So, Sony didnt spend a dollar in customizing a byte?.
Thanks, if there's any truth to this guy's story, it look like the Sony delay is not related to MS (could be because Sony change from steamroller to jaguar, because of steamroller cancellation). It says Orbis was already close to mass production. So maybe they shifted engineers on Durango simply because Sony's chip is practically ready. I agree it sounds like Durango had more custom stuff but we already know that, I doubt the esram is an "easy" additionYeah I found the original quote (I think) as I never really saw it myself and only remembered people talking about it. It seems that Sony's delay is open to interpretation as to why it happened.
The note about supercomputer may refer to Durango's one-off customization to maximize compute power. They are moving to use more mainstream GPGPUs to save $$$ these days.
Orbis was delayed for 6 months. It's also evolved from the old Steamroller and 2GB memory rumor. There was probably a Cell-based early work as well.
They wanted to put steamroller in with its super fpus and finally changed to vanilla netbook fpus based cores?.The strangest thing is they for sure knew MS was going to customize its jaguar fpus.And is not like they added cus on change because the 18 CUs were there from the beginning.So if something, apart from the 4 GB upgrade Orbis was downgraded from initial target specs.
I wonder if vgleaks is going to have an article about durango's cpu tomorrow?
The note about supercomputer may refer to Durango's one-off customization to maximize compute power.
These systems are designed to work the way their makers want them to work and not because of some reactionary whim. The fact that some people think its done this way is not only wrong but I find it rather amusing.
If they have a target in mind, they will find something else to reach the target. This may be why we heard custom compute unit rumors for Orbis.
DF's compute unit comment is somewhat different from VGleaks. There may be further changes after the CPU switch.
The 1.84Tflop figure was there since 2011,so if the 4 special cus were there to compensate the weak fpus they still are detraining resources from the target gpu power.
Maybe they were originally thinking of the CPU doing a lot of heavy-lifting, but switched to DSPs and decompression blocks?They wanted to put steamroller in with its super fpus and finally changed to vanilla netbook fpus based cores?
No no, don't put words in my mouth. I just said that there would be no documentation bearing the words "Jaguar". And it's one L, not two. I am not an irish beerAre you saying bgassassin is lying? Wrong? Trolling us? Has bad intel? What are YOU assuming in the process? Honestly bkillian's reply to my question there seems to strongly imply it's not Jaguar at all.
Indeed, if Apple are using the fan tech they developed for their macbooks, it's brilliant. It still won't work out at the distances required for a console. The Kinect pipeline gets around 40DB noise reduction, but takes 1/8th of a 360 core to do it. And directional mics, while they _may_ be useful on a mac mini, although I'd doubt it, would not work on a gaming machine where the field of play is something like 100 degrees wide.I guess you've never seen and not heard a MacMini? Because I have a quadcore i7 and two HDD RAID array in mine and you know what? I never hear it - including the fan. By the way, the types of mics they'd use are directional. They current mics are directional because they don't want to pick up the ambient noise around them, like from A/V equipment and the TV These types of problems were solved years ago.
Directional mics, as I pointed out above, would not work for a large playspace.So when bkilian says they have the world's first and only open mic with a fan attached, and mentions a static tone remover would work but be daft (it won't deal with non-constant vibrations from HDD and optical), is he lying? Or just being an idiot, because all Kinect needs is a directional mic and that solves all background noise and vibration issues?
Yep, for multiple reasons. One of them being that MS always owns Dev kits, this way, if a game company goes bankrupt, MS can claim back their property without it being able to be auctioned off.Utter bullshit. This has been debunked by BKilian on these B3D forums. MS never provides a spec-sheet and tells the developers to go build a dev-kit.
Phone mics use a simple form of noise reduction. They don't have to be amazing because your mouth is literally an inch from the mic, and it's mono. Some phones use a second mic to try remove ambient noise. I don't know how good those are, but again, the second mic is 5 times further away from your mouth as the first, but almost the same distance from ambient noises. I could probably write a filter that isolates the voice in a system with that configuration pretty easily. The Kinect stuff, on the other hand is a different ball of wax. For instance, your speakers are most likely much closer to the mics than the player, and have a lot more energy. The audio tech behind kinect is the CS equivalent of rocket science, and took a good six months of 12 hour days by a reasonable sized team, including a full time researcher, to get right. From the research page on AEC:The MacMini sits on a glass desk, the same one I use. And yeah, of course it vibrates. Slightly. Again directional mics to capture sounds ahead and not indiscriminately. This isn't expensive technology, it is, and has been, used in all sorts of consumer gear. Phones, tablets, portable gaming machines. It's already cheap.
Stereo, and later multichannel (for surround sound systems), acoustic echo cancellation is not a trivial problem and sparked a lot of interest in the research community. The problem is that the stereo channels are highly correlated which leads to infinite number of solutions. Only one of them is the real solution, for all others the AEC has to readapt when something changes in the stereo signal.
I don't know. Is the Steamroller rated higher than the Jaguars in FLOP count ?
No. 8 1.6 GHz Jaguars would be the same max theoretical flop count as 4 3.2 GHz Steamrollers. Jaguar can issue one 128-bit FPADD and 128-bit FPMUL per cycle per core whereas Steamroller could issue 2 128-bit FPMAD per module (two cores). So in terms of flop count, both can execute 8 single precision flops per clock per core.
I think the biggest performance difference between Jaguar and Steamroller is the number of instructions that can be decoded per core in the frontend. For Jaguar it's two and for steamroller it will be four.
No. It doesn't. I've taken the specs as they've been reported and explained what function they would have in that system. The motivations behind the design decisions don't enter into my argument whatsoever, no matter how many times you attempt to drag it into the equation.
Cute.It's obvious you are not equipped to even understand, let alone dispute the substance of my argument, so keep in mind this will be my last response to you.
You didn't actually respond to the points I raised specifically on your assumption here. You don't even know what Durango's setup is capable of at this point yet you are eager to gloss over points raised by others just to assert that whatever works well for Durango must just automatically work as well or better for Orbis? I am certain I have read ppl on this very forum lay out criticisms of exactly this assumption. Maybe they were wrong. The same counter argument was made in that article from that Spanish blog that made the rounds a week or two ago.So it smartly avoids limitations that exists in the Durango to give near peak performance, while simultaneously improving performance on Orbis by minimizing unnecessary memory access.
Oh goodie, a tantrum.It wasn't a rhetorical gimmick. I was calling you out on your obvious hypocrisy you ass.
No, my 'entire argument' is critiquing your assumptions, which is absolutely based on a premise that you've yet to justify with any evidence. According to what you said earlier the extra kit isn't there to leverage virtualized asset handling in a meaningful way, it's just there to reduce a performance gulf between Durango and a competitor's console that MS didn't even know about when these decisions were likely being made.YOUR entire argument assumes I must be irrational, and hinges on a misreading I've already corrected for you.
Quote me saying this please. Nobody is attacking you. There is no need to have a meltdown trying to defend your arguments with personal attacks and arrogant, condescending assertions left and right. You can defend your premise without that emotional baggage. It's OK to note an assumption you have made and then carry arguments based off of that through to conclusions. There is nothing wrong with starting with an axiom. Us physicists do it all the time. But pretending like it's not there isn't helpful and only misleads ppl when your posts are so bloated with arrogance.YOUR first post accused me of being blind to reality.
Ooook guy. Moving on now...you clearly would rather have an emotional meltdown and feign victimhood than engage in skeptical inquiry. Yikes.This is rich, considering you are challenging me on technical aspects you clearly don't understand. YOUR entire tactic has been to attack my character and invent false motivations to draw my analysis (which you couldn't follow) into doubt.
No no, don't put words in my mouth. I just said that there would be no documentation bearing the words "Jaguar".