Your source for hardware that's NOT just x86

I’m thoroughly unimpressed. I was holding off on getting a good CPU because I thought a Phenom II X6 would be pretty good, but it doesn’t provide enough of a performance advantage over the Phenom II X4, and what little advantage it ekes out over a Core i7 920 (really, the price point is at a 930 level) can be easily mitigated by overclocking the i7 920 to 3GHz. The Phenom II X6 1090T BE costs one buck shy of 300EURs, and doesn’t perform like one. OK so nobody was expecting a Nehalem killer, but with 2 more cores, can’t it at least put more distance?

To be specific, it barely manages to fend off the Core i7 920 in heavily threaded apps, which costs 220EURs. In Anand’s x264 encode, in the second pass, the X6 1090T BE is faster than a i7 860 and 920, but only ~2fps faster. In this case, it’s about 40% performance increase for a 50% increase in core count, which sounds good but still doesn’t look good next to the i7. Again, i7s are VERY overclockable. And the Core i7s are back on top in the 3DS Max rendering tests. It manages to eke out a 10% performance advantage over the Core i7 920/860 in Cinema 4D, but again, that’s not very impressive when you have 2 extra cores and 600MHz more.

And in the archiving tests, the Core i7s remain firmly on top.

Bit-tech wasn’t impressed, and I think they’ve got the right impression. It might be due to the 6MB L3 cache staying at the same size, despite the core count increasing – then again one only needs to look at the Athlon series to figure that the K10 architecture isn’t too sensitive to cache sizes. It’s only a 512KB drop for the Phenom II X6. There has to be something at fault here – I can’t believe 6 cores from AMD running at 600MHz more than their quad core Intel counterparts can only manage to break even in the most heavily threaded of workloads (rendering, archiving, Photoshop and video encoding), despite a very real IPC disadvantage. And yet Anand said “Applications like video encoding and offline 3D rendering show the real strengths of the Phenom II X6.” The biggest gain I ever saw over the Phenom II X4, again, was 40% (max theoretical gain is 50%), and that level of performance has already been achieved with quad core Core i7s, which have been out longer and are subsequently selling for less at the moment. Where’s the strength in this?

According to everyone around the internet, those rumours about Apple’s tablet (I won’t call it a tablet PC, because those Mac folks might frown) were true, because Jobs just announced the iPad. Everybody hopes this will rejuvenate the tablet PC market, which so far has been quite disappointing after everyone jumped on the bandwagon 2 years or so ago and failed to catch any fruit on it.

The amazing thing about the iPad is that it’s a device that has 16GB of flash memory, 802.11n wifi, a freaking IPS panel, and it starts at just USD 499 which is cheap for anything from Apple. And it supposedly is compatible with iPhone apps. It’s also designed like an iPhone, but bigger. No camera, though. And for a lot more, you can have the ones with 3G support. Those come with a plan from AT&T, though, and they don’t seem to be that worth it.

So with it being able to run iPhone apps, it must clearly be an ARM CPU right? Well, all Apple says about the CPU is that it’s an “Apple A4″ chip running at 1GHz. Yeah, as if we believe it’s all made by Apple. Anandtech guesses that it’s probably a Cortex A8/A9 with PowerVR SGX video, and adds that the video format support is quite paltry. Hopefully that will be fixed with 3rd party apps.

Also, Apple claims 10 hours of web browsing on it, and over a month of standby power. Wow, 10 hours? 10 hours? Jesus, Jobs. This had better be true, because I’m actually thinking of buying it.

As a lot of people must know by now, the famous iPhone jailbreaker George Hotz, 20 years old, younger than me by 8 months (fuck), has hacked the PS3. Specifically, he has read/write access to real memory in the PS3 under OtherOS, ironically with help from the hypervisor that the OtherOS runs under. Needless to say, this doesn’t work on the new PS3 Slim. Suckers.

You can download the source code for the exploit and instructions here. You need some way to lower the voltage between the two pins in the photograph for 40ns repeatedly to glitch the memory bus, so that means you need to start cracking open your fat PS3 and solder some shit to it. I think he used a Spartan3 FPGA to do the trick. Sometimes it works, sometimes it doesn’t, as it says in the readme in the zipfile. Which reminds me of the 5.03 ChickHEN exploit for the PSP-3000 – I always laugh at my friend when his PSP-3000 shuts down and he always has to go through the TIFF thing a hundred times before getting the green screen. All the while I’m playing Dissidia on my PSP-2000.

geohot: well actually it’s pretty simple
geohot: i allocate a piece of memory
geohot: using map_htab and write_htab, you can figure out the real address of the memory
geohot: which is a big win, and something the hv shouldn’t allow
geohot: i fill the htab with tons of entries pointing to that piece of memory
geohot: and since i allocated it, i can map it read/write
geohot: then, i deallocate the memory
geohot: all those entries are set to invalid
geohot: well while it’s setting entries invalid, i glitch the memory control bus
geohot: the cache writeback misses the memory
geohot: and i have entries allowing r/w to a piece of memory the hypervisor thinks is deallocated
geohot: then i create a virtual segment with the htab overlapping that piece of memory i have
geohot: write an entry into the virtual segment htab allowing r/w to the main segment htab
geohot: switch to virtual segment
geohot: write to main segment htab a r/w mapping of itself
geohot: switch back
geohot: PWNED
geohot: and would work if memory were encrypted or had ECC
geohot: the way i actually glitch the memory bus is really funny
geohot: i have a button on my FPGA board
geohot: that pulses low for 40ns
geohot: i set up the htab with the tons of entries
geohot: and spam press the button
geohot: right after i send the deallocate call

I hope to goodness there will be a software exploit and that those nouveau guys can finally start porting their 3D driver code to PS3 Linux. I bought my PS3 thinking that it was gonna be a cheap replacement for a computer and all I got was some restricted piece of crap. I want the hardware I paid money for.

All code executing on the Cell in OtherOS seems to be encrypted. Supposedly the reserved SPU under OtherOS is doing all the encrypting, although this exploit allows one to get past that. Hotz isn’t working on the PS3 full time now, but he does say that he’ll try to find the root key for the Cell, which would be really awesome if he did, because I’m sure Sony will try to patch this ASAP. For an easy to understand explanation, see Digital Foundry.

EDIT: I take that back, root labs rdist has a much better explanation of the exploit that goes in-depth.

Close enough to studying, I guess.

I have no clue how Artefact and Danny Choo manage to do this every day.

EDIT:

I also forgot – I will be forfeiting the incredible rise in visitors I’ve had because of those damn exams. Sob.

X Japan goes to the USA! And instead of sitting around on their asses and doing absolutely nothing, they wrote ONE new song called Jade!!!



OK, so that’s not really much. And I.V. was a kinda weak song. They really need to come up with more old school X songs like Stab Me in the Back, Orgasm, Standing Sex, Sadistic Desire, and cool stuff like White Poem I and Voiceless Screaming. And of course, more ballads. But I’m just glad they’re back. I almost cried manly tears when I watched The Last Live’s Endless Rain. MANLY TEARS. But I wasn’t in touch enough with Real Life that I missed X Japan coming to Germany last year (NOOOOOOOOOOOOOO!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!).




Great, now I’m depressed.

Thanks ANN.