Μια ματιά στον επερχόμενο Sandy Bridge της Intel

[nem52]

Πιο 1156;

Για τους Sandy Bridge μιλάμε.

Σορρυ εχεις δικιο.. απλα ειδα στην το 115χ και πιστευψα οτι ο φιλος παραπανω αναφεροτανε σε αυτους..

[grimpr]

Η Intel παρουσιάζει την laptop πλατφόρμα επόμενής γενιάς Sandy Bridge, στην Computex 2010.

 

http://www.engadget.com/2010/06/01/intel-briefly-demos-next-gen-sandy-bridge-laptop-platform-at-com/

 

Σχετικό άρθρο.

 

8.2 μονάδες επάνω η AMD στις κάρτες γραφικών.

http://www.insomnia.gr/forum/showthread.php?t=380049

 

Η AMD αποκάλυψε τον Llano.

http://www.insomnia.gr/forum/showthread.php?t=360748

 

 

[nikos21aek]

 

Ένα καλύτερο βιντεάκι με την παρουσίαση.

[grimpr]

8 Ιουνίου 2010.

 

[ame]http://www.youtube.com/watch?v=TkP4rEV_MTE[/ame]

 

2 Ιουλίου, 2010.

Η Intel Math Kernel Library 10.3 με υποστήριξη AVX ήδη σε beta κυκλοφορία.

 

http://software.intel.com/en-us/articles/intel-math-kernel-library-103-beta/

 

29 Ιουνίου 2010.

Η νέα έκδοση της Intel Kernel Math Library, συνεχίζει την παράδοση....

 

New version is still crippling Intel's competitors

 

Intel have released a new version of their Math Kernel Library (v. 10.3) in beta test.

 

I have tested the new libraries and found that the CPU dispatching works basically the same way as before. The standard math library, vector math library, short vector math library and the 64-bit version of other math kernel library functions still use an inferior code path for non-Intel processors.

 

I have found the following differences from previous versions:

 

* Many functions now have a branch for the forthcoming AVX instruction set, but still only for Intel processors. This will increase the difference in performance between Intel and AMD processors on these functions. Both Intel and AMD are planning to support AVX in 2011.

 

* The CPU dispatcher for the vector math library has a new branch for non-Intel processors with SSE2. Unlike the generic branch, the new non-Intel SSE2 branch is used only on non-Intel processors, and it is inferior in many cases to the branch used by Intel processors with the same instruction set. The non-Intel SSE2 branch is implemented in the 32-bit Windows version and the 32-bit Linux version, but not in the 64-bit versions of the library.

 

* A new Summary Statistics library uses the same CPU dispatcher as the vector math library.

 

Obviously, I haven't tested all functions in the library. There may be more differences that I haven't discovered. But it is clear that many functions in the new version of the library still cripples performance on non-Intel processors. I don't understand how they can do this without violating the legal settlement with AMD.

 

http://www.agner.org/optimize/blog/read.php?i=49#107

 

22 Ιουνίου, 2010

Intel, FTC in talks to settle antitrust case.

 

 

Intel Corp. and the Federal Trade Commission are in talks to settle an antitrust case against the chip maker, a move that could make it more difficult for rivals to pursue damages.

 

In December, the FTC filed charges against Intel, seeking to end what it described as decades of illegal sales tactics that have hampered competitors and kept prices for computer chips artificially high.

 

This week, the FTC and Intel agreed to suspend administrative trial proceedings as they work on hashing out a settlement.

 

The FTC accused Intel of strong-arming computer makers into exclusive deals, manipulating technical data to make its chips look more powerful than those from competitors and blocking rivals from making its chips work with Intel's.

 

Intel, which disclosed the settlement talks late Monday, has disputed the charges and said it has merely been offering discounts.

 

A settlement would be at least a partial victory for Intel, said Robert Lande, director of the American Antitrust Institute at the University of Baltimore. If Intel loses in court, rival chip makers such as Nvidia Corp. would be able to pursue damages without having to again prove the antitrust charges. By contrast, settlements often come without any admission of wrongdoing.

"Once you've lost the case in federal court you can't deny the charges anymore," Lande said. "The only question is, 'Did you hurt our clients and by how much.'"

 

Although a settlement with the FTC would ease Intel's legal troubles in the U.S., the company still faces trouble abroad. A harsh settlement against Intel could bolster the case for stronger measures in European and Asian courts, where Intel is fighting penalties for anticompetitive practices. Lighter terms could discourage plaintiffs outside the U.S, Lande said.

 

A key question remains about whether the cases will affect computer prices. Intel says its sales strategies help keep chip prices low. The FTC argues that prices haven't fallen as much as they could have.

 

Roger Kay, president of Endpoint Technologies Associates, said Intel's practices could harm consumers by destroying competitors that have innovative technologies.

 

"Intel controls so many levers in the ecosphere that they can crush their competitors if they choose to," Kay said.

 

Intel, which declined further comment, is the world's No. 1 maker of microprocessors, the "brains" of personal computers.

 

The FTC case is particularly important because the FTC has said it wants to change Intel's behavior, instead of merely issuing fines, as the European Union and South Korea have done. Last year, Intel paid a record $1.45 billion fine to European regulators but the company is appealing.

 

Intel has said the FTC was seeking major changes to the way Intel sells chips.

 

The two sides had been deep in settlement talks before the FTC filed charges. But those talks broke down over what Intel's general counsel, Doug Melamed, described at the time as "unprecedented remedies ... that would make it impossible for Intel to conduct business," such as restrictions on how much Intel can charge for its chips.

 

Intel's prices have been at the center of complaints by Advanced Micro Devices Inc., a rival whose lobbying of regulators led to charges against Intel around the world.

 

AMD has complained that Intel has paid computer makers to avoid using AMD chips.

 

AMD quoted a manager from Toshiba Corp. comparing Intel's payments for not using AMD's chips to "cocaine." In a 2005 lawsuit against Intel, AMD also quoted an executive from Gateway complaining that Intel's threats of retaliation for working with AMD beat them "into guacamole."

 

Intel says its payments to computer makers are simple volume discounts that help them pass on lower prices to consumers. Microprocessors can be 20 percent of a computer's price.

 

AMD settled the lawsuit with Intel last year, with Intel agreeing to pay AMD $1.25 billion and the companies entering into a new, five-year cross-licensing deal.

 

But the FTC later filed its own complaint.

 

The two sides have 30 days to reach an agreement. If not, the case will continue and go to trial before an administrative law judge in September.

 

The New York Attorney General's office is also investigating Intel.

 

The FTC's lawsuit contains the most wide-ranging allegations yet against Intel, the world's largest chip maker. If the FTC wins, the case will have a broad impact because it affects two key markets dominated by Intel — central processing units used in personal computers and graphics chips.

 

Shares of Intel, based in Santa Clara, Calif., fell 21 cents, or 1 percent, to close Tuesday at $20.98.

 

http://articles.sfgate.com/2010-06-22/business/21921775_1_intel-controls-rival-chip-makers-ftc

 

 

[ahristos]

Βαρέθηκα και την AMD πλέον να μην κάνει τίποτα. Δεν έχεις τους πόρους να γράψεις τους δικούς σου compilers, ζήτα ευγενικά από την Intel να τροποποιήσεις τους δικούς της. Αν αρνηθεί θα ξέρουμε.

Η Intel γράφει τον κώδικα ώστε να ευνοεί τους επεξεργαστές της, η nVidia ώστε να μην χρησιμοποιεί όλες τις δυνατότητες των χ86 επεξεργαστών γενικά, το μάθαμε το παραμύθι.

[grimpr]

Η Intel σκοπεύει να περιορίσει το Overclocking στους Desktop Socket 1155 Sandy Bridge.

Θες να κλοκάρεις στην αποτυχημένη πλατφόρμα μου; σκάστα για ξεκλείδωτους Sandys (300 ευρώ)...

 

When you're overclocking, you want to be able to push certain frequencies, such as the Base Clock and memory clock, but leave others, such as SATA, completely stable as they're very sensitive to adjustment. Current motherboards allow multiple bus speeds because external clock generators are programmable via the BIOS.

 

According to one Taiwanese motherboard company, on a Sandy Bridge system, the fact that all the busses are linked means that turning up the Base Clock by just 5MHz caused the USB to fail and SATA bus to corrupt.

 

 

We chatted about possible work-arounds but at the moment the few 'asynchronous' setups tried were currently not working. It's been claimed to use out-of-the-box the design was deliberately limited with the intention to simplify board design and lower costs. This obviously has the 'unfortunate' side effect that enthusiasts will be unable to manually overclock Sandy Bridge CPUs to their limits, but the CPU's own internal overclocking, TurboBoost, will still work and Intel will offer some controlled multiplier overhead for enthusiasts as a token gesture.

 

http://www.bit-tech.net/news/hardware/2010/07/22/intel-to-limit-sandy-bridge-overclocking/1

 

 

 

[grimpr]

 

 

 

[Thresh]

αυτοί (στον 1ο πίνακα) είναι για socket 1155 έτσι?

[grimpr]

Socket 1155 ναι. Από ότι βλέπεις, τα 32nm της Intel για το 2011 απλά τα σπάνε, ειδικά στην σειρά S.

[qsis]

το οτι ριξανε τη cache στους i5 θα επηρεασει την αποδοση? πχ σε σχεση με τον δικο μου που τον εχω βαλει να δουλευει στα 3.8?

 

ανεβαζουν παρα πολυ με turbo παντως, ακομα και οι φορητοι

[grimpr]

Δεν νομίζω ότι υπάρχει σοβαρός λόγος για να ανησυχείς και εσύ αλλά και πολλοί άλλοι με oced συστήματα γύρω στα 4ghz. Στην θέση σου θα περίμενα και για τις υπόλοιπες πλατφόρμες από Intel αλλά και AMD μέχρι το Q3 2011.

[qsis]

ε καλα εννοειται , αυτη τη δουλεια θα κανουμε? τουλαχιστον για 3 χρονια θα κρατησω το μηχανημα μου

και οταν το αλλαξω θα φροντισω ο αντικαταστατης του να ειναι δραματικα πιο γρηγορος , απλα απο εγκυκλοπεδικο ενδιαφερον αναρωτηθηκα

 

o αντιστοιχος i5 παντως που ενδεχομενως αγοραζα στο μελλον (αν δεν ειχα αγορασει ηδη) ισως με εκανε να ξανασκεφτω το κλοκαρισμα και να τον αφηνα στοκ, λογω τετοιου turbo

[grimpr]

o αντιστοιχος i5 παντως που ενδεχομενως αγοραζα στο μελλον (αν δεν ειχα αγορασει ηδη) ισως με εκανε να ξανασκεφτω το κλοκαρισμα και να τον αφηνα στοκ, λογω τετοιου turbo

 

Χεχεχε, για αυτό ακριβώς τον λόγο το Overclocking θα γίνετε όλο και ποιο ακριβό hobby, η Intel προφανώς το γνωρίζει και σκοπεύει να το εκμεταλλευτεί... Έχουμε περάσει ήδη στην εποχή που εξειδικευμένα υποσυστήματα όπως το PCU (Power Control Unit/System Agent) στους μοντέρνους επεξεργαστές της Intel αποτελούν πλέον την πραγματική καρδιά του κτήνους με πλήρη έλεγχο σε όλα τα υπόλοιπα συστήματά μιας μοντέρνας CPU και με επέκταση έως και σε προγραμματιζόμενα TDP...

 

Ανέμενε παρόμοιες και καλύτερες εκπλήξεις στα σημεία και από AMD μεριά...

[MiDi_MiLiZ]

Δηλαδη ρε παιδια αυτο το 1156 "που δε θα ηταν futureproof γιατι θα σταματαγε η intel να το υποστηριζει" βγηκε καλουτσικο τελικα?

[ahristos]

Δηλαδη ρε παιδια αυτο το 1156 "που δε θα ηταν futureproof γιατι θα σταματαγε η intel να το υποστηριζει" βγηκε καλουτσικο τελικα?

 

Χμ... Μήπως δεν κατάλαβες ότι μιλάμε για 1155

Καλούτσικο βγήκε το 1156 επειδή συνδυάστηκε με το διαμάντι τον I5-750. Intel σε τιμές AMD