On Linux using the pepperflash plugins, lots & lots of zombie processes get created and aren't even killed after you exit the browser. When I noticed 5GB of memory usage on an empty desktop, I realized that Chromium is a pro-zombie browser.

If you bothered to RTFA, you'd see that 500 superhornets are in active service right now. The "Superhornet" isn't really that new and it has issues such as it is still too-short ranged although an improvement over the original F-18, and it suffers from the jack-of-all-trades-master-of-none capability profile. However, it is real, it is proven, and it can likely receive some halfway decent upgrades without costing anywhere near as much as the F-35.

Did I ever say the infrastructure is fine? NO
I said: Obama wanted "stimulus" to "rebuild infrastructure" He got it. Now all of the sudden it's like that stimulus never happened and he has amnesia about it. Could it be that it was completely wasted on things like.. overpriced routers... instead of being spent on the precious "infrastructure" like we were promised? Could it be that West Virginia's government didn't want to use all that taxpayer money it got for something useful? Could it be that handing gobs of cash to unaccountable politicians is just as bad an idea as giving liquor and card keys to teenagers?

I'm holding the government bureacrats and Cisco's feet to the fire on this one BUT... if Cisco hadn't sold "oversized" routers to those government agencies, we'd just be hearing about how Cisco intentionally pawned off "insufficiently specced gear" that would be "obsolete overnight" on the West Virginia government instead... there's no way to win with an inherently bad idea.

Speaking of which, after over $1 Trillion in stimulus to "rebuild infrastructure" (yeah right), the State of the Union address was asking for more money to... rebuild infrastructure. Did it already fall apart again?

Cortex-M operates in a completely different market than the stated market that Intel wants to reach... and BTW that market isn't insanely profitable for even an IP-only company like ARM. The Intel-fanboy response to your ARM-fanboy propaganda would be to say that ARM has no products that are even a bad imitation of Intel's high-end Xeon chips, which is equally as true and equally as irrelevant.

As for the Krait == A15 business, that is complete and utter bunk. Review after review after review have shown these things:
  1. Krait's power efficiency is somewhat better than Medfield's (not miraculously better).
  2. Medfield definitely outperforms Krait.
  3. The A15's are way ahead of Krait in both performance, but also in having a much higher power draw.
    4. Once and for all, just because Qualcomm licensed the ARM instruction set does *not* mean that Krait == A15. Krait == Krait and that is all. It would be just as correct (read: completely wrong) to say that the Atom chips are "Haswell styled" because they have the same core instruction set... no they are not.

Fact 2: Compilers exist.
Fact 3: x86 compilers exist.
Fact 4: NDK for x86 exists.
Fact 5: ARM fanboys who say that ARM will win simply be recompiling Linux while claiming that it is physically impossible to run Linux on an x86 architecture are almost as amusing as they are ignorant.

Myth 1: Intel can't make x86 power/performance competitive with ARM: Being busted as we speak.

Myth 2: ARM can't scale up performance: Beginning to be busted with the A15, more to come with the 64-bit chips.

Myth 3: ARM can just press a button and get Intel level performance without using any extra power: Busted wide open by ARM themselves with the whole "littleBIG/BIG/little/etc" approach and by the conspicuous lack of high-end A15 chips in smartphones (note Tablet != Smartphone, and look at the Cortex-A9 based Tegra4i for the latest example of manufacturers not putting high-clocked A15s in smartphones).

There are 2 different "Tegra4" chips. The Tegra4i is the only one that will ever end up in your smartphone, and it only includes Cortex A9 derived CPU cores and a GPU that, while quite good, includes far few processing units than the big-brother Tegra 4.

So basically: Of course the full-blown Tegra 4 beats smartphone chips. Guess what: So does Haswell, but you won't be seeing Haswell in smartphones either. There's a price to all that performance and Nvidia is targeting the heavy-weight version of Tegra 4 at tablets because the power draw won't fly in a smaller smartphone platform, where the Tegra4i will be Nvidia's offering.

This thing would obviously only be $2.99 + S&H if it weren't for the Microsoft tax! I'm tired of M$ driving up the price of hardware with ... interruption... whispering .... uh... I'm tired of the GOOGLE TAX!

Those are *not* Bulldozer cores! They are more similar to the lower-end Jaguar cores that are going into AMD's tablet & netbook products. They are still a major step up from the actual cores in the Cell (those SPE things are not really "cores"), but even a Bulldozer core will be more powerful than these things on a clock for clock basis.

The good news is the GPU is pretty nice for this type of system and the power consumption should be quite good, so heat won't be an issue. Definitely a huge step up from PS3 hardware, and "console ports" won't suck so bad since this thing is basically a real PC.

17 December 2008.

5 years? Might as well write a white paper on the benefits of drum memory over mercury delay lines.

The AC is dead-on right. At 25nm the endurance for high-quality MLC cells is about 3,000 writes. That's a relatively conservative estimate so you are pretty much guaranteed to get the 3K writes and likely somewhat more, but it's a far far cry from the 100K writes you can get from the highly expensive SLC chips. Intel & Micron claimed that one of the big "improvements" in the 20nm process was hi-K gates that are claimed to maintain the 3K write endurance at 20nm, which otherwise would have dropped even more from the 25nm node.

The author of the article went to all the time & trouble to do his mathematical analysis without spending 10 minutes to find out the publicly available information about how real NAND in the real world actually performs....

Obviious Troll is Obvious but... while SSDs can & do fail (just like old hard drives can & do fail), the reason for SSD failure in the real world is very rarely due to flash memory wear. Hint: If your flash drive suddenly stops working one day, that ain't due to flash wear, which would manifest as gradual failure over time.

Blah Blah Blah.

I'm well aware that the Government or hackers* could compromise Slashdot and find out who I am. I also don't care. I can also lie to you about who I am on Slashdot. Assuming you use the handle "Kenja" on any other online forums I can probably show you fun-filled research papers from people who are really good at data mining who could probably track you down with a very high probability just based on the content of your publicly-available posts and some educated cross-reference guessing. It's life, deal with it.

*Slashdot still runs on Apache 1.3 you know, and the "infrastructure" if you can call it that hasn't been updated since the late '90s. It's been compromised in the past and I feel that hackers don't bother with doing more damage because there isn't any money in it.

Controversial title but here's why: With the ability to use nicknames, you can delude yourself into thinking you have privacy when you really don't. With a real-name policy you are having your lack of privacy rubbed right in your face so you don't forget it and do something stupid under an "assumption" of privacy.

You want real online privacy? Don't use Facebook.

You think this violates the "anonymity" of the Internet? The Internet was never anonymous.. it's just that the Internet made it (and still makes it) difficult to verify that the other person at the end of the pipe is actually who he says he is and isn't lying to you. Don't confuse lack of authentication with privacy, they ain't the same thing.