24 April, 2013


The previous post was about AMDs current bugdet-line mobile processors. Before that, I posted (in Norwegian) about their plans for future mobile solutions.

They are basing their new budget-line / mobile processors on the "Jaguar" microarchitecture (the pre-decessing C- and E-series are "Brazos" based), and will sport 64-bit dual- or quad-core ("Bobcat" x86-cores) with integrated Radeon 8000 graphics (80-core GPU), significantly outperforming any "Atom"-chips with integrated graphics from Intel.

The new "AMD Embedded G-series System-on-Chip" sets will be made in two versions (or architectures): x86 and ARMDetermining which architecture a chip is based on, is as easy as looking at the CPU-logo; the ARM-version will have an "A" in the lower right corner of the logo and the x86-version will have an "X" (as shown in the picture below).

AMDs "Embedded G-series SoC" seems to be direct competition against Intels "Atom" budget-line. CPU-vendor competitiveness is, as we all know (at least the one's paying attention), a good thing :-P

It will be exciting to see how the combination ARM+Radeon will operate and behave performance-wise :-) since I've been an avid embedded (ARM ) fan for years, it'll also be quite interesting to see AMD play with both CISC- and RISC-based designs in the same series.

AMD press-release:

Ultrabook for my sister

I was recently asked by a family-member to look for a decent study-laptop for my sister. I'll admit I was kind of thrilled. I was playing around with the idea of finding a low-budget laptop, with decent specifications. Having done a machine-architecture report for school (x86/x64) last year, I had become thoroughly aware of what AMD has been doing lately.

I have always been a fan of AMD. Mainly because of their support for hardware-hacking (NOTE: this voids any warranty) and / or chip overclocking (in many cases also voids any warranty), and their clever innovation of techniques (amd64/APUs). And last but not least, their open and full support of FOSS.

In my report, I wrote about AMDs budget-line processors; the "FusionC-series and E-series, direct competitors to Intels "Atom" budget-line processors. In fact, the C-chips match similar Atom-chips in both TDP (Thermal Design Power) and clock-frequency (GHz). Whereas the E-series is higher clocked, resulting in a higher TDP, but also making it more similar to an Atom-chip system paired with a dedicated GPU (e.g. ION2).

My sisters computing-needs do not require massive number-crunching abilities, but it shouldn't be sluggish in operation either, so I went for the high-end mobile-solution: AMD E-450 (codename: Zacate) APU (Accelerated Processing Unit) used in a 13,3" LED-screen Asus model U32U laptop with 4GB DDR3 RAM.

The AMD E-450 is a 64-bit dual-core ("Bobcat" low-power x86-cores @ 1,65GHz) processor with an integrated Radeon HD 6320 GPU on the same die (which AMD markets as a so-called APU-chip), with a total TDP of 18 watts.

(The basic idea behind this type of system, is that: the stripped, low-power x86-cores (x2) does most of the general-purpose processing, but, hands over floating-point unit calulations to the GPU-cores on the same die, thereby eliminating both process load-balancing and CPU-GPU inter-communication delays.)

Asus product-link:

When unboxing the machine, I was pleasantly surprised :-) with the battery firmly locked-in the thing didn't weigh more than 1,55 kg! Didn't take up much space when closed either so it was a perfect carry-on companion and study-tool.

First boot took a few minutes, Windows always does at "first-boot".

Running Windows Update took around 4,5 hours to complete from a fresh install, after downloading everything that is (on a 2Mbps cable-DSL connection no less :-P *shrug*).

Installing Microsoft Office 2010 (with the help of an external USB 2.0 CD-ROM drive) actually didn't take all that long, and it ran like a dream.
Overall the CPU-response was acceptable when installing and pretty good in operation. It won't blow away anything with 4GB of RAM, but in combination with AMDs Fusion (referred to as a: Heterogeneous System Architecture by AMD) chipset and APU-chip, it didn't do half bad for an ultrabook-like laptop.

This was my first ultrabook-like experience, and I must say, I am impressed :-) I want one! Strictly speaking, the U32U does not fit the ultrabook-specification (being an Intel-trademark, and the U32U being AMD-based).

Personally (speaking as a poweruser), I would probably cram at least 8GB of RAM into the machine afterwards. And replace the 320GB HDD with an SSD for both system-speed and machine-weight. But for my sister, it was perfect! :-)

I might also add that I actually own two machines with Atom-processors, and the E-450 blew them both away on graphics and processing-power.

I had to add a total of 8GB DDR3 RAM modules to the machine. She complained that it was getting sluggsish after a period of time. I had apparently not anticipated her usage-patterns good enough. 4GB of RAM was a little weak, especially if you want to run office-suites and similar resource-hogging software.

Funny comment :-P

This was a response to "Correct me if I'm wrong" posted by an irrate Linux FUD-er. http://www.zdnet.com/six-open-source-security-myths-debunked-and-eight-real-challenges-to-consider-7000014225/ )

You're wrong and probably never worked for a software company.

I've been through this a couple times before, but I'll describe it again.

Software versions are charactized by version numbers, in my case, when I worked for Bentley Systems, a version was designated by numbers like

Software, like Firefox makes versions for Windows and Linux among others. Basically the core parts of the program operate in the same manner and the API is different to work on Linux and adjustments are made for libraries, directories, etc.

The dirty little secret here at ZDNet and among the shills is that they blame an application for allowing an intrinsic problem or vulnerability with the OS to be accessed. Shills, Ed, and ZDNet are great at blaming the application, such as Chrome or Firefox for the problem and not addressing the core Windows vulnerability. Then, they read documentation and without knowing anything about how things are done, blame the Linux counterparts, because they are listed.

The problem is that items present in the application allow the core Windows vulnerability to be used to infect Windows. The application issue may also be present in the Linux version, but because Linux is so much more secure than Windows, there is no problem or infection with Linux. The only way Linux could be infected is if the malware could read the mind of the user and get his password.

Developers review the Windows version issue and make adjustments so it does not allow the Windows vulnerability to be addressed and also make the change across the board to all sister versions to maintain consistency. Because you are naive and see Ubuntu listed as affected, it does not mean Ubuntu ever had a security issue at all, the Ubuntu version is just having the code changed for consistency. In other words, no application for Windows is ever going to fully prevent all the Windows critical flaws from being accessed. Those application characteristics causing the Windows issues may be present the Linux version, but can't be used to attack Linux, but are being changed anyway. In most cases, the change may be an operating improvement and be more efficient.

It's so silly to ZDNet pull the same BS over and over again, year after year. If you want to believe it, you are only following the ZDNet propaganda trail, Do yourself a favor, pour yourself a strong one, and install Ubuntu or Mint on a second machine, run it as a Live DVD, or install it as a dual boot and your primary computer. Then, install, Chrome, Opera or any other open source program you like and try to get infected. Then come back here and post the Website and how you got infected. That's something that no one, in all these years of accusations has ever been able to do. Once you see that you don't get infected you;ll begin to see how ZDNet twists information and is just a stooge for Microsoft.

As far as you referencing Linux Torvalds and the linux.com issue it was related to stolen passwords. Anyone who gets poorly secured passwords an attacks a system can't be stopped. Most times the admins are storing their login information on a Windows box, that gets easily hacked by a zero-day or a crafted emai that allows access. Remember the big ZDNet push for articles about Google, which runs 100% Linux getting hacked? Well, two Chinese employees were storing data on a Windows notebook and it easily got hacked. Since then, Google forbids employees from using Windows or work. you don't hear about that anymore here, do you? Forbidding employees to do company work on Windows is the single most important any manager can make.

If you dig deeply into these articles against open source and Linux, you will find, as I have, that the core problem is Windows and you will see a critical update down the road, at a later time to silently correct the Windows problem. But that is never brought up here.
  • Prove that Linux gets hacked. It's something never done here.

    If you feel that strongly about it, post how and where you got hacked. I'm waiting.

    Test the theory yourself

    Get a box and harden Windows, use ANY anti-malware you like.
    Get a box and install any Linux distro, any of them, pick the weakest one you can think of.

    Go look at any web site you like with Firefox, Opera or Chrome on both ... NO IE (not supported on Linux).

    (hint: adult content and gaming sites that are 2nd or 3rd tier are reportedly famous for infections), try google searching "most dangerous web sites".

    The rules of the game are:

    * only following links or using the back button icon of the browser are allowed
    * if windows pop up you are not allowed to touch them anywhere (including the X to close).
    * if the back button is not usable or the browser is non-responsive, close the browser with task manager.

    The object is to visit infected sites and return without touching anything.

    See which system is left running after 1 hour.

    Please report your results HONESTLY.

    * ( ... no clicking (X)
    * If the browser locks up ... use the "task manager" to kill it.


23 April, 2013

AMD, SoC og ARM.

AMDs nye satsing på bærbare enheter heter "AMD Embedded G-series System-on-Chip" (kort: SoC), basert på AMDs nye Jaguar mikroarkitektur.

AMDs pressemelding:

Det er nok den nye strømgjerrige serien som skal konkurrere direkte med Intels "Atom"-serie (benyttet i NAS-bokser, nettbrett, mini-laptoper, osv.).

Brikkene skal lages i to arkitekturer: ARM og x86. Om brikken er ARM- eller x86-basert ses nede i høyre hjørne av logoen hvor det blir trykt en "A" for ARM, og en "X" for x86 (bildet ovenfor viser en x86-basert brikke). De vil også komme med to eller fire kjerner som vil kjøre på en klokkefrekvens mellom 1 til 2 GHz. Grafikkmessig stiller AMD seg bedre enn Intel. Disse brikkene har integrert GPU basert på Radeon 8000 som i følge AMD gir opptil 5 ganger bedre grafikkytelse enn tilsvarende "Atom"-prosessorer fra Intel.

Personlig liker jeg veldig at AMD leker med å blande CISC / RISC i en brikke-serie, og det vil bli voldsomt spennende å se hvordan brikke-kombinasjonen ARM / Radeon vil oppføre seg i bruk.


16 April, 2013


Blir nok heller satsning på Vishera mikro-arkitekturen hvis jeg skal oppgradere kjerne-systemet :-) FX-8350s etterkommer :-) ekte "octa-core" 5GHz ulåst FX-prosessor :-p