get better utilization of those compute engines popular in ai chiplet cloud can bring the cost of llms way down july 12 2023 18 nvidia picks up the pace for datacenter roadmaps october 12 2023 18 the crazy eights of large language models april 10 2023 17 amd s instinct gpu business is coiled to spring november 1 2023 17 when push comes to shove google invests heavily in gpu compute may 11 2023 15 popular in compute arm is the new risc unix risc v is the new arm september 22 2022 24 amd is the undisputed datacenter gpu performance champ for now december 6 2023 20 the first peeks at the doe post exascale supercomputers october 2 2023 19 arm s v9 architecture explains why nvidia needs to buy it march 30 2021 18 crazy move 1486 what if intel buys vmware may 13 2021 18 popular in store the future of system memory is mostly cxl july 5 2022 16 the world has changed why haven t database designs march 25 2021 15 what do we do when compute and memory stop getting cheaper january 18 2023 14 compute is easy memory is harder and harder december 13 2022 13 if hyperconverged storage is so good why is it not pervasive and profitable september 2 2021 10 popular in ai chiplet cloud can bring the cost of llms way down july 12 2023 18 nvidia picks up the pace for datacenter roadmaps october 12 2023 18 the crazy eights of large language models april 10 2023 17 amd s instinct gpu business is coiled to spring november 1 2023 17 when push comes to shove google invests heavily in gpu compute may 11 2023 15 popular in compute arm is the new risc unix risc v is the new arm september 22 2022 24 amd is the undisputed datacenter gpu performance champ for now december 6 2023 20 the first peeks at the doe post exascale supercomputers october 2 2023 19 arm s v9 architecture explains why nvidia needs to buy it march 30 2021 18 crazy move 1486 what if intel buys vmware may 13 2021 18 popular in store the future of system memory is mostly cxl july 5 2022 16 the world has changed why haven t database designs march 25 2021 15 what do we do when compute and memory stop getting cheaper january 18 2023 14 compute is easy memory is harder and harder december 13 2022 13 if hyperconverged storage is so good why is it not pervasive and profitable september 2 2021 10 about the next platform is published by stackhouse publishing inc in partnership with the uk s top technology publication the register it offers in depth coverage of high end computing at large enterprises supercomputing centers hyperscale data centers and public clouds read more newsletter featuring highlights analysis and stories from the week directly from us to your inbox with nothing in between subscribe now rss twitter facebook linkedin email the editor about contributors contact sales newsletter books events privacy ts cs cookies do not sell my personal information all content copyright the next platform
|