Post has shared content
What do you want to be in future
No votes yet
-
votes visible to Public
0%
Software engineer
0%
Doctor
0%
Richest man
0%
Politician
0%
Real estate

Kingsley the name of your community is saying another thing

Where is the programming

Post has attachment
Qualcomm announces 48-core Falkor CPUs to run Microsoft Windows Server
Today at the Open Compute Summit, Qualcomm announced a significant partnership with Microsoft that could have ramifications for the entire server industry. Not only has Qualcomm created new server designs around Microsoft’s Project Olympus specifications and its own Centriq 2400 SoCs, it’s working with Microsoft to bring ARM support to Windows Server.

Last fall, Microsoft announced a new initiative, dubbed Project Olympus. According to Microsoft, Project Olympus “applies a model of open source collaboration that has been embraced for software but has historically been at odds with the physical demands of developing hardware.” The software firm pledged to release its specifications for cloud hardware designs when they are 50% complete rather than waiting to finalize them. Qualcomm has apparently been one of Microsoft’s major partners, at least as far as ARM servers are concerned.

Qualcomm has been collaborating with Microsoft to bring software support for its platforms to market with Windows Server support. Here’s how the company describes these efforts:

The Qualcomm Centriq 2400 Open Compute Motherboard pairs QDT’s recently announced 10nm, 48-core server processor with the most advanced interfaces for memory, network, and peripherals enabling the OCP community to access and design ARM-based servers for the most common cloud compute workloads. It fits into a standard 1U server system, offering system vendors the flexibility to create innovative, configurable designs for compute-intensive data center workloads. It can be paired with compute accelerators, multi-host NICs, and leading-edge storage technologies such as NVMe to optimize performance for specific workloads.

Qualcomm-Centriq

Qualcomm is hyping up the fact that Falkor is built on 10nm, but with so little additional information on how the core is built or how it performs, there’s little more we can say on that front. Packing 48 cores into a single chip is impressive, but making all of those cores scale and communicate effectively is a difficult challenge on its own. To date, Qualcomm hasn’t said much about how it handled these challenges or offered public data on its benchmark results.

One key caveat to the company’s PR blast is that these servers are apparently intended for Microsoft’s own internal use, and it isn’t clear when Microsoft will launch a commercial Windows Server on ARM product. Still, the prospect of a resurgent ARM backed by a Qualcomm / Microsoft partnership could prove a thorn in Intel’s side. To date, we haven’t seen much movement in this market, and AMD’s decision to put K12 on ice was, in our opinion, the right one — the ARM market simply isn’t mature enough to justify AMD bringing a new server part to market while they try to scale up Naples, its just-announced 32-core architecture. Still, a company Qualcomm’s size could provide the cash influx needed to get the market moving.

Intel, of course, has contended with such threats before. It launched new low-cost Xeon servers based on Atom to fend off competition from the now-defunct Calxeda several years ago, and it successfully reclaimed virtually all of AMD’s server market share. Calxeda, however, was a minor startup, not a multi-billion dollar company in its own right — and AMD clearly thinks it has a chip it can use to take Chipzilla down a peg. Intel could soon find itself facing a war on two fronts, from a resurgent AMD on one side, and a freshly-minted Qualcomm core on the other.






Photo

Post has attachment
Nvidia, Microsoft announce new HGX-1 hyperscale GPU accelerator for AI workloads
Nvidia and Microsoft announced a new system form factor today at the Open Compute Summit for the Open Compute Project. Unlike the venerable ATX standard, this one is designed for data centers and aimed at maximizing GPU performance as part of Microsoft’s Project Olympus initiative.

According to Nvidia, HGX-1 is designed “to meet the exploding demand for AI computing in the cloud — in fields such as autonomous driving, personalized healthcare, superhuman voice recognition, data and video analytics, and molecular simulations.”

Microsoft’s Project Olympus has been pulling in headlines, with hardware launches from Intel, AMD, Qualcomm, and now Nvidia as well. Each of these platforms is intended to accelerate a specific type of workload or scenario. According to Microsoft, Intel’s work with Project Olympus fielded support for Skylake processors, with future versions expected to add support for FPGA accelerators or Intel Nervana solutions. AMD solutions are Naples-centric, as you might expect, and Qualcomm’s focus on its own upcoming 48-core CPU.
HGX-1

An HGX-1 chassis

Nvidia’s Project Olympus platform will pack eight Pascal GPUs (GP100s) into a single chassis, all connected via NVLink, Microsoft said. Up to 32 GPUs can be supported by linking four HGX-1 systems together (it isn’t clear which standard is used to link the systems themselves).

As Patrick Moorhead, of Moor Insights & Strategy points out, the ATX comparison represents how ambitious Nvidia is being with this push. The rollout of ATX in 1995 between Microsoft and Intel gave the computer industry a single, unified form factor to design against. It helped set the stage for an era where system components could be assumed to be compatible with system chassis simply by conforming to the ATX standard. If the HGX-1 standard takes off similarly, future HPC GPUs or CPUs would be able to take advantage of the same kind of guarantees. The HGX-1 standard is designed to allow CPUs and GPUs to connect in whatever ratio suits the workload, all via the NVLink interconnect.

Nvidia doesn’t mention which CPUs this effort is compatible with, and it’ll be interesting to see if we see any AMD-Nvidia team-ups in this area in the future. AMD has its own server infrastructure and graphics division, but the RTG division within AMD operates much more autonomously now than it did in the past, and Nvidia has a vastly larger share of the HPC market than AMD does.

It might make sense for AMD’s server CPU team and Nvidia’s HPC division to work together to expand the cloud computing market, even if the companies are rivals in other business segments. AMD announced its own AI products, dubbed Radeon Instinct, late last year, but has yet to announce any major hardware partners or system designs.

You Might Also LikePowered By ZergNet
Movies That Seriously Damaged Actors' Bodies
Heroes and Villains That Have Completely Destroyed Batman
The Folding Gun That Looks Like a Smartphone is Worrying Police
The Terrifying Girl From 'The Ring' Grew Up To Be Gorgeous
Actors Who Ruined Their Career in a Matter of Seconds
Movies That Seriously Damaged Actors' Bodies
Tagged In

microsoft
amd
nvidia
gpu
qualcomm
ai
hpc
naples
Falkor
NVLink
Radeon Instinct
HPX-1
Project Olympus

Post a Comment 0 Comments
ExtremeTech Newsletter

Subscribe Today to get the latest ExtremeTech news delivered right to your inbox.
Subscribing to a newsletter indicates your consent to our Terms of Use and Privacy Policy.
More Articles

Zelda BOTW The Legend of Zelda: Breath of the Wild really struggles on the Wii U [UDPATE] Mar 8
File Explorer Windows 10 Microsoft now puts ads in Windows 10 File Explorer, because of course Mar 8
ET FreedomPop Sponsored ET deals: Get unlimited data with a Netgear Fuse LTE hotspot from FreedomPop Mar 8
The CIA may be hacking cars, as well as phones and TVs, according to WikiLeaks Mar 8
Qualcomm Centriq Qualcomm announces 48-core Falkor CPUs to run Microsoft Windows Server Mar 8


Photo

Post has attachment
Normally it’s a pretty solid assumption that whatever “revolutionary new battery chemistry” has just hit the news is going to crash and burn — sometimes literally. Maybe the thing uses some outlandishly expensive metal as a catalyst, or it has to be supercooled to be well-behaved, or you have to fastidiously mind the outgassing lest the thing explode. Now there’s a new battery chemistry in town, and it comes from the mind of John B. Goodenough (pictured below): the same guy who came up with the cobalt-oxide cathode that powers the lithium-ion battery chemistry we know and love. Goodenough predicts that the new chemistry will have triple the energy density of lithium-ion cells.Lithium-ion batteries have been in the news lately because things Samsung puts them in tend to explode. They even had some washing machines that apparently blew up out of sympathetic embarrassment. So it’s probably wise to reserve judgment until something can be manufactured at scale. Other battery chemistries have come before, and failed. Lithium-air batteries are a great example of a very interesting battery chemistry that we can’t use, because its development has been hamstrung by engineering problems we can’t yet solve.

This new chemistry has one important difference from the lithium-ion model: It uses sodium instead of lithium. Sodium and lithium are both alkali metals, with the same +1 charge. But sodium is a whole lot more abundant than lithium, which could make the new battery chemistry less expensive than lithium-ion cells.Then there’s the exploding. Lithium-ion batteries are plagued by the formation of metallic lithium “dendrites” that “spread like kudzu” between the anode and the cathode, which causes a runaway reaction and shorts out the cell. Bang. To avoid this, Goodenough’s new battery chemistry uses an annealed glass matrix as an electrolyte. This presents several possible advantages over most liquid electrolytes, namely that it won’t splash horrible battery liquid all over you if the casing is somehow breached. The glass mats also defy the formation of dendrites, because the anode never reacts with the mats.

I’m a big fan of glass mats, because they have extremely shiny physical properties. Goodenough and colleagues used fiberglass sheets as the electrolyte matrix, and electroplated them with metallic sodium (or lithium) as the anode. Their build then packed the remaining cavities with carbon. The contact surface between the metallic anode and the glass mat is so tight that it’s actually classified as wetting.

Goodenough is betting that the new battery chemistry will be, well, good enough [I was waiting for this. -Ed]. In his words, a “safe, low-cost, all-solid-state cell with a huge capacity giving a large energy density and a long cycle life suitable for powering an all-electric road vehicle or for storing electric power from wind or solar energy.”
Photo

THE WORLD'S FIRST SMART CONDOM
The newly announced i.Con “smart condom” is the odds-on heavyweight contender for terrible idea of the week. It is not smart, and it is not a condom, and those are only two of the many facepalm-inducing problems with this product. You can only loosely call it a product, anyway, because it’s so early in development that they aren’t even taking money for preorders — just collecting email addresses to “register one’s interest.”

For a thing presenting itself as a data-enabled smart condom and having the box art it’s got, they sure are taking poetic liberties in calling it a condom in the first place. It’s not a condom. The i.Con is a synthetic rubber “condom ring” that’s meant to sit outside of and atop a condom, in order to use its “Nano-Chip Technology” to perform feats of techno-sexual sophistication. And you still have to buy condoms anyway.

The device is supposed to log “calories burnt, duration of intercourse, how many thrusts (averaged), girth measurements, and various other pieces of data.” Then you’ll have the option of sharing your results with “friends, or, indeed the world. You will be able to anonymously access stats that you can compare with i.Con users worldwide.”

“Have you ever wondered how many calories you’re burning during intercourse?” British Condoms wants to know. Their curiosity is burning. “How many thrusts? Speed of your thrusts? The duration of your sessions? Frequency? How many different positions you use in the period of a week, month or year? Ever wondered how you stack up to other people from around the world?”

No, British Condoms, I never had. I would have made it to 30 without ever having visited upon me the specter of paying someone for the privilege of crowdsourcing my sexual insecurity and/or narcissism. I was doing fine without having sent that data, presumably in plain text, to your server somewhere in the cloud from where it will no doubt be immediately sold to your “trusted partners.”

Just what everyone wants: a device capable of compromising your phone and leaking enough data to triangulate your present location, social security number, and, now, penis size. It’s like buzzwords have achieved sentience and are now mocking us. Can you even imagine the advertising herpes that’s going to result from this? Will it have a little Bluetooth conference with your smart fridge and smart toaster over some common protocol, so they can all collude to warn you via SMS that you’ll need to pick up eggs for breakfast the morning after? What happens when your tech-savvy ex hacks your smart condom? How will this data factor into next-gen credit scores?

But that’s not all they want to do. No no. There’s more.

Lead engineer and spokesperson Adam Leverson said in a statement that the device will feature “built-in indicators to alert the users to any potential STI’s present.” But wait. Doesn’t that make it a wearable medical device?

Naturally, the company hasn’t so much as hinted at the actual device specs, other than saying it has an adjustable carbon fiber band, because I guess carbon fiber is high-tech and that should balance out the squick factor of “synthetic rubber.” You can guess it’d have to include an accelerometer, and they probably mean to include a six-axis [the jokes are writing themselves at this point -Ed] because otherwise it’ll be tough to tell what positions you’ve been in. But how exactly are you supposed to sanitize something with a micro-USB port? Do you have to buy a new one for every new partner? What microfluidics wizardry (ahem) are they intending to use to make this theorized STI indicator refillable or reusable? What database security are they going to use to ensure their STI results aren’t leaked to the public? How are they going to get this device past the British version of the FDA?

One of two things is going to happen, and we’re going to find out which soon. Either this is somebody’s cute hoax or concept design, in which case bully for them, or this product is going to be canceled before it ever hits the shelves.
Wait while more posts are being loaded