ISPgeeks.com

No, Canada! ISPs ordered to out users in Hurt Locker case

Published a day ago - by Jacqui Cheng | Posted in: Law & Disorder

If you thought North American P2P lawsuits were limited to the US, you thought wrong. Three Canadian ISPs have been ordered to disclose the identities of accused file sharers, bringing the IP address debate to the Great White North.

Canadian Internet users are getting a taste of the P2P lawyering that had previously been confined to the US and UK, as Hurt Locker lawsuits have begun moving up to the Great White North. Three ISPs have already been ordered to disclose the identities of users accused of downloading the film, and if the ISPs decide to comply, there could be plenty more lawsuits.

As noted by University of Ottawa law professor Michael Geist, the producers behind Hurt Locker, Voltage Pictures, filed suit against a number of John and Jane Does in Quebec towards the end of August. The suit accuses the anonymous users of downloading, copying, and distributing the film via P2P networks—the users are only identified by IP address, and the Voltage Pictures maintains that the only way to proceed with the case is to obtain names and addresses from their ISPs.

On August 29, the Federal Court in Montréal issued an order to the three ISPs in question—Bell Canada, Cogeco Cable, and Videotron—giving them two weeks to cough up the identities of the users associated with those IPs. Those two weeks expire next Monday, September 12, and the ISPs have yet to give any sort of signal that they won't comply.

Voltage Pictures first began suing users in May of 2010 in the US for allegedly torrenting Hurt Locker—a case that started out with a massive 5,000 users, but quickly bloomed into a much larger (and precarious) monster of P2P lawsuits all over the country. Ever since then, the Washington D.C. District Court has been bombarded with letters claiming innocence, highlighting some of the major problems with using IP address to identify users.

Among other things, the accused have argued that their routers—identified by a single IP address—are used by numerous guests or simply that they don't even know how to download illegal music or movies. Though the claims are difficult to vet, it's clear that using an IP address in an attempt to pinpoint a single individual is problematic to say the least.

Geist pointed out that the prospect of thousands of P2P file sharing suits making their way to Canada makes him all the more grateful for Canada's Bill C-32. The copyright modernization bill makes the distinction between commercial and noncommercial infringers, and limits statutory damages on noncommercial infringers to a maximum of CAN$5,000 for all infringements. "This case confirms that mass lawsuits with the threat of thousands in liability is a real possibility in Canada and why changes to the law are needed," Geist wrote.

Privacy-violating, useless AVG antivirus app pulled from Windows Phone Marketpla

Published a day agoLast updated a day ago - by Peter Bright | Posted in: One Microsoft Way

An anti-virus application surprisingly appeared on Windows Phone Marketplace earlier this week. It's now been pulled, amidst claims that it violates user privacy and doesn't actually do anything useful.

Last updated a day ago

To the surprise of many, an antivirus application was published on the Windows Phone Marketplace earlier in the week. The publication of AVG Mobilation for Windows Phone was peculiar for two main reasons. The first is that Windows Phone simply doesn't have any viruses to scan for. Second, Windows Phone applications are sandboxed; they have no access to the system files or other applications. Even if a virus were to be developed for the platform, the virus scanner would not be able to detect or remove it.

AVG was apparently undaunted by these obstacles, and developed the free, but ad-supported, Mobilation regardless. Though Windows Phone gives applications no ability to access most files on the system, there are some exceptions. Third-party software can access photos and music stored on the device, and so, for lack of anything better to scan, this is what Mobilation examines.

Or at least, what it pretends to examine. Analysis by Rafael Rivera suggested that it was not even particularly thorough in the scanning it did perform: it checked the names of the files it scanned and if they matched the string "eicar" or "עברית" it flagged them as "suspicious". "eicar" is a reference to the EICAR file used to test anti-virus software. "עברית" is Hebrew written in Hebrew. Its significance is not obvious.

Being merely useless is not, however, against the terms of the Windows Phone Marketplace, and even with its restrictions, there is the possibility that the platform does attract viruses. Some of these might even use music and image files to propagate. AVG putting the infrastructure in place to scan these files is not entirely unreasonable.

However, further investigation showed that the software was not merely useless. Ex-Microsoft employee Justin Angel decompiled the program and found that it collects a range of information—including the phone's unique ID, the network operator, the owner's e-mail address—and then sends this information, along with the phone's GPS location, to AVG's servers.

The relevance of this information to a virus-scanner appears to be negligible.

The evidence of shady behavior led to Microsoft pulling the application late last night, so that it could investigate fully.

AVG in turn made a lengthy blog post in support of its product. However, the post is remarkably quiet about any of the allegations made of Mobilation. It talks about another security product for Windows Phone that AVG released at the same time, a "Safe Search" application that verifies the safety of URLs to block access to malicious websites—but this is irrelevant to the complaints being made.

The post also talks about the malware threat on Android and attempts to use this to justify the development of the Windows Phone product; it describes it as a "step in the right direction." That there are marked differences between Windows Phone's curated Marketplace and Android's free-for-all Market has been ignored.

Finally, the post concludes with a statement that AVG takes privacy very seriously, that it won't sell the data it collects, and that nobody should have anything to worry about. It avoids explaining why it needs the data in the first place.

Repeated requests to the company for further information have gone unanswered.

Update: AVG says that the data collection is for the phone tracking feature that their software provides. It's enabled by default, they say, because they don't want users to be in the position of having lost their phone without enabling the feature. Quite how it works on a platform with no multitasking is unclear, as is the necessity of a second location tracking feature: Windows Phone itself already includes this functionality.

As well as talking about its own software, AVG also took great pains to claim that the software was produced with the full knowledge and involvement of Microsoft, and that it made alterations at Microsoft's request. Exactly what this means is unclear; it could mean simply that the application was initially rejected from Marketplace and that AVG had to make changes in order to have it approved for publication at all, or it could mean that Microsoft was actively involved in the product's development. We are awaiting comment from Microsoft to clarify the situation.

Update: Microsoft tells us only that "AVG's app has been removed from Windows Phone Marketplace while we work with AVG to ensure that the app is in full compliance with our published policies."

If the product was indeed written with Microsoft's involvement, that raises further questions. The presence of a virus-scanner, even if useless, does Windows Phone no favors. The image of desktop Windows is seriously tarnished by the malware specter, and allowing a (largely bogus) anti-virus program onto the phone platform serves only to taint that, too.

P2P lawyer fined $10,000 for "staggering chutzpah"

Published about 19 hours agoLast updated about 15 hours ago - by Nate Anderson | Posted in: Law & Disorder

Texas lawyer Evan Stone hoped to cash in by suing accused file-swappers of downloading porn. Instead, he owes $10,000 after sending subpoenas and settlement letters without the judge's permission.

In September 2010, Stone brought suit on behalf of Mick Haig Productions against 670 accused file-swappers, and he asked permission to take early discovery. Judge David Godbey said no; instead, Godbey brought in the Electronic Frontier Foundation and Public Citizen to represent the interests of the Does, since none of them had yet been named and therefore had no counsel to speak for them. EFF and Public Citizen lawyers soon began hearing from people who said that Verizon had turned over their information to Stone, information generally obtainable only by subpoena.

The lawyers asked Judge Godbey to find out what was going on, and to sanction Stone if he had in fact issued subpoenas without the court's permission. Turns out that he had—at least four times. Godbey ruled (PDF) yesterday that Stone "grossly abused his subpoena power," obtained subscriber names he was not entitled to learn, and then, "almost unbelievably, Stone used the information he received to contact an unknown number of potential Does, presumably in the form of demand letters and settlement offers."

This wasn't even the first time Stone had run into subpoena problems. In a separate Texas lawsuit over anime, Stone sent a subpoena more than a month after the judge in that case withdrew permission to do so; even more shockingly, "Stone issued the subpoena on the same day that he voluntarily dismissed the underlying case," according to Godbey.

The judge's anger at Stone burns clearly throughout his order, especially when he sums up the entire situation:

To summarize the staggering chutzpah involved in this case: Stone asked the Court to authorize sending subpoenas to the ISPs. The Court said “not yet.” Stone sent the subpoenas anyway. The Court appointed the Ad Litems [EFF and Public Citizen] to argue whether Stone could send the subpoenas. Stone argued that the Court should allow him to - even though he had already done so - and eventually dismissed the case ostensibly because the Court was taking too long to make a decision. All the while, Stone was receiving identifying information and communicating with some Does, likely about settlement. The Court rarely has encountered a more textbook example of conduct deserving of sanctions.

Judge Godbey fined Stone $10,000, which he hopes will be enough to "deter similar misconduct and adequately reflects the gravity of the circumstances." In addition, Stone has to pay the attorney fees for EFF and Public Citizen, and must tell the court if he has settled with anyone in the Mick Haig case, and if so, for how much.

Finally, Stone has to send a copy of Judge Godbey's order to each judge overseeing any federal or state case in which Stone currently has a hand—whether or not it has to do with file-sharing.

Those are serious penalties, but as Judge Godbey noted, "The adage 'it is easier to ask forgiveness than it is to get permission' has no place in the issuance of subpoenas."

Stone has not yet responded to our request for comment.

Update: Stone tells me by e-mail, "After three rewrites, I finally decided I'm just going to have to let Justin Bieber do my quoting for me: 'Whenever you knock me down I will not stay on the ground.'"

Judge calls $1.5M file-sharing judgment "appalling," slashes to $54,00

by Nate Anderson | Published 2 days agoLast updated about 24 hours ago

Jammie Thomas-Rasset, the first US resident to have the file-sharing lawsuits against her go all the way to trial and verdict back in 2007, “lied in her trial testimony," said federal judge Michael Davis today. And her “past refusal to accept responsibility for her actions raises the need for strong deterrence."

But that deterrence won't come courtesy of a jury, which last year found Thomas-Rasset liable for $1.5 million dollars—$62,500 for each song she was accused of sharing on the KaZaA peer-to-peer network. That case was her third time through a trial; the first two trials had ended with Thomas-Rasset on the hook for $222,000 and $1.92 million, respectively. In each case, Judge Thomas has altered or set aside the jury's verdict, and he did so again this morning.

Davis has found that the $1.5 million award was unconstitutional; he slashed it to $54,000—$2,250 per song. And even then, the amount was "a higher award than the Court might have chosen to impose in its sole discretion.”

Davis has done this before. After the second trial and its $1.92 million verdict, he also reduced the award to $54,000, but he elected not to do so on constitutional grounds. That decision led to Thomas-Rasset's third trial, but today's constitutionally based decision should end the matter at the District Court level. If this case is pursued—and we suspect it will be—the trial phase will end and a federal Appeals Court will take over.

Sections of the verdict are worth quoting in full; they illustrate Judge Davis' deep common sense about the case and provide a worthwhile framework for thinking about similar P2P cases.

The Court concludes that an award of $1.5 million for stealing and distributing 24 songs for personal use is appalling. Such an award is so severe and oppressive as to be wholly disproportioned to the offense and obviously unreasonable. In this particular case, involving a first-time willful, consumer infringer of limited means who committed illegal song file-sharing for her own personal use, an award of $2,250 per song, for a total award of $54,000, is the maximum award consistent with due process.

This reduced award is punitive and substantial. It acts as a potent deterrent. It is a higher award than the Court might have chosen to impose in its sole discretion, but the decision was not for this Court to make. The Court has merely reduced the jury’s award to the maximum amount permitted under our Constitution.

Davis had nothing positive to say about Thomas-Rasset, who he faulted for lying and for denying responsibility by “casting possible blame on her children and ex-boyfriend.” As for her argument that she caused no harm to the music industry, Davis “rejects her suggestion" and calls for a penalty in order to enforce copyright law, compensate the record labels, and “deter future copyright infringement.”

But Davis was at his most Old Testament when blasting the huge damage awards handed down.

There is no doubt that a multimillion dollar penalty is overkill to deter a private individual from obtaining free songs online… Although Thomas-Rasset played a role in the web of online piracy, she played a minuscule role… It cannot be that she must pay the damages caused by millions of individuals because she was one of two users caught, sued, and subjected to a jury trial… [Joel Tenenbaum in Massachusetts was the other.] This award constitutes the maximum amount a jury could award, consistent with the due process clause.

His $2,250 per song figure is three times the minimum statutory damages of $750, though Davis does acknowledge the difficulty of drawing exact lines between fair and unfair judgments. “Any specific dollar amount will appear to be somewhat arbitrary,” he wrote. “Why is an award of $2,251 per song oppressive while an award of $2,250 is not?" The "3x" multiplier seemed to him a fair one, as such multipliers are routinely used to punish willful offenses (the same amount was chosen in the Tenenbaum case when the judge likewise gutted that verdict).

Thomas-Rasset has made clear to me at her trials that she can't and won't pay even a reduced $54,000 amount, and she has apparently rejected music industry settlement offers that would have her pay even less. Where we go from here is therefore anyone's guess—but continued court action looks likely.

(Sidenote for copyright wonks: Judge Davis declined to include in his injunction against Thomas-Rasset a clause that would bar her from "making available" the songs at issue here. "The Copyright Act does not provide a making-available right," he wrote.)

Update: An RIAA spokesperson tells Ars, "We disagree with this decision and are considering our next steps."

Kiwi Camara, a lawyer representing Jammie Thomas-Rasset, called today's decision "an excellent result. I am sure that the recording industry plans to appeal; when they appeal, we will file a cross-appeal also."

As for Camara's one-time plan to file a class action lawsuit against the entire music industry, I asked him if that was still in the works. "As for the class action," he replied, "you know [Harvard Law's] Charlie [Nesson] and you know us—anything could happen!"

Roku adds gaming to streaming media with sub-$100 Roku 2

by Chris Foresman | Published 3 days agoLast updated about 24 hours ago

Roku announced a new line of compact streaming video players on Wednesday, dubbed Roku 2. Aimed squarely at competing with the second-generation Apple TV, the new Roku 2 is smaller, lighter, uses less power, and, unlike the Apple TV, can play Angry Birds.

Apple announced the Apple TV 2 at the beginning of September last year. The Lilliputian device was essentially a fourth-generation iPod touch packed in a tiny case with a power supply. It runs iOS, and offers an optimized 10-foot interface. Because of its iOS-underpinnings, it was speculated that Apple would allow developers to build and sell apps for the device, including popular casual games. However, Apple has yet to offer developers an SDK to build apps for the big (flat) screen.

Roku followed up three weeks later with a line of three new players that cut the size down considerably, though still not as small as the Apple TV. This new line ditched standard definition resolution, offering 720p at the low-end, but essentially offered the same features as the older Roku hardware they replaced.

The Roku 2 is smaller than the already-tiny Apple TV 2.

Roku

But Roku didn't rest on its laurels for very long. The new Roku 2 is 3.3 inches by 3.3 inches by 0.9 inches thick (9.8 cu. in.) versus Apple TV's 3.9 inches by 3.9 inches by 0.9 inches thick (13.7 cu. in.), and weighs just 3 ounces—less than a third of the Apple TV's 9.6 ounces. The tiny device consumes as little as 2 watts when streaming HD content, compared to the Apple TV's 6 watts.

Roku also offers a wider variety of content. Apple TV can stream content from iTunes, Netflix, and YouTube, and can access subscriptions for streaming MLB and NBA games. Roku has, among its 300 channels, Netflix, Amazon Video on Demand, Hulu Plus, Crackle, YouTube, and more. It also streams music from Pandora, Last.fm, Rdio, and others, and you can buy subscriptions for NBA, MLB, NHL, and UFC sports coverage.

The new Roku "enhanced remote" with motion sensor.

Roku

Topping that, Roku 2 is now a gaming platform. All three updated devices now include Bluetooth for connecting to a motion sensor-equipped, $30 enhanced remote for a Wii-esque "casual gaming" experience. Roku worked directly with Rovio to bring on Angry Birds as a launch title, and that's currently the only game being highlighted. However, Roku said that more games are coming soon.

Roku 2, like its predecessors, comes in three different models. All three are equipped with 802.11n WiFi, HDMI and composite output, a microSD slot for storing game and channel data, and Bluetooth 3.0 for connecting to the game controller remote. The HD costs $59.99 and is limited to 720p resolution. The XD costs $79.99 and boosts resolution to a full 1080p. The XS, for $99.99, adds a wired Ethernet port, USB port to playback video, music, and photos from a storage device, a game controller remote, and a free copy of Angry Birds.

The new family of Roku 2 devices. From left to right: The $60 HD, $80 XD, and deluxe $100 XS.

Roku

Users of the lower-end models can add an enhanced gaming remote, bundled with a 2GB microSD card, for $29.99. It looks as though a single Roku device may be able support up to two game controllers simultaneously for head-to-head gaming (though Roku didn't respond to our request for more information). Either way, the $100 Roku 2 XS seems like the better value if you're interested in playing some games on a second TV or don't already have a Wii or other console.

Samsung-funded study finds 3D video causes extra eye strain, fatigue

by Casey Johnston | Published 2 days ago

3D displays cause extra eye fatigue, according to a study published by the Journal of Vision today that was funded in part by Samsung's R&D arm. A group of researchers from the University of Califonia-Berkeley found that when test subjects watched 3D displays, they reported more eye strain and fatigue and less vision clarity afterward than when they watched 2D video. The disparity between the depth of the screen and the depth of the 3D image caused the most problems, though researchers also found the relationship between image depth and nearness of the screen also played a role in eye strain.

Twenty-four participants in the study were shown 3D and 2D video at various viewing distances, and then responded to questionnaires on their eye fatigue, neck and back pain, and vision clarity. In the video clips, the authors were varying the focal point, which is the surface of the screen, and the vergence distance, which is where in the image the eyes are trying to focus on. For 2D video, these points are always one and the same, but in 3D video the vergence distance varies, and can be either deeper than the surface of the screen or in front of it.

The participants responded that they experienced more eye strain and fatigue from the video with different vergence and focal distances, a feature of 3D that has long been supposed to cause eye strain. The self-reported differences between 3D and 2D were not drastic, but they were significant.

A second part of the study found that though 3D was fatiguing in general, the participants had more problems with distant displays showing an image with a vergence distance deeper than the screen and with near displays showing images popping out of the screen.

This research is highly relevant to 3D content designers, who could determine what to recess or pop out of the screen based on the expected viewer distance. However, the study also means that 3D video that is more comfortably viewed in a movie theater is necessarily much more uncomfortable to look at when viewed in a living room.

Unfortunately for Samsung, financial supporters of the study and manufacturers of 3D TVs, the research appears to indicate that they're hurting the eyes of their customers.

Will VMware's new licensing scheme open the door for Microsoft?

by Peter Bright | Published 10 days agoLast updated 10 days ago

VMware announced vSphere 5 yesterday, which will bring greater scalability and robustness to VMware's virtualization platform. The new version will support larger virtual machines—up to 1TB of RAM and 32 virtual processors each—faster I/O, simpler high-availability, easier deployment, and more. These announcements were somewhat overshadowed, however, by the launch of a new licensing scheme for the software.

For vSphere 4.x, the current version, pricing is based on a combination of the number of physical CPU sockets, physical cores, and physical memory installed in a server. Leaving aside the "Essentials" versions, as they operate on a different pricing model, there are four tiers: Standard, which gives you one socket, six cores, and 256GB memory; Advanced, which is 1 socket, 12 cores, 256GB memory; Enterprise, which is 1 socket, 6 cores, 256GB memory, and extra functionality; and Enterprise Plus, which is 1 socket, 12 cores, unlimited memory, and even more functionality. Additional sockets, cores, and memory required purchase of additional licenses.

Neutrino transformation could help explain mystery of matter

by Dave Mosher, wired.com | Published about 6 hours ago

Two research teams have found new evidence of transformations in elusive elementary particles called neutrinos. The findings may finally help explain why the universe didn’t vanish shortly after its birth.

“These results are just the beginning of the story for neutrinos,” said physicist Robert Plunkett of Fermilab in Chicago. “They could be a clue, a leftover from the early universe, and tell us why there’s now far more matter than antimatter.”

Most neutrinos are emitted by the sun, and are so small and ghostly that billions pass through our bodies every second. Most go right through Earth without hitting anything. But some human-built devices—big chambers of oil or water lined with photon detectors, or detector arrays plunged into seawater or Antarctic ice—can record the blip of light when a neutrino occasionally slams into an atom.

Using these detection events, physicists have identified three types of neutrino, called muon, tau and electron neutrinos. Further discoveries suggested that each type can transform into another, with muon-to-tau and tau-to-electron neutrino transformations being dominant. Researchers proposed a third and weaker change, that of muon-to-electron neutrinos, but until now lacked evidence for its existence.

On June 14, the Japanese Tokai-to-Kamioka experiment reported the significant detection of muon-to-electron neutrino changes. On June 24, the Main Injector Neutrino Oscillation Search (MINOS) experiment at Fermilab reported the same. While the ranges of their data varied, the basic claims were the same.

“[The values] differ because we used different techniques and distances, but they overlap at one part. They’re complementary,” said Plunkett, a co-spokesperson of MINOS.

With a more complete understanding of neutrino transformation in hand, Plunkett said physicists can now design experiments to investigate larger questions about the universe. The largest among them: why there’s far more matter than antimatter.

Matter and antimatter particles annihilate when they meet. Each type is thought to have appeared in equal proportions shortly after the Big Bang, yet the matter-rich universe as we know it still exists. As a result, physicists are seeking evidence of “asymmetries,” in which matter-antimatter encounters end up emitting more matter particles.

Some matter-favoring asymmetry shows up in the annihilation of quarks, though the effect is relatively meager. But physicists say a muon-to-electron neutrino transformation supports the possibility of more significant asymmetries.

“We now have a good enough handle on neutrinos to design experiments and try to address such a big mystery,” Plunkett said.

Google TV developer hardware has Honeycomb-based software

by Ryan Paul | Published 3 days ago

The Google TV platform was welcomed by a chorus of chirping crickets and not much else when it launched in 2010. The lackluster launch hardware and limited software capabilities made Google's Internet-enabled set-top solution a no-go in the living room. Despite the rough start, Google has continued to push the platform forward.

Some details have materialized about a very early version of the next-generation Google TV software platform, which was reportedly shipped out to a select group of developers on a reference hardware prototype, codenamed Fishtank. Some hardware photos and software screenshots were published by gadget blog geek.com. The software interface has the same "holographic" look and feel as Honeycomb tablets.

As Google promised last month at the Google I/O developer conference, the new version of the Google TV software is built on Android 3.1. The search giant also said that the Android Market would be opened up to Google TV devices, allowing end users to install and run third-party software. There is already some evidence that this work is underway—the AndroidCentral blog reported this week that Google TV devices have started showing up in the device listings on the Android Market website, albeit without support for remote installation yet.

The potential to run third-party Android applications on Google TV opens up some new possibilities and could help make the set-top platform more compelling. Imagine, for example, being able to run Plex's Android client on a Google TV device. It would also open the door for turning the Google TV into a budget gaming console—pairing low-cost Android games with the newly added support for game controller input devices in Android 3.1.

Intel unveils 50-core supercomputing coprocessor plans

by Jon Stokes | Published 7 days ago

Intel's Larrabee GPU will finally go into commercial production next year, but not as a graphics processor. Instead, the part will make its debut in a 50-core incarnation fabbed on Intel's 22nm and aimed squarely at one of the fastest growing and most important parts of NVIDIA's business: math coprocessors for high-performance computing (HPC).

When Intel's ambitious, hybrid software/hardware GPU effort failed in late 2009 due to delays, Intel insisted that the silicon side of the project would live on in some form. The next year, the company announced that Larrabee had morphed into the Knight's family of HPC coprocessors, which the company began shipping in very limited quantities as a research testbed. Intel also began calling the basic architecture of the Knight's family is Many Integrated Core (MIC) architecture.

Today's announcement is the official unveiling of Intel's broader plan to commercialize the MIC-based Knight's family, starting with the 50-core Knight's Corner chip on 22nm. Intel is also announcing partnerships with SGI and other system integrators that plan to build commercial HPC systems around the MIC silicon.

The MIC products will compete directly with NVIDIA's Tesla line, making MIC a threat to NVIDIA's growth prospects in a world where integrated processor graphics (IPGs) like Sandy Bridge and AMD's Llano are eating the discrete GPU market from the bottom up.

The main advantage that Intel touts vs. Tesla is that because MIC is just a bunch of x86 cores, it's easy for users to port their existing toolchains to it. (When using Tesla, researchers must port to NVIDIA's proprietary but well-loved CUDA platform.)

In a previous discussion of the MIC vs. Tesla issue, I suggested that Intel was massively overselling this ease of porting, since applications must be redesigned anyway. But having learned a bit more in the intervening year, I'm not quite as certain that this is the case. Ease of development and porting do seem to matter for both budget-constrained academic labs and the kinds of high-frequency trading and other finance applications that prize speed of deployment along with fast computation.

Even if it does turn out that x86 gives Intel a big advantage, it's not all smooth sailing for MIC. One thing that's missing from the press materials that Intel sent was a set of relative performance claims vs. Tesla, and that's probably because the more specialized Tesla would crush it on the kinds of codes for which Tesla is commonly used. It's the age-old general-purpose (slower, easier to use) vs. specialized (faster, harder to use) tradeoff, and Intel is betting that since Tesla has so far been the only real option there are plenty of potential users out there who are in the market for something less specialized.

None of the press materials I saw indicated that Intel is changing its interconnect architecture for Knight's Ferry, so it looks like the chipmaker will be hanging 50 cores off a single, high-performance yet power-hungry ring bus. This is surprising, since I would have expected the company to move to a tile architecture (à laSCCC) for a core-count this high. A tile-based design is no doubt in MIC's future, but there are certainly issues that remain to be worked out.