The only problem with this is Intel's SpeedStep implementation in the P4 / Xeon isn't aggressive. The best you'll see a SpeedStep-enabled processor drop down to is 2.8 GHz at 1.2v. this gives you a maximum %20 power reduction from frequency, and a maximum %25 power reduction from dropping the voltage from 1.4v to 1.2v, for a total potential drop of 45%.
This is in contrast to all Cool 'n Quiet-enabled Athlon 64 processors, which drop to 1GHz @ 1.1v. This provides plenty of processing power to do mundane tasks, as well as deliver snappy performance to servers at low usage levels.
The agressive reduction in frequency and voltage gives you a processor that sips 3w at idle, and less than 10w at light load (1.0 GHz).
IIRC both AMD and Intel are realising what VIA and Transmeta did years ago--that the clock-speed race is completely pointless now. Not only do both companies invest heavily in mobile technologies--that technology is finding its way into desktops and servers as well. About bloody time I think. It'd be great if this turns into a golf-score style contest of watts-per-flop race.
VIA didn't realize that the clock speed race was pointless. VIA went "low-power" out of sheer necessity.
The Winchip architecture was never intended as a power-house, and unfortunately VIA didn't get their hands on it until it had languished for a few years (and Intel and AMD had pulled far ahead). By that time, they could nave catch up in terms of performance per clock, or performance per watt.
You seem to forget that the VIA Epia never really had ANY performance per watt lead, and it never had a significant lead on other low-power x86 solutions. For example, the P3 Tualatin LV, a CPU released around the SAME TIME as the Epia platform, ran as fast as 800 MHz and consumed 7w max. Despite having the same limited memory bandwidth as the Epia platform, it performed twice as fast clock-for-clock.
Today, we have LV Pentium M derivatives that blow the Epia out of the water. They're just not as popular as the Epia platform because the chips and boards aren't cheap.
In addition, people just aren't as willing to look past pre-built solutions to find a low-power platform. Take a look at AMD's DESKTOP processors. Every single one of them offers the ability to run at 1.0GHz at 1.1v (Cool 'n Quiet mode). It can been shown that an AMD 64 DESKTOP part running at 1.0GHz and 1.1v produces 10w under full load, and around 3w idle...VERY competitive with VIA's specs.
Basically, VIA is doing everything they can to carve a niche for a CPU that really should have been beaten into the ground long ago. The performance is pathetic, so they leverage the "embedded, low-power x86 compatibility" angle and try to make ends meet. The thing is, the market for the Epia is so tiny relatively that not even AMD feels it is worth tackling.
But just because Intel and AMD don't want to tackle the tiny "x86 low-power kiosks and hobby platforms" market, that doesn't mean we'll have to forgo low power. My desktop AMD system already uses the same power at idle as your anemic Epia system, and costs less. Once Intel starts offering Conroe, they'll have the same.
When it comes to CRTs, Sony was the best only three years ago. When Mitsubishi started pushing their "DiamondTron" flat-CRTs, totally ripped off from Sony's Trinitron technology.
You act like Mitsubishi was stealing candy from a baby. You also act as if competition is a bad thing.
Sony had their chance at the Trinitron. They've been making it since 1969, so they've gotten more than their money's worth out of the technology.
Normally such a market lead as Sony had with such a sophisticated product would preclude other companies entering the market, for fear of never catching up. But Sony was complacent, never offering significant improvements on the Trinitron beyond the 70s, and always charging a premium. When Mitsubishi saw an opportunity in the market in the 1990s, they caught Sony with their pants down.
Mitsubishi was the first Aperture Grille manufacturer to produce a completely flat, distortion-free tube (marketed as Diamondtron NF), and they pushed Sony to release their own version (FD Trinitron) in response. Who knows how long we would have had to wait for Sony to do it without competition, especially with the explosion of LCDs looming on the horizon?
The best part is, Mitsubishi has worked to reduce the price of their Dirmondtron NF line, effectively removing the premium long associated with the Trinitron. The end result is, you can still buy a quality Aperture Grille tube, even though Sony can't seem to make theirs cost-effective. Sony no longer matters.
I've been reading through the discussion, and I've been thinking of responses, but it's all a muddied mess out there. so, I've decided to lay out the basic discussion points and my thoughts as one post.
First of all, why do we need faster-response LCD screens, when we already have 4ms?
There are a few key reasons for this. For starters, the 4ms number doesn't mean much. It is the time the panel takes to turn a pixel from black to white, then back to black. In a traditional panel, this is usually the fastest transition possible...and all other tranitions (Grey to Grey) are MUCH slower. Sometimes GTG transitions can be as much as 3x slower than the Black-White-Black number.
The industry has concocted a possible solution to this called Overdrive.
Overdrive takes advantage of the fast transition in Black-White-Black. Every time an input pixel changes color, the pixel on-screen is bootsted up to white, and allowd to fall back down to the new color.
This is slightly slower than the Black-White-Black transition time, but it's much faster than going Grey-to-Grey.
Unfortunately, Overdrive has a drawback that is DIRECTLY tied to the response time. Every time a pixel changes, it is overdriven WHITE for a fraction of a second, until it settles down to the target color. In darker scenes, or in cases where where colors are almost uniform, as pixels change these white pixels are painfully obvious. Better response times are the only thing that can remove this annoying artifacting.
This link to Tom's also addresses the other issue discussed in this thread:
What's wrong with 18-bit color?
The dithering algorithms used by panels to simulate 24-bit color are not all that bad, but they have a serious drawback:
Dithering yields poor quality in scenes which require high contrast. Foggy, smoky or dark scenes, which tend to have subtlte color transitions, look like crap on an 18-bit panel. The panel is constantly changing pixels that are VERY close to each other in color, resulting in a muddy image. Unfortunately, the only way to avoid such artifacts it to buy an MVA panel with true 24-bit color (and sacrifice response time).
It was developed as an open standard to ensure industry-wide acceptance. Intel does not need a license to develop and sell an x86-64-compatible chip, and neither do you.
According to Marc Miller of AMD: x86-64, now named amd64, was developed as an open standard, therefore it's not something we can license.
The open standard is x86-64. AMD's trademarked implementation is called AMD64. Intel's trademarked implementation is called EMT64.
- Divx, even the unlocked discs, were not playable on DVD players (nor could you take unlocked discs to another divx player and play them for free) . These discs, on the other have, have zero restrictions.
- Divx still required you to drive to the store and put down a whopping $5 for the first 48-hr rental (with typical unlocking fees of $20-25). This has much lower total fees, and does not require you to leave the house.
Honestly, if they allowed me to pay regular PPV price, then make my decision to buy after the movie shows, I would actually try this service out. But knowing Comcast, they'll try to make you buy the whole thing up-front.
you see the whole movie and than you get the hard copy. hey $17 isnt it bit costly . so why do you need the hard copy. is it for the video library or what. huh.
Ahh, I've broken his secret alien code! If you carefully re-arrange the words, you get this:
Huh? What you have isn't the whole costly movie. You see what the hard copy is after the video library seen it.
So, you get $17, so why do you need the hard copy?
Hey, is it for or and what than?
See, it all makes sense now! The ton foil hats, the satellite dishes, the voices in my head!
Yeah. The article talks about Kahuna featuring improved "basic" functionality. How about this for "basic":
Outlook allows you to open multiple emails at once, so that I can parse through them quickly. Hotmail, on the other hand, makes you click-through to read each email (and each time you do this, the inbox disappears). In addition, the Hotmail UI prevents the opening of emails into their own tabs / windows, so I can't open all the emails I'm interested in at once.
If they're going to work on the "basics," and have a rich applications framework, I sure hope they're planning on providing "basics" available to Outlook Express users for the last decade. I'd actually READ my Hotmail inbox more than twice a week if it were as easy as reading my Outlook inbox.
* Patiently waiting for the next generation in web email since I signed up for my Hotmail account in 1997.
While I certainly don't agree with the outright gifting of cars, there's no reason cars can't be loaned to kids if the circumstances are right.
I put myself through college mostly on my own dime. I worked my ass off in high school to get good grades, which turned into merit + need based financial aid. I worked full-time summer, and part-time during the year.
For college, I took my high school economics teacher's advice: that a car was a money pit. I didn't need a car living so close to campus, so for the first two years I didn't have one. But I did need a car to move home during the summer and winter breaks, and my parents got tired of that.
When they bought a new car, instead of trading in their old 91 Corolla, they loaned it to me. Since I didn't own it, they stipulated that they would pay the insurance if I handled the upkeep, and transported myself. They figured it would solve their problem while giving me more freedom.
(And my economics teacher was right. That car was STILL a money pit)
It was a good compromise. They didn't come back to my campus until graduation, and shortly thereafter got my own car. The Corolla was later loaned to my sister.
As another example, I don't think the modern game hardware industry owes much to Silicon Graphics - their hardware was ahead of PC hardware for a long time, but at a price most people wouldn't pay. When the time was right, graphics hardware became widespread, but it was no thanks to SGI who were trying to maintain the old prices.
Are you kidding me? The modern game hardware industry owes a lot to SGI.
First, let's talk about SGI's direct effect on the industry:
SGI designed the N64. It was basically all the best bits of a $10,000 Indy workstation, shrunk down and sold for $250. Sure, it was a little memory-deprived, but that was a result of the pricepoint. Its 3D prowess was unmatched by anything consumer-level until the Voodoo Graphics was released for the PC a year later (which still cost more, intro-ed at $300), and was unchallenged in the console arena until the Dreamcast.
SGI's '$250 Indy' featured hardware support for mip-mapping, billinear filering, anti-aliasing...tons of features typically only seen on SGI workstations.
Now, let's talk about the indirect effects SGI had on the industry.
SGI developed OpenGL. This had the following indirect, but lasting effects:
* 3dfx based their GLide API on a subset of OpenGL, removing some of the professional-only instructions, as well as the transform and lighting hardware support, in order to implement a cheap 3D pipeline. 3dfx was the most influential hardware company in the early consumer 3D industry.
* OpenGL made Microsoft react and make Direct3D a solid development platform, and encouraged MS to innovate even when they surpassed OpenGL. The ability of OpenGL to use unofficial extensions kept Microsoft on their toes. SGI's industry clout was key in getting OpenGL support built into Windows 95b / 98.
SGI lead the way in making 3D a commodity, but once 3D became a commodity they lost the reins.
But in REALITY, we know that there are white-hat, and black-hat hackers.
Once the community finds the vunerability, it is free game for the white-hats (firmware downgrader) and black-hats (firmware toaster). You should feel lucky that a remote exploit hasn't been found (yet).
Sony has provided a fix for the vunerability, because they have TWO significant financial interests. Most of you think there's only one (locking down the system to enforce licensing fees from developers), but there is another very serious one: keeping the system secure to avoid lost sales and a class-action lawsuit once users are left holding a bunch of hacked-up bricks.
Since the console is wireless, security is extrodinarily important.
They simply read Robin Cook's "Contagion", which was released back in 1996.
In the story (minor spoiler),
A hospital has recurring internal infection problems with exotic disesases, including plague and Spanish flu.
(major spoiler)
Certain parties that enjoyed collecting and cultivating viruses come across specimens in the yukon permafrost frozen to death, infected with Spanish flu. These parties eventually become involved in a plot to discredit a hospital by planting viruses to cause internal infections, and use the Spanish flu to cause panic (and unknowlingly threaten an epidemic).
And thats the lowest that proc will ever be, though it may be replaced with a slightly better one in that tiers slot?
Yup, you got it.
The XP filled in for about two years as AMD's value line. Now that XP production has stopped, AMD has retired the Socket A platform. The Sempron (Athlon 64 on socket 754 with less cache) is the new budget processor.
The Sempron fills the $140 and less market, and the Athlon 64 socket 939 is priced above that. Due to the fact that AMD holds the performance crown, this will not change anytime soon unless Intel starts a price war. AMD is already having trouble meeting demand with their current prices, so they have no reason to lower them.
EVERY console that has been worth its salt has rivaled the arcades on release.
The Atari VCS was incredible when it came out, capable of quite accurate renditions of hits like Space Invaders and Bezerk. Sure, by the time Pac Man was popular (4 years later), it was faltering, but that's what happens.
The NES produced surprisingly accurate versions of arcade hits like Ikari Warriors, Super Dodge Ball and Double Dragon, just to name a few.
The Sega Genesis was BASED on Sega's System 16 arcade board, which was the platform for great games like Altered Beast, Alien Syndrome, Golden Axe and Shinobi (link has full games list). Even the later Street Fighter II and Mortal Kombat ports were quite complete (but the graphics suffered due to the aging Genesis' limited color depth).
The SNES, that's already been well-covered by another poster.
It has a poor reference card heatsink and heatsink attachment design, the heatsink likes to break its seal with the die on the GPU making the card overheat and lockup.
So did the 9800 Pro. So has EVERY midrange card. It comes with the pricepoint: manufacturers are trying to pack as much performance as they can into the cheapest card they can, because the midrange typically has the worst profit ratio. Some manufacturers cut corners on design, and you end up with cheaply-applied epoxy that breaks loose. Next time don't buy the CHEAPEST card.
But you know what? There are TONS of manufacturers who offer NON-REFERENCE HEATSINKS these days preinstalled for a little more money...like, for one instance, the card featured in this article.
The memory bandwidth is insufficient, the 6600GT has a 128-bit memory bus and a 1000MHZ memory clock. Consequently, there is not enough memory bandwidth to turn on antialiasing and anisotropic filtering.
You would THINK this if you were nothing but a spec whore. But if you actually looked at performance comparisons in games, the latest ATI 6600 GT killer (the x800 GT) actually has NO BETTER PERFORMANCE WITH AF / AA, even though it sports TWICE the memory bandwidth of the 6600 GT.
Did it ever occurr to you that AA actually takes additional GPU passes, in addition to the extra memory bandwidth? Did it ever occurr to you that higher memory bandwidth on cards that can't utilize it is a great but worthless selling point, just like putting 256MB ram on a 6200? The 6600 GT and x800 GT suck at 4x AA at higher resolutions because they lack the GPU horsepower.
But both are perfectly capable of providing a reasonable balance of AA / AF. With my 6600 GT, I ran HL2 at 1280x960 with 2xAA and 8xAF, or 1024x768 with 4xAA and 8xAF. As for multiplayer, where I need more fps, I run CS: Source at 1152x864, 2xAA, 8xAF, and Battlefield 2 at 1024x768, 2xAA. Both play over 60fps for me, even in heavy matches. This is not bad for a year-old card.
As for HDR, it's really only usable by the 7800 series. Even the 6800 GT and Ultra choke on it. But that's no surprise, really...most new features aren't that great the first time out (T&L on GeForce 256, for example).
I also forget to mention the wretched Vertex Shader performance of the 6600GT even when only Vertex Shader 2.0 is used
Again, we get back to the "usefulness" versus "selling point" argument. ATI has 6 vertex shaders on all their parts (even the ones that can't make use of them, except in unrealistic benchmarks), but the 6600 GT was specifically designed with three because it doesn't need more to fill those pipelines.
I'll bet you didn't know that the 6600 series actually only has 4 pixel rendering pipelines. The other 4 pipes are only put to work doing multi-pass rendering techniques, which is a given for all games released since 1999. You have probably not noticed this because it doesn't harm real-world gaming performance at all.
YES, the 6600 GT is a midrange card. NO, it can't perform like a high-end card. But YES, it is VERY competitive. Stop bashing it like it's some six-fingered man who killed you father.
You are exactly correct. Witness that we still have to choose between buying the widescreen and 'adjusted for your screen' version, whereas when I first started buying DVDs, they were frequently flipper disks with both screen formats.
Yes, there were a couple reasons early DVDs were mostly flipper discs, and most of them came with both versions.
Early on the industry had not perfected the dual-layer concept, so the dual-sided, single-layer disc was common. But the industry was smart - they knew that discs that required flipping in the middle of the movie would not only piss people off, but would remind people of Laserdisc.
So, with the exception of longer movies like Goodfellas, most double-sided, single-layer early discs cut the video bitrate a little and packed a whole movie onto each side. This also gave the industry the perfect "extra" feature to sell the new DVD format: widescreen and standard on the same disc.
Witness that DTS and Dolby Digital used to be two separate disks (and still are in some cases).
I didn't like those early dual-sided discs. Some movies that hit the 2-hour mark look like shit because the video bitrate is down in the 4Mbit range. It was a compromise.
Nowadays, the number of different versions available reflect the truth about DVD:
DVD was sold to the public on the concept that it could hold EVERYTHING, including multiple 9Mbit versions of the movie, Dolby Digital, DTS, extras, you name it. But the capacity to do this was truely only available in DVD18, which was expensive to make, and still required that annoying flipping by the user.
The fact is, DVD is most consumer-friendly in the single-sided, dual-layer package (and it's also convenient that single-sided, dual-layer discs are cheaper to make these days). Thus, you have 9GB to work with, which means ONE high-bitrate version of the movie with Dolby Digital in english and spanish, and maybe DTS or a few behind the scenes extras.
I mean, SURE studios invent reasons for re-releases, but the compartmentalization of features is mostly due to the limitations of DVD. Most people don't realize this, but if you crank up the DVD bitrate to maximum, you'll eat up almost an entire layer in an hour. We need more space...and hence, the new format.
This time around, the hybrid disc looks like a winner, because you don't have to cut any corners: dual layer DVD on one side, dual-layer HD-DVD on the other. Unfortunately, the movie industry will capitalize on this killer combination, and will probably charge 20-50% more for these new features (much like a special-edition DVD today).
My hope is that BD-ROM will be out around the same time period. It may cause confusion, but the format war has to be fought sooner-or-later. With two formats on the field, artificial upcharges will disappear.
I am an electrical engineer. I suck at picking up theoretical math. I'm not bad with the relationships once I absorb the concept through examples, but if you just show me a page full of equations, my mind takes a long time to absorb the information.
Convolution? Might as well shoot me.
So, how did I get my degree? Simple: I went to a college with instructors who cared. NOTE that I stressed the word college, as opposed to university.
My first-year classes were big, but my instructors were all PhDs in their resprective fields. Some were poor in the classroom, but none failed to communicate the concepts by example in their numerous office hours (with the exception of one nutso chem teacher). My professors did research, but they were mostly instructors first.
I had a very similar situation to the author of this article in that I had trouble adjusting to the load of engineering - my first year I barely scraped a 2.6. But because the system was available, open and friendly, I learned to use it. I only made below 3.0 one other semester, and by the time I graduated my average was a 2.9.
And no, the lack of dynamite theoretical math skills hasn't hurt my career. Makes me glad I took a chance with the small enginnering college instead of going to the big, well-known, research-oriented engineering university.
Yeah, my college had a major just for people like that: AB Engineering.
Reduces the technical requirements by about 1/3, and encourages a concentration in Gov/Law or Economics for the "AB" part of the degree.
Most of the people who graduated AB Engineering were EE/ME/CE/ChemE burnouts who weren't smart enough to jump ship by sophomore year. So long as you made the jump by the start of your senior year, you could graduate with your class.
Right. The drop-off rate is pretty bad, and it really sucks to be stuck in your third year and suddently come to the conclusion that you hate all the math.
My college had a major designed specifically for this situation called AB Engineering. As you might expect, the degree reduced the number of required upper enginnering classes, and replaced them with government / law / economics / business. Basically, it was a one-way ticket into engineering management, but it was better than throwing away those lovely credits.
I had a friend who, at the start of his senior year, decided he didn't like EE anymore. He switched to AB Engineering and graduated with his class. I don't think he went off to do anything special with that degree, but at least he didn't have to start from scratch.
In addition, I don't see what everyone is getting so excited over.
Let's say they do fully crack firmware 2.0...no big deal. Sony will then launch a knockout punch: GTA: Liberty City Stories automatically installs the brand-new, patched 2.10 firmware.
Crack the 2.10 firmware? No problem, essential games like Madden 2007 and Lumines 2: "The Bloodening" will come with the spiffy new 2.50 firmware.
If you ever intended to play official games at all, emulation and homebrew on the PSP will be a losing battle. This isn't like hacking the Xbox, where hardware versions changed only occasionally...Sony holds all the cards here because the vast majority of PSP owners want to play games.
If you really want an open, hackable platform, there are many to choose from...but they don't come with Sony's games. CHOOSE ONE OR THE OTHER, damnit.
That's weird, I'm VERY concious of the refresh rate (can tell the difference between 60, 72, 75, 85 easily), but even I have to admit that 85Hz is plenty fast enough on any decent CRT.
Oh well, at least your monitor was born to run at such high speeds (also a proud 454 owner).
No. LCD pixels aren't just "more persistent" than CRT pixels.
LCD pixels hold their current state until the input signal changes. There is no scan period on an LCD, the 60Hz signal is simply a convenient way to bridge the gap between raster scan displays and active-matrix displays.
So, if you send an LCD a set of successive white screens, after the initial white screen no pixels will change, ever. A CRT, on the other hand, will write a white pixel to every part of the screen once every 1/60th of a second...and while the beam is not concentrated on a particular pixel, its brightness will fade.
It has nothing to do with LCDs' slower reponse time, as you implied.
Also, it would have a great HTPC, if Apple opened APIs for the MPEG2 acceleration hardware onboard. With that, it would be capable of HDTV video playback. Without that acceleration, it's not fast enough to keep up with 1080i video.
Huh? Any onboard MPEG2 acceleration on the Mac Mini comes from the Radeon 9200 chip, which has hardware iDCT and motion compensation. I thought the specifications on the 9200 were well known in the open-source community, so what's stopping them from accessing it directly?
Slashdot Mirror
The only problem with this is Intel's SpeedStep implementation in the P4 / Xeon isn't aggressive. The best you'll see a SpeedStep-enabled processor drop down to is 2.8 GHz at 1.2v. this gives you a maximum %20 power reduction from frequency, and a maximum %25 power reduction from dropping the voltage from 1.4v to 1.2v, for a total potential drop of 45%.
This is in contrast to all Cool 'n Quiet-enabled Athlon 64 processors, which drop to 1GHz @ 1.1v. This provides plenty of processing power to do mundane tasks, as well as deliver snappy performance to servers at low usage levels.
The agressive reduction in frequency and voltage gives you a processor that sips 3w at idle, and less than 10w at light load (1.0 GHz).
IIRC both AMD and Intel are realising what VIA and Transmeta did years ago--that the clock-speed race is completely pointless now. Not only do both companies invest heavily in mobile technologies--that technology is finding its way into desktops and servers as well. About bloody time I think. It'd be great if this turns into a golf-score style contest of watts-per-flop race.
VIA didn't realize that the clock speed race was pointless. VIA went "low-power" out of sheer necessity.
The Winchip architecture was never intended as a power-house, and unfortunately VIA didn't get their hands on it until it had languished for a few years (and Intel and AMD had pulled far ahead). By that time, they could nave catch up in terms of performance per clock, or performance per watt.
You seem to forget that the VIA Epia never really had ANY performance per watt lead, and it never had a significant lead on other low-power x86 solutions. For example, the P3 Tualatin LV, a CPU released around the SAME TIME as the Epia platform, ran as fast as 800 MHz and consumed 7w max. Despite having the same limited memory bandwidth as the Epia platform, it performed twice as fast clock-for-clock.
Today, we have LV Pentium M derivatives that blow the Epia out of the water. They're just not as popular as the Epia platform because the chips and boards aren't cheap.
In addition, people just aren't as willing to look past pre-built solutions to find a low-power platform. Take a look at AMD's DESKTOP processors. Every single one of them offers the ability to run at 1.0GHz at 1.1v (Cool 'n Quiet mode). It can been shown that an AMD 64 DESKTOP part running at 1.0GHz and 1.1v produces 10w under full load, and around 3w idle...VERY competitive with VIA's specs.
Basically, VIA is doing everything they can to carve a niche for a CPU that really should have been beaten into the ground long ago. The performance is pathetic, so they leverage the "embedded, low-power x86 compatibility" angle and try to make ends meet. The thing is, the market for the Epia is so tiny relatively that not even AMD feels it is worth tackling.
But just because Intel and AMD don't want to tackle the tiny "x86 low-power kiosks and hobby platforms" market, that doesn't mean we'll have to forgo low power. My desktop AMD system already uses the same power at idle as your anemic Epia system, and costs less. Once Intel starts offering Conroe, they'll have the same.
When it comes to CRTs, Sony was the best only three years ago. When Mitsubishi started pushing their "DiamondTron" flat-CRTs, totally ripped off from Sony's Trinitron technology.
You act like Mitsubishi was stealing candy from a baby. You also act as if competition is a bad thing.
Sony had their chance at the Trinitron. They've been making it since 1969, so they've gotten more than their money's worth out of the technology.
Normally such a market lead as Sony had with such a sophisticated product would preclude other companies entering the market, for fear of never catching up. But Sony was complacent, never offering significant improvements on the Trinitron beyond the 70s, and always charging a premium. When Mitsubishi saw an opportunity in the market in the 1990s, they caught Sony with their pants down.
Mitsubishi was the first Aperture Grille manufacturer to produce a completely flat, distortion-free tube (marketed as Diamondtron NF), and they pushed Sony to release their own version (FD Trinitron) in response. Who knows how long we would have had to wait for Sony to do it without competition, especially with the explosion of LCDs looming on the horizon?
The best part is, Mitsubishi has worked to reduce the price of their Dirmondtron NF line, effectively removing the premium long associated with the Trinitron. The end result is, you can still buy a quality Aperture Grille tube, even though Sony can't seem to make theirs cost-effective. Sony no longer matters.
* Proud user of the Diamondtron NF since 1999.
I've been reading through the discussion, and I've been thinking of responses, but it's all a muddied mess out there. so, I've decided to lay out the basic discussion points and my thoughts as one post.
First of all, why do we need faster-response LCD screens, when we already have 4ms?
There are a few key reasons for this. For starters, the 4ms number doesn't mean much. It is the time the panel takes to turn a pixel from black to white, then back to black. In a traditional panel, this is usually the fastest transition possible...and all other tranitions (Grey to Grey) are MUCH slower. Sometimes GTG transitions can be as much as 3x slower than the Black-White-Black number.
The industry has concocted a possible solution to this called Overdrive.
Overdrive takes advantage of the fast transition in Black-White-Black. Every time an input pixel changes color, the pixel on-screen is bootsted up to white, and allowd to fall back down to the new color.
This is slightly slower than the Black-White-Black transition time, but it's much faster than going Grey-to-Grey.
Unfortunately, Overdrive has a drawback that is DIRECTLY tied to the response time. Every time a pixel changes, it is overdriven WHITE for a fraction of a second, until it settles down to the target color. In darker scenes, or in cases where where colors are almost uniform, as pixels change these white pixels are painfully obvious. Better response times are the only thing that can remove this annoying artifacting.
Read about these artifacts at Tom's, who did the first review ever on Overdrive panels in May.
This link to Tom's also addresses the other issue discussed in this thread:
What's wrong with 18-bit color?
The dithering algorithms used by panels to simulate 24-bit color are not all that bad, but they have a serious drawback:
Dithering yields poor quality in scenes which require high contrast. Foggy, smoky or dark scenes, which tend to have subtlte color transitions, look like crap on an 18-bit panel. The panel is constantly changing pixels that are VERY close to each other in color, resulting in a muddy image. Unfortunately, the only way to avoid such artifacts it to buy an MVA panel with true 24-bit color (and sacrifice response time).
Yeah, I about laughed out loud when I read that.
:D
This time in the year 2000, ATI was releasing the Radeon.
Yes, THE ORIGINAL Radeon (later named 7500). Two pipes, 3 textures per pipe and a T&L unit.
Come on people, get your ATI release dates straight
2000: Radeon SDR/DDR
2001: Radeon 8500 (later reworked as 9000/9100/9200)
2002: Radeon 9500 series, 9700 series
2003: Radeon 9600 series, Radeon 9800 series
2004: Radeon x300/x600/x700/x800 series
2005: Radeon x1300/x1600/x1800 series
AMD64 / EMT64 / x86-64 is NOT PROPRIETARY.
It was developed as an open standard to ensure industry-wide acceptance. Intel does not need a license to develop and sell an x86-64-compatible chip, and neither do you.
According to Marc Miller of AMD: x86-64, now named amd64, was developed as an open standard, therefore it's not something we can license.
The open standard is x86-64. AMD's trademarked implementation is called AMD64. Intel's trademarked implementation is called EMT64.
"Wherever you go, there you are!"
Except that it's not the same thing.
- Divx, even the unlocked discs, were not playable on DVD players (nor could you take unlocked discs to another divx player and play them for free) . These discs, on the other have, have zero restrictions.
- Divx still required you to drive to the store and put down a whopping $5 for the first 48-hr rental (with typical unlocking fees of $20-25). This has much lower total fees, and does not require you to leave the house.
Honestly, if they allowed me to pay regular PPV price, then make my decision to buy after the movie shows, I would actually try this service out. But knowing Comcast, they'll try to make you buy the whole thing up-front.
what after you have seen it
you see the whole movie and than you get the hard copy. hey $17 isnt it bit costly . so why do you need the hard copy. is it for the video library or what. huh.
Ahh, I've broken his secret alien code! If you carefully re-arrange the words, you get this:
Huh? What you have isn't the whole costly movie. You see what the hard copy is after the video library seen it.
So, you get $17, so why do you need the hard copy?
Hey, is it for or and what than?
See, it all makes sense now! The ton foil hats, the satellite dishes, the voices in my head!
Yeah. The article talks about Kahuna featuring improved "basic" functionality. How about this for "basic":
Outlook allows you to open multiple emails at once, so that I can parse through them quickly. Hotmail, on the other hand, makes you click-through to read each email (and each time you do this, the inbox disappears). In addition, the Hotmail UI prevents the opening of emails into their own tabs / windows, so I can't open all the emails I'm interested in at once.
If they're going to work on the "basics," and have a rich applications framework, I sure hope they're planning on providing "basics" available to Outlook Express users for the last decade. I'd actually READ my Hotmail inbox more than twice a week if it were as easy as reading my Outlook inbox.
* Patiently waiting for the next generation in web email since I signed up for my Hotmail account in 1997.
While I certainly don't agree with the outright gifting of cars, there's no reason cars can't be loaned to kids if the circumstances are right.
I put myself through college mostly on my own dime. I worked my ass off in high school to get good grades, which turned into merit + need based financial aid. I worked full-time summer, and part-time during the year.
For college, I took my high school economics teacher's advice: that a car was a money pit. I didn't need a car living so close to campus, so for the first two years I didn't have one. But I did need a car to move home during the summer and winter breaks, and my parents got tired of that.
When they bought a new car, instead of trading in their old 91 Corolla, they loaned it to me. Since I didn't own it, they stipulated that they would pay the insurance if I handled the upkeep, and transported myself. They figured it would solve their problem while giving me more freedom.
(And my economics teacher was right. That car was STILL a money pit)
It was a good compromise. They didn't come back to my campus until graduation, and shortly thereafter got my own car. The Corolla was later loaned to my sister.
As another example, I don't think the modern game hardware industry owes much to Silicon Graphics - their hardware was ahead of PC hardware for a long time, but at a price most people wouldn't pay. When the time was right, graphics hardware became widespread, but it was no thanks to SGI who were trying to maintain the old prices.
Are you kidding me? The modern game hardware industry owes a lot to SGI.
First, let's talk about SGI's direct effect on the industry:
SGI designed the N64. It was basically all the best bits of a $10,000 Indy workstation, shrunk down and sold for $250. Sure, it was a little memory-deprived, but that was a result of the pricepoint. Its 3D prowess was unmatched by anything consumer-level until the Voodoo Graphics was released for the PC a year later (which still cost more, intro-ed at $300), and was unchallenged in the console arena until the Dreamcast.
SGI's '$250 Indy' featured hardware support for mip-mapping, billinear filering, anti-aliasing...tons of features typically only seen on SGI workstations.
Now, let's talk about the indirect effects SGI had on the industry.
SGI developed OpenGL. This had the following indirect, but lasting effects:
* 3dfx based their GLide API on a subset of OpenGL, removing some of the professional-only instructions, as well as the transform and lighting hardware support, in order to implement a cheap 3D pipeline. 3dfx was the most influential hardware company in the early consumer 3D industry.
* OpenGL made Microsoft react and make Direct3D a solid development platform, and encouraged MS to innovate even when they surpassed OpenGL. The ability of OpenGL to use unofficial extensions kept Microsoft on their toes. SGI's industry clout was key in getting OpenGL support built into Windows 95b / 98.
SGI lead the way in making 3D a commodity, but once 3D became a commodity they lost the reins.
But in REALITY, we know that there are white-hat, and black-hat hackers.
Once the community finds the vunerability, it is free game for the white-hats (firmware downgrader) and black-hats (firmware toaster). You should feel lucky that a remote exploit hasn't been found (yet).
Sony has provided a fix for the vunerability, because they have TWO significant financial interests. Most of you think there's only one (locking down the system to enforce licensing fees from developers), but there is another very serious one: keeping the system secure to avoid lost sales and a class-action lawsuit once users are left holding a bunch of hacked-up bricks.
Since the console is wireless, security is extrodinarily important.
No,
They simply read Robin Cook's "Contagion", which was released back in 1996.
In the story (minor spoiler),
A hospital has recurring internal infection problems with exotic disesases, including plague and Spanish flu.
(major spoiler)
Certain parties that enjoyed collecting and cultivating viruses come across specimens in the yukon permafrost frozen to death, infected with Spanish flu. These parties eventually become involved in a plot to discredit a hospital by planting viruses to cause internal infections, and use the Spanish flu to cause panic (and unknowlingly threaten an epidemic).
And thats the lowest that proc will ever be, though it may be replaced with a slightly better one in that tiers slot?
Yup, you got it.
The XP filled in for about two years as AMD's value line. Now that XP production has stopped, AMD has retired the Socket A platform. The Sempron (Athlon 64 on socket 754 with less cache) is the new budget processor.
The Sempron fills the $140 and less market, and the Athlon 64 socket 939 is priced above that. Due to the fact that AMD holds the performance crown, this will not change anytime soon unless Intel starts a price war. AMD is already having trouble meeting demand with their current prices, so they have no reason to lower them.
EVERY console that has been worth its salt has rivaled the arcades on release.
The Atari VCS was incredible when it came out, capable of quite accurate renditions of hits like Space Invaders and Bezerk. Sure, by the time Pac Man was popular (4 years later), it was faltering, but that's what happens.
The NES produced surprisingly accurate versions of arcade hits like Ikari Warriors, Super Dodge Ball and Double Dragon, just to name a few.
The Sega Genesis was BASED on Sega's System 16 arcade board, which was the platform for great games like Altered Beast, Alien Syndrome, Golden Axe and Shinobi (link has full games list). Even the later Street Fighter II and Mortal Kombat ports were quite complete (but the graphics suffered due to the aging Genesis' limited color depth).
The SNES, that's already been well-covered by another poster.
It has a poor reference card heatsink and heatsink attachment design, the heatsink likes to break its seal with the die on the GPU making the card overheat and lockup.
So did the 9800 Pro. So has EVERY midrange card. It comes with the pricepoint: manufacturers are trying to pack as much performance as they can into the cheapest card they can, because the midrange typically has the worst profit ratio. Some manufacturers cut corners on design, and you end up with cheaply-applied epoxy that breaks loose. Next time don't buy the CHEAPEST card.
But you know what? There are TONS of manufacturers who offer NON-REFERENCE HEATSINKS these days preinstalled for a little more money...like, for one instance, the card featured in this article.
The memory bandwidth is insufficient, the 6600GT has a 128-bit memory bus and a 1000MHZ memory clock. Consequently, there is not enough memory bandwidth to turn on antialiasing and anisotropic filtering.
You would THINK this if you were nothing but a spec whore. But if you actually looked at performance comparisons in games, the latest ATI 6600 GT killer (the x800 GT) actually has NO BETTER PERFORMANCE WITH AF / AA, even though it sports TWICE the memory bandwidth of the 6600 GT.
Did it ever occurr to you that AA actually takes additional GPU passes, in addition to the extra memory bandwidth? Did it ever occurr to you that higher memory bandwidth on cards that can't utilize it is a great but worthless selling point, just like putting 256MB ram on a 6200? The 6600 GT and x800 GT suck at 4x AA at higher resolutions because they lack the GPU horsepower.
But both are perfectly capable of providing a reasonable balance of AA / AF. With my 6600 GT, I ran HL2 at 1280x960 with 2xAA and 8xAF, or 1024x768 with 4xAA and 8xAF. As for multiplayer, where I need more fps, I run CS: Source at 1152x864, 2xAA, 8xAF, and Battlefield 2 at 1024x768, 2xAA. Both play over 60fps for me, even in heavy matches. This is not bad for a year-old card.
As for HDR, it's really only usable by the 7800 series. Even the 6800 GT and Ultra choke on it. But that's no surprise, really...most new features aren't that great the first time out (T&L on GeForce 256, for example).
I also forget to mention the wretched Vertex Shader performance of the 6600GT even when only Vertex Shader 2.0 is used
Again, we get back to the "usefulness" versus "selling point" argument. ATI has 6 vertex shaders on all their parts (even the ones that can't make use of them, except in unrealistic benchmarks), but the 6600 GT was specifically designed with three because it doesn't need more to fill those pipelines.
I'll bet you didn't know that the 6600 series actually only has 4 pixel rendering pipelines. The other 4 pipes are only put to work doing multi-pass rendering techniques, which is a given for all games released since 1999. You have probably not noticed this because it doesn't harm real-world gaming performance at all.
YES, the 6600 GT is a midrange card. NO, it can't perform like a high-end card. But YES, it is VERY competitive. Stop bashing it like it's some six-fingered man who killed you father.
You are exactly correct. Witness that we still have to choose between buying the widescreen and 'adjusted for your screen' version, whereas when I first started buying DVDs, they were frequently flipper disks with both screen formats.
Yes, there were a couple reasons early DVDs were mostly flipper discs, and most of them came with both versions.
Early on the industry had not perfected the dual-layer concept, so the dual-sided, single-layer disc was common. But the industry was smart - they knew that discs that required flipping in the middle of the movie would not only piss people off, but would remind people of Laserdisc.
So, with the exception of longer movies like Goodfellas, most double-sided, single-layer early discs cut the video bitrate a little and packed a whole movie onto each side. This also gave the industry the perfect "extra" feature to sell the new DVD format: widescreen and standard on the same disc.
Witness that DTS and Dolby Digital used to be two separate disks (and still are in some cases).
I didn't like those early dual-sided discs. Some movies that hit the 2-hour mark look like shit because the video bitrate is down in the 4Mbit range. It was a compromise.
Nowadays, the number of different versions available reflect the truth about DVD:
DVD was sold to the public on the concept that it could hold EVERYTHING, including multiple 9Mbit versions of the movie, Dolby Digital, DTS, extras, you name it. But the capacity to do this was truely only available in DVD18, which was expensive to make, and still required that annoying flipping by the user.
The fact is, DVD is most consumer-friendly in the single-sided, dual-layer package (and it's also convenient that single-sided, dual-layer discs are cheaper to make these days). Thus, you have 9GB to work with, which means ONE high-bitrate version of the movie with Dolby Digital in english and spanish, and maybe DTS or a few behind the scenes extras.
I mean, SURE studios invent reasons for re-releases, but the compartmentalization of features is mostly due to the limitations of DVD. Most people don't realize this, but if you crank up the DVD bitrate to maximum, you'll eat up almost an entire layer in an hour. We need more space...and hence, the new format.
This time around, the hybrid disc looks like a winner, because you don't have to cut any corners: dual layer DVD on one side, dual-layer HD-DVD on the other. Unfortunately, the movie industry will capitalize on this killer combination, and will probably charge 20-50% more for these new features (much like a special-edition DVD today).
My hope is that BD-ROM will be out around the same time period. It may cause confusion, but the format war has to be fought sooner-or-later. With two formats on the field, artificial upcharges will disappear.
I disagree with your analysis.
I am an electrical engineer. I suck at picking up theoretical math. I'm not bad with the relationships once I absorb the concept through examples, but if you just show me a page full of equations, my mind takes a long time to absorb the information.
Convolution? Might as well shoot me.
So, how did I get my degree? Simple: I went to a college with instructors who cared. NOTE that I stressed the word college, as opposed to university.
My first-year classes were big, but my instructors were all PhDs in their resprective fields. Some were poor in the classroom, but none failed to communicate the concepts by example in their numerous office hours (with the exception of one nutso chem teacher). My professors did research, but they were mostly instructors first.
I had a very similar situation to the author of this article in that I had trouble adjusting to the load of engineering - my first year I barely scraped a 2.6. But because the system was available, open and friendly, I learned to use it. I only made below 3.0 one other semester, and by the time I graduated my average was a 2.9.
And no, the lack of dynamite theoretical math skills hasn't hurt my career. Makes me glad I took a chance with the small enginnering college instead of going to the big, well-known, research-oriented engineering university.
Yeah, my college had a major just for people like that: AB Engineering.
Reduces the technical requirements by about 1/3, and encourages a concentration in Gov/Law or Economics for the "AB" part of the degree.
Most of the people who graduated AB Engineering were EE/ME/CE/ChemE burnouts who weren't smart enough to jump ship by sophomore year. So long as you made the jump by the start of your senior year, you could graduate with your class.
Right. The drop-off rate is pretty bad, and it really sucks to be stuck in your third year and suddently come to the conclusion that you hate all the math.
My college had a major designed specifically for this situation called AB Engineering. As you might expect, the degree reduced the number of required upper enginnering classes, and replaced them with government / law / economics / business. Basically, it was a one-way ticket into engineering management, but it was better than throwing away those lovely credits.
I had a friend who, at the start of his senior year, decided he didn't like EE anymore. He switched to AB Engineering and graduated with his class. I don't think he went off to do anything special with that degree, but at least he didn't have to start from scratch.
In addition, I don't see what everyone is getting so excited over.
Let's say they do fully crack firmware 2.0...no big deal. Sony will then launch a knockout punch: GTA: Liberty City Stories automatically installs the brand-new, patched 2.10 firmware.
Crack the 2.10 firmware? No problem, essential games like Madden 2007 and Lumines 2: "The Bloodening" will come with the spiffy new 2.50 firmware.
If you ever intended to play official games at all, emulation and homebrew on the PSP will be a losing battle. This isn't like hacking the Xbox, where hardware versions changed only occasionally...Sony holds all the cards here because the vast majority of PSP owners want to play games.
If you really want an open, hackable platform, there are many to choose from...but they don't come with Sony's games. CHOOSE ONE OR THE OTHER, damnit.
That's weird, I'm VERY concious of the refresh rate (can tell the difference between 60, 72, 75, 85 easily), but even I have to admit that 85Hz is plenty fast enough on any decent CRT.
Oh well, at least your monitor was born to run at such high speeds (also a proud 454 owner).
No. LCD pixels aren't just "more persistent" than CRT pixels.
LCD pixels hold their current state until the input signal changes. There is no scan period on an LCD, the 60Hz signal is simply a convenient way to bridge the gap between raster scan displays and active-matrix displays.
So, if you send an LCD a set of successive white screens, after the initial white screen no pixels will change, ever. A CRT, on the other hand, will write a white pixel to every part of the screen once every 1/60th of a second...and while the beam is not concentrated on a particular pixel, its brightness will fade.
It has nothing to do with LCDs' slower reponse time, as you implied.
Also, it would have a great HTPC, if Apple opened APIs for the MPEG2 acceleration hardware onboard. With that, it would be capable of HDTV video playback. Without that acceleration, it's not fast enough to keep up with 1080i video.
Huh? Any onboard MPEG2 acceleration on the Mac Mini comes from the Radeon 9200 chip, which has hardware iDCT and motion compensation. I thought the specifications on the 9200 were well known in the open-source community, so what's stopping them from accessing it directly?