"The computers will not have artificial intelligence, and scientists remain many years away from building one that matches even the abilities of a simple mouse brain."
As a last resort, can we give the gentle push to the Moon instead, knocking the asteroid away (aside/to kingdom come) with it? Surely Mother Earth would be much confused after, lack of tidal effect and all, but at least alive. Needless to say, this is quite a one-off method.
This Shuttle case is very noisy. The "liquid cooling" -- as Alienware calls it -- is in reality a heatpipe based CPU heatsink offloading to the 80mm fan at the back. And, unfortunately, the fan is very noisy. Also, the case doesn't suppress hard drive noise at all. (Saw this in a magazine reviewing the same Shuttle hardware a couple of weeks ago. I'm sure Google comes up with reviews on web rags.)
Unless Alienware managed to get a quiet(er) fan in there -- which I doubt because of the system specs, all hot stuff -- this is not a system for the living room.
Even the GF4 series incorporated similar technology to the Kyro chips
Beg your pardon? GF3/4 may employ some tiling to access video memory smarter, and do some occlusion culling with bounding boxes, but they don't (defer the scene and) bin the polygons for visibility sorted (and, in this case, tile based) rendering -- Kyros are totally different beasts.
Agreed.
Once your Nvidia card gets a year old, the unified drivers start crapping on it. The optimisations for latest models start actually hurting older models. (But you still need the newer drivers to keep up with OS releases et al.)
I like ATI's model better. It's clearer and the result is more predictable, now that they actually produce pretty good drivers over there.
I could remember this wrong, but I think Carmack asked for 64-bit color, 16-bit per component. GFFX (quite doesn't roll off the tongue, does it?) offers this and also 32-bit per component. I don't think anyone has seriously requested 64-bit per component -- Pixar et al. are still at 16, and SGI hardware is at 12. It'll be at least Doom IV before 64-bit per component, i.e. 256-bit color is mandatory:-)
But their war chest is filled and they'll keep at it until Symbian is just a distant memory (hello OS/2, Wordperfect, Palm, Nintendo)
But what if Nokia and Sony/Ericsson take Symbian (in some form) to every damn cell phone they manufacture, making even the cheapest consumer models have some functionality? Microsoft can't really fight them there unless it too is in the Top 2 or 3 of cell phone manufacturers. And I don't see that happening, at least not in 5 years! While Microsoft may gain ground in PDAs and smartphones, the low end is out of reach for them, and that may be significant. I'm no expert here tho.
Just to clarify:
Sony has managed to fit the entire Emotion Engine + CPU + sundry other parts onto a single chip
The EE (which consists of the MIPS CPU core and the sundry other parts) and the GS are discrete chips, so PS2 is not a single-chip design.
they've gone and asked IBM for a radically different kind of chip
"CELL" (now "Grid") was originally Ken Kutaragi's funky (if vague) brainchild. Now Grid is being co-developed by IBM, Sony, and Toshiba (the designer and fabber of EE). It's kinda funny that IBM makes the Gecko CPU for the GameCube whily working for PS3... but for sure they have their own projected uses (networking gear, for example) for Grid technology!
This detail is slightly incorrect. PS2 does rely heavily on the FP power of the Emotion Engine CPU, especially for T&L (which Xbox can leave to the XGPU) -- the Graphics Synthesiser is just triangle setup and rasteriser. The MIPS core in the EE is nothing to write home about, but if you count in the accompanying co-processor and especially the muscular Vector Units, with a grand total of 10 FMAC units (each does a FMAC per cycle), well thats' FP power roughly equal to a 3GHz P4, IIUC. It's massively parallel, in short.
(Otherwise I agree with your post. PS2 just is a weird beast out there.)
At some time, Enermax PSUs tended to have uneven voltage regulation under high loads. Haven't tested the newer models though. My own 431W sometimes has problems keeping the 12V rail up to snuff and specs, but has never failed to spin-up the drives; thus appears good enough. (Regarding a previous post, certainly the 5V rail has nothing to do with hard drives.)
AFAIK, it was the first batch from the brand-new Hungarian fab that caused all these problems.
When they got the particle-filtering et al. over there up to snuff, the 75GXP drives have been good as usual.
I had two of them fail on me, with large losses of data. Both replacements have behaved, tho. (Irish, both.) All this ate IBMs good rep tho, I agree.
Amen, bro. When you're going to a party and the host has asked for your help with the music, it's just easier to bring a HD than burn a dozen CDs trying to figure out what the audience is this time.
(-1: Ungeek)
So, out of courtesy, you mention it on Slashdot, hopefully not advertising it to any of SR's audience?;-)
Just amused, no offense intended. I'll be sure to check out SF, although I'll be shy to reg 'n post.
[i]Certainly, there are lots more drives out there, but they are failing at a higher rate.[/i]
Is that rate "failures per drive" or "failures per gig"? IOW, if I wanted to be safe, should I buy two smaller "very old" drives or one larger "quite old" drive? Mind you, I'm not making a joke here; I don't currently posses the cavalier mindset requisite of jocularity. I'm thinking replacing my current Hungary-made IBM GXP75's with a reliable system, preferably before these buggers fail too.
Um, size does matter -- 120mm won't take all the hurt away, but it lowers the noice to much more tolerable frequencies. Whip out your dremel for a small case mod, tho.
Slashdot Mirror
From the article:
"The computers will not have artificial intelligence, and scientists remain many years away from building one that matches even the abilities of a simple mouse brain."
Let alone a Microsoft Intellimouse!
No, it's "Keep your eggs in one basket, and watch that basket!" (--Mark Twain)
As a last resort, can we give the gentle push to the Moon instead, knocking the asteroid away (aside/to kingdom come) with it? Surely Mother Earth would be much confused after, lack of tidal effect and all, but at least alive. Needless to say, this is quite a one-off method.
Arthur Dent would call that "The Beowulf Cluster".
This Shuttle case is very noisy. The "liquid cooling" -- as Alienware calls it -- is in reality a heatpipe based CPU heatsink offloading to the 80mm fan at the back. And, unfortunately, the fan is very noisy. Also, the case doesn't suppress hard drive noise at all. (Saw this in a magazine reviewing the same Shuttle hardware a couple of weeks ago. I'm sure Google comes up with reviews on web rags.)
Unless Alienware managed to get a quiet(er) fan in there -- which I doubt because of the system specs, all hot stuff -- this is not a system for the living room.
Even the GF4 series incorporated similar technology to the Kyro chips
Beg your pardon? GF3/4 may employ some tiling to access video memory smarter, and do some occlusion culling with bounding boxes, but they don't (defer the scene and) bin the polygons for visibility sorted (and, in this case, tile based) rendering -- Kyros are totally different beasts.
Agreed. Once your Nvidia card gets a year old, the unified drivers start crapping on it. The optimisations for latest models start actually hurting older models. (But you still need the newer drivers to keep up with OS releases et al.) I like ATI's model better. It's clearer and the result is more predictable, now that they actually produce pretty good drivers over there.
I could remember this wrong, but I think Carmack asked for 64-bit color, 16-bit per component. GFFX (quite doesn't roll off the tongue, does it?) offers this and also 32-bit per component. I don't think anyone has seriously requested 64-bit per component -- Pixar et al. are still at 16, and SGI hardware is at 12. It'll be at least Doom IV before 64-bit per component, i.e. 256-bit color is mandatory :-)
But their war chest is filled and they'll keep at it until Symbian is just a distant memory (hello OS/2, Wordperfect, Palm, Nintendo)
But what if Nokia and Sony/Ericsson take Symbian (in some form) to every damn cell phone they manufacture, making even the cheapest consumer models have some functionality? Microsoft can't really fight them there unless it too is in the Top 2 or 3 of cell phone manufacturers. And I don't see that happening, at least not in 5 years! While Microsoft may gain ground in PDAs and smartphones, the low end is out of reach for them, and that may be significant. I'm no expert here tho.
Although it's got GSM/GPRS, I'd still put it on the PDA side of the PDA/phone divide. Matter of taste, perhaps...
Plus the design really has legs.
The above mess taught me to use the Preview button.
Just to clarify: Sony has managed to fit the entire Emotion Engine + CPU + sundry other parts onto a single chip The EE (which consists of the MIPS CPU core and the sundry other parts) and the GS are discrete chips, so PS2 is not a single-chip design. they've gone and asked IBM for a radically different kind of chip "CELL" (now "Grid") was originally Ken Kutaragi's funky (if vague) brainchild. Now Grid is being co-developed by IBM, Sony, and Toshiba (the designer and fabber of EE). It's kinda funny that IBM makes the Gecko CPU for the GameCube whily working for PS3... but for sure they have their own projected uses (networking gear, for example) for Grid technology!
Sony gets away with using a ~300mhz (IIRC) cpu
This detail is slightly incorrect. PS2 does rely heavily on the FP power of the Emotion Engine CPU, especially for T&L (which Xbox can leave to the XGPU) -- the Graphics Synthesiser is just triangle setup and rasteriser. The MIPS core in the EE is nothing to write home about, but if you count in the accompanying co-processor and especially the muscular Vector Units, with a grand total of 10 FMAC units (each does a FMAC per cycle), well thats' FP power roughly equal to a 3GHz P4, IIUC. It's massively parallel, in short.
(Otherwise I agree with your post. PS2 just is a weird beast out there.)
I agree. Shame on me that I bought, lemme see, a few Athlons while waiting. Sorry. X86 sucks.
Bhawahaha, but I'll wanr my RDRAM anyway. Evil, sorry.
At some time, Enermax PSUs tended to have uneven voltage regulation under high loads. Haven't tested the newer models though. My own 431W sometimes has problems keeping the 12V rail up to snuff and specs, but has never failed to spin-up the drives; thus appears good enough. (Regarding a previous post, certainly the 5V rail has nothing to do with hard drives.)
AFAIK, it was the first batch from the brand-new Hungarian fab that caused all these problems. When they got the particle-filtering et al. over there up to snuff, the 75GXP drives have been good as usual. I had two of them fail on me, with large losses of data. Both replacements have behaved, tho. (Irish, both.) All this ate IBMs good rep tho, I agree.
Amen, bro. When you're going to a party and the host has asked for your help with the music, it's just easier to bring a HD than burn a dozen CDs trying to figure out what the audience is this time. (-1: Ungeek)
So, out of courtesy, you mention it on Slashdot, hopefully not advertising it to any of SR's audience? ;-)
Just amused, no offense intended. I'll be sure to check out SF, although I'll be shy to reg 'n post.
[i]Certainly, there are lots more drives out there, but they are failing at a higher rate.[/i] Is that rate "failures per drive" or "failures per gig"? IOW, if I wanted to be safe, should I buy two smaller "very old" drives or one larger "quite old" drive? Mind you, I'm not making a joke here; I don't currently posses the cavalier mindset requisite of jocularity. I'm thinking replacing my current Hungary-made IBM GXP75's with a reliable system, preferably before these buggers fail too.
Um, size does matter -- 120mm won't take all the hurt away, but it lowers the noice to much more tolerable frequencies. Whip out your dremel for a small case mod, tho.
Messier Peltier has a massage for you.
No, my background color remains intact. Pathetic?
Thank you very much! All clear now :)