rumblings of something huge going on with SEGA

x*1 is not necessarily x, and x*0 is not necessarily 0. The bottom bit of
the mantissa is pretty random.

Perversely, 1*x IS always x, and 0*x IS always 0. So the order matters.
That sounds disgusting. Is that the MIPS core or the VUs? Admittedly I did get something similar once on an Intel P4. Why the **** can't I do a proper, accurate x*y, where x and y are just whole numbers(but not int)? I still can't figure that out, to this day. But since that project was procrastinated indefinitely, I let it rest.
 
Simon, I sympathize with the right to be bitter :p, but are you actually telling me that hw in DC, CLX and SH4 in particular, were completely bug free?
Because I can tell you for certain that there are quite a few more stories like the PS2 FP issue, and they are on other consoles too.
But I honestly don't know about DC, so I can't argue there.

I mean seriously, you sound a little bit like AMD/Cyrix fans of old when that pentium FP bug was unveiled. Given it was pretty common for Intel cpus to ship with 70-80 bugs back then(I don't know what's it like today with growing complexity and all), were they spending less on development then PR too?
 
passerby: VUs I think.
Fafalada said:
Simon, I sympathize with the right to be bitter :p, but are you actually telling me that hw in DC, CLX and SH4 in particular, were completely bug free?
SH4: I couldn't say but, as I assume it went into other products as well, I guess it was relatively clean. It's not exactly a "big" chip.

CLX: I don't recall anything in the major functionality, but obviously I wouldn't want to stake my life on it. There may have been a bug, say, in the "on-the-fly, render without a framebuffer"(!) mode, but that was pretty esoteric functionality!



The only reason it was fresh in my mind was that we were discussing FPU behaviour and accuracy on the GDA mailing list and I said that you could guarantee commutivity, but not associativity, and someone said that the PS2 didn't even give you the former.

OK, I'll stop with this now.
 
Simon F said:
What I mean is that you see things like...
Tom F (on GD Algorithms) said:
x*1 is not necessarily x, and x*0 is not necessarily 0. The bottom bit of
the mantissa is pretty random.

Perversely, 1*x IS always x, and 0*x IS always 0. So the order matters.
(er... it may be the other way around - I forget - not that you can usually
control the order when coding in C).
:rolleyes: :rolleyes: :rolleyes:
And this is just one of many examples of the fine features of the PS2. One wonders if they spent more money on spin than they did on development.

Maybe ( as a feeble form of defence ) I would have to point out that this is also a symptom shown by the single precision fmac on the sh4 - To claim that the result is random though is to show a certain lack of understanding about multiplication and rounding ;)

( Quick - hide this argument before DM or Lazy8s picks up on it... :D )
 
PC-Engine said:
Fox5 said:
Evil_Cloud said:
london-boy said:
sytaylor said:
I want a new console from SEGA.. Since the days of the DC ended ive lost my love of gaming :(


Heh, since the days of DC, it seems Sega has lost their love in games........

Their focus is more spread... Making games for one platform is easier then making three optimised versions. Look at Smilebit, they've made only Xbox games, and have made some extremely impressive games. (Ok, Smilebit is/was also one of Sega's top development studios :))

Uh, very few of sega's games(besides sports) go multiplatform, so their focus isn't really move spread. They just make mostly crap, AM2, Smilebit,(overworks still there?), and AV still make fairly high quality stuff. Sonic team has been lacking, but they weren't that great on dreamcast in many people's opinions, and their one original game this generation was bad.... mainly because a lot of the level designs made no sense, it still had bad camera angles, and there was a complete lack of story or cohesion.....sonic and knuckles did a much better job of tieing the levels together.

I thought PSO on GCN was great.

PSO wasn't an original game....sure they added the extremely short episode 2, but no one actually played that anyhow. It was pretty, but it seemed most people felt the enemies and quests were way too hard for the meek rewards they gave.
 
Sega's big announcement on the first day... I thought I knew it before, but now I'm not so sure, what with someone so well-informed as Enterbrain's Hirokazu Hamamura talking about how "it'll surprise you". My wild guess: Mr. Yuji Naka shows up at Nintendo's conference the day before to announce a brand-new Sonic game for the DS that's some kind of revolutionary and demonstrates how the two screens can be used and (more importantly) is almost complete, and then they show it at this announcement.

http://www.video-fenky.com/
 
No, I'm asking you to guess.

Ofcourse, I know whether it's gonna be new hardware or not. I work there.

But I'm asking you to continue your guess work.

Let the speculation continue!
 
How are we supposed to believe you when a Sega employie(Damn, my spelling sucks at times.) has stated that you don't work at Sega?
 
SegaR&D said:
Yes, but how would he know me?

He works at Sega of America right?

I don't work there.


Alright then, which Sega of do you work for? Japan? Europe? Edit: Wouldn't mentioning future plans in the method yo're using break about 1500 NDAs anyway?
 
Did I say he didn't work there? I can't really recall, but I do not believe he works at any division of SEGA.
 
segrd, taking advantage of people looking for any news at all doesnt proove you are smart or anything ... just that you are an asshole, fuck off.
 
Back
Top