I’ve noticed one thing on the Internet, that stands out above almost all others: most people on the Internet have no clue what they are talking about. Case in point, a lot of ricers and gamerz like to say that DDR is lower latency than DDR2 because DDR2 takes more cycles to do things; except they forget one important thing: cycles are not a measurement of time, they are a measurement of iterations.
That said, there is only one case where DDR actually manages to be lower latency than DDR2 (and this doesn’t mean it has higher performance, or effects benchmarks in any measurable way in favor of DDR), and that is with DDR400 memory vs. DDR2-400 memory: latency is theoretically lower, but you pay a penalty for giving up DDR2’s larger prefetch buffer and better power efficiency. Also, no one actually uses DDR2-400 memory, only 667 and 800. DD2-800 compared to DDR400, latency ends up being similar in impact, and the actual performance is at least twice as much as DDR400, probably even more.
Another thing people say is that DDR2 is slower because it takes more cycles to do things. Yet another thought that hasn’t been fully thought out, and in a similar manner to the whole latency problem (infact, they are directly related; faster timings usually decrease latency across the same memory archetecture). As I said earlier, cycles do not measure time; however, cycles combined with cycles per unit of time measure time. DDR2 in most, if not all, situtations simply performs better.
So, to anyone out there that says that DDR2 is a step backwards: You’re an idiot.
How many people out there know how to properly describe a “DSL modem”? DSL modems do not modulate/demodulate like analog phone modems, they simply bridge some form of Ethernet or ATM over unused frequencies on POTS lines and unused ATM lines. They are simply bridges.
Modern DSL modems also act as single port switches, and deal with the PPP tunnel and whatnot now… and I have came up with an appropriate name for these: a swidge. DSL modems with built in firewall/PPP/switch capabilities are swidges, those four port home LAN switches that can run OpenWRT are swidges, and anything that bridges you to another network, and switches packets, is a swidge.
So, what do we call a swidge that is also a WiFi access point? A Wiswidge.
Caution: This math may be wrong.
In episode 6 of Neon Genesis Evangelion, NERV borrows a positron cannon from a government research lab to deal with the latest threat. This cannon takes all of Japan’s power output, for 37 seconds, to charge the rifle for one shot. According to the CIA World Factbook, Japan produced 1.017 trillion kWh for the whole year in 2003.
1.017 trillion kWh per year/365 days/24 hours = 116 million kWh per hour
So, Japan produces 116 million kWh per hour. However, Makoto Hyuga states that it will take atleast 180 million watts to pierce the AT field of the angel Ramiel. Seeing as it takes 37 seconds to charge the rifle*…
180 million kW per 37 seconds * (3600 seconds / 37 seconds) = 17460 million kWh per hour
So, if my calculations are correct, the rifle would need 17460 million kWh to charge for an hour; unfortunately, this is quite a lot higher than the 116 million kWh per hour Japan produces, infact, it is 150 times more. So, if Evangelion’s Japan in the year 2015 is to be able to power the rifle, it needs to produce 150 times more power than 2003 Japan does.
* Counting from the ejection of the first shot’s fuse, to the pulling of the trigger for the second shot.
And now part 2 of my lengthy rant. Read part 1 here.
Oh Cassini, help us!
The mention of the Sega Saturn makes me cringe. I mean, Sega finally produced a new console, and everyone wanted to love it. Hell, I owned one, and I loved it; but Sega decided to kill it for me. In 1994, Sega released the Saturn, and in 1997 they quit production of it.
Internally, the Saturn reminds me a lot of the 32x: Dual 28MHz Hitachi SH2 processors and a pair of video processing units modeled after the video processing unit in Sega’s System 24 and System 32 arcade systems. The memory included 1MB fast and 1MB slow system memory, two sets of 256KB for the framebuffer, two sets of 512KB for use of each of the video processing units, 512KB for sound sample caching, and a larger CD cache than what was in the SegaCD.
So, that doesn’t sound like a bad system, right? Well, for one, the pair of SH2s is slower than what was in the 32x, though the faster memory and faster main bus more than makes up for it; the 32x was held back by the slow memory and main bus. Main system memory in the Saturn was eight times what it was in the 32x, and could be expanded using external expansion cards (which very few games used).
Also, the sound hardware is setup using a Motorola 68k paired with a Yamaha FH1 DSP. Compared to the Genesis, the 68k and FH1 pairing is far more powerful than the Genesis’ Z80 and Yamaha YM2612, plus it can do anything the 32x and SegaCD sound hardware can do, including PCM sample playback and many channels of complex music synth and sound effects.
So, ultimately, the Saturn is sorta like if they took the 32x and SegaCD, the 3D hardware from their System32 arcade system, and a more powerful sound setup than anything seen before, and combined it all together to produce a new console.
Hardware wise, it beat the pants of Sony’s Playstation and Nintendo’s N64, but looking back at history, we know hardware does not make a successful platform. Sega failed with the Saturn in five major ways:
- Sony gave the Software Development Kit (SDK) to all third party developers, but Sega only used it internally withholding it from all outsiders.
- The Saturn ‘s opening day price was over US$400, vs the Playstation’s US$300.
- The Saturn suffered from development problems the same way the 32x did, few developers knew how to efficiently use dual processors in a Slave/Master configuration (unlike modern PCs which use a symmetric multi-processing (SMP) setup).
- Sega promised that they’d port many Japanese games to English, especially all the popular titles.
- The substandard construction and poor support of the SegaCD and 32x bred distrust in the Sega fandom, and many people converted to Nintendo and Sony, never to look back.
But that isn’t what put the final nail in Saturn’s coffin: The then CEO of Sega of America decided to not support the Saturn, and refuse to port games here. I mean, not just unpopular ones, but any. This violated Sega’s promise to the few fans who were left that they’d bring many games here to America.
This prompted Sega of Japan to evict the CEO of SoA and install a new one that would actually follow the wishes of SoJ; however, this came too late, and the Saturn was quite dead in the American market by then, even though the Saturn was selling well in the Japanese market. At this point, Sega started work on the Sega Dreamcast, the most powerful and final console of Sega’s empire.
As the American market fell apart, this caused the cancellation of the Saturn’s installment of the Sonic series, Sonic X-treme, and they also decided not to bring Panzer Dragoon Saga to America, which is arguably one of the best games ever produced for the Saturn and one of the best games ever to hit a Sega console.
The Saturn died, and in it’s place, Sega offered the Dreamcast.
ドリームキャスト: It’s thinking
The American release of the Dreamcast was September 9th, 1999, or as the marketing put it, 9/9/99. Five years ago, in Janurary 2001, the Dreamcast marked the end of Sega as we knew it, Sega soon afterwards became just a software vendor and merged with Sammy, another well known vendor of games.
Even though the Dreamcast was Sega’s last console to be produced, it cannot be blamed for the downfall of Sega. The Dreamcast is considered Sega’s best effort to remain in the market, and a threat to Nintendo and the newcomer Sony. What really killed Sega was the failure of the 32x and the SegaCD, which in some sort of chain reaction, took the Saturn out as well.
That said, the Dreamcast did well in the market. On opening day in America, the Dreamcast sold over 225,000 machines, with 200,000 of them being pre-ordered, and over 500,000 of them being sold in the first week; Sega made $98.4 million in profit with the 9/9/99 launch. These records were held until Sony released the Playstation 2 in late 2000.
The Dreamcast was well received, and clearly was defeating it’s competition, Nintendo’s N64 and Sony’s Playstation, but Sony in turn defeated the Dreamcast with the Playstation 2. Once again, Sega produced a superior platform (the only thing missing on the Dreamcast that the Playstation 2 did have was a DVD drive).
The Dreamcast was basically the Saturn done correctly. Powered by a 206Mhz Hitachi SH4 (which was far more powerful than the aggregate power of the Saturn’s dual SH2) with a built-in special-use computational engine; a PowerVR2 CLX2 (ie, a real 3D engine) to handle all the 3D rendering work; and a 47MHz ARM7 with a customized Yamaha DSP to do sound, the Dreamcast was very powerful.
It also had a fair bit of system memory at a total of 16MB, eight times what the Saturn had. The Dreamcast’s 8MB of video memory (2.5x of the Saturn) and 2MB of sound memory (4x of the the Saturn) wasn’t something to sneeze at either.
Sega didn’t repeat any mistakes on this one, it had a single processor that didn’t require any special (read as: convoluted) programming, it could do modern 3D processing comparable to a PC (only the XBox, Gamecube, and Playstation 3 share that distinction with the Dreamcast), and it had Sonic Adventure ready for release day. Everything went perfectly.
Yup, you heard me, I can’t say Sega did anything wrong with the Dreamcast. It even came with a built in Internet gameplay device, something no other platform before ever did, compared to the Playstation 2 that didn’t even have useful online gameplay until 2002.
The Dreamcast isn’t even really dead, even though Sega no longer produces or supports them; many third party publishers continue to produce and release games, some even sold in Japanese retail stores. Even now, Radilgy was just released for the Dreamcast (which Sega is now selling refurbished Dreamcasts for), and Under Defeat is being released real soon now.
The Dreamcast may be the most perfect console ever produced.
Conclusion
Ultimately, Sega failed because they tripped up on two add-on devices for the Genesis. The Genesis was a great console, the Saturn wasn’t a bad console, and the Dreamcast is one of the best consoles ever produced.
This stands as a lesson to future game companies: if your target market doesn’t trust you, or you’ve previously abused their trust, or simply if you confuse them by releasing multiple similar devices, they might find someone else to buy things from, and won’t come back to you when you figure out you’ve done wrong.
However, the Sega of the old days will always be remembered as they left their mark on video game history, and even now they still continue to produce games, and the Sonic series of games has found itself a new home on Nintendo’s Gamecube (with Sonic Riders soon to be released), Gameboy Advance (featuring the Sonic Advance series of games), and DS (with Sonic Rush).
I played, and finished, Sonic CD for the first time today. It wasn’t that bad of a game; it was no Sonic 3 and Knuckles, but it was about on par with Sonic 2. Unfortunately, even though the CD audio background music and the small number of full motion video was cool, it did little to enhance the gameplay.
While playing it, I thought about how Sega failed not only itself, but all Sega fans everywhere. I’ve decided to compile my thoughts, so here’s a warning, this entry is quite long, and may take awhile to read; it’s also full of technical terms, and delves into the technical side of game platform development.
Also, I’ve decided to split this into two halves, part one is about the Genesis, SegaCD, 32x, and the Neptune; part two will be about the Saturn, and Dreamcast.
In the beginning…
In 1988, Sega released the famous sequel to their lackluster Sega Master System: the Sega Megadrive, or as we call it here in America, the Sega Genesis. This console was Sega’s foray into the world of 16-bit game consoles, and they managed to make a successful platform out of it.
Between the Sonic the Hedgehog series of games, and smart marketing, it put Sega on the map. Sega continued to be widely recognized as a console manufacturer until they quit the console business with the utter and complete failure with the Saturn, and then the Dreamcast.
The hardware inside the console was very powerful; between the fast 8Mhz Motorola 68k, the giant helping of 64KB system and 64KB video memory, the well designed custom video processing unit, the Genesis was very powerful. (Remember, this was 1988, people were still using 386s and running DOS.)
The best selling game, Sonic 2, sold millions of copies via one of the best marketing schemes, Sega’s famous ‘Sonic 2uesday’. You can’t hope for anything better than this. So, you, my faithful readers, are probably asking yourselves, “Wow, Sega rocks! What could possibly have gone wrong?” Everything.
The Genesis kept selling well, and both Sonic 3 and Sonic and Knuckles also sold well. Sega knew that Nintendo’s Super Nintendo Entertainment System (SNES) managed to make up the gap in popularity, and was beating the Genesis in sales and popularity, so Sega decided to make something new; actually, they ended up making two new things, the SegaCD and Sega 32x add-ons to the Genesis.
“150KB/sec is enough for anyone!”
Oh boy, what the hell was Sega thinking. Its like after the creation of the Genesis they decided to turn off their brains, and just produce crap for the next ten years. The SegaCD was simply an add-on 1x CD drive for the Genesis, and the Genesis plugged into it. It was released in 1991.
Inside the SegaCD was a faster 12.5Mhz 68k, a second audio processor (with 64KB of RAM dedicated to it) to mix high quality stereo PCM and CD audio, which was combined with the Genesis’s own audio processor to do traditional Genesis audio work as well as enhanced PCM sample playback, an additional 768KB of system memory, and another video processing unit that would add allow the Genesis to do effects similar to the SNES’s Mode 7 (however, don’t expect something like Mario Kart, it wasn’t powerful enough) and to decode full motion video (which it wasn’t very good at).
But as I said earlier, the CD drive itself was a mere 1x, and many games suffered from severe lag from loading data from the CD; unlike cart-based games where nothing needed loaded, and could be directly accessed/executed. Particularly, Mortal Kombat CD would pause for several seconds when doing complex moves, because it needed to load data off the CD.
Because the CD drive was so slow, most games decided just to have a few full motion video sequences, some CD audio background music, and fit everything else they needed in the 768KB of system RAM. In comparison, Genesis games couldn’t do CD audio/PCM sample playback or full motion video, but they had up to 4096KB to store anything they needed in the cart itself, and in my opinion, the trade-off was worth it.
Between that, and the release day selling price of Â¥49,800 (vs the Genesis’ own release day selling price of Â¥21,000), the add-on was a failure; and the first true killer app for it, Sonic CD was released three years after the release of the SegaCD, in 1994, right about the same time the 32x was released.
Now with thirty-two times the suck!
Okay, so the SegaCD was a failure, that shouldn’t mean anything, right? Wrong. It only started a long standing tradition of Sega consoles sucking. The Sega 32x, released in 1994 for about Â¥18,000 (at least they got the damn price right), was an add-on that plugged in as a cart slot, and allowed you to plug both normal Genesis and 32x games into it.
The 32x was almost like a completely different console. Where the Genesis was powered by a single 68k, the 32x was powered by twin 32Mhz Hitachi SH2 processors; Sega changed to a completely incompatible processing architecture, totally unlike the 68k, and developers had to learn the new processors from scratch; or rather, some did, a vast majority of others simply ignored the platform or incorrectly used it because of that..
In addition to the twin processors, it had a new linear framebuffer design with a simple graphics co-processor, allowing it to composite alpha-blended graphics, and also composite 32x and Genesis graphics in the same scene.
The 32x did all scaling, rotation, and 3D graphics in software instead of using some hardware accelerated method; this arguably made the 32x as powerful as a SNES doing Mode 7 with a SuperFX math co-processor. However, this did not bode well seeing as the SNES was released in 1991, and the first SuperFX game was released in 1993.
Rounding the 32x off, was 256KB of main system RAM; a third of what was added by the Sega CD. Also lacking from the SegaCD was the ability to play back and mix PCM samples, the 32x simply added another synth (which like the SegaCD) was paired with the Genesis’s main one.
Unlike the the SegaCD, the 32x had no killer app. The Sonic title for the 32x didn’t even have Sonic in it, just Team Chaotix in their self-titled game Knuckles’ Chaotix, which had horrible and very annoying game play. I’ve played the game myself, and I can’t stand it; I feel it was rushed, badly planned, and even though it was on the 32x, it was no more graphically intense than Sonic 3 was.
The Neptune
At that time, I remember thinking, why don’t they just release a Genesis that has the 32x and SegaCD built in for little more than what it costs to buy a plain Genesis. About two years ago, I learned about Project Neptune, Sega’s half-hearted answer to that way of thinking.
Originally, the 32x was supposed to be a self-standing platform, a sequal to the Genesis instead of an add-on, as requested by Hayao Nakayama, CEO of Sega of Japan. This original concept project was called Project Jupiter. A few engineers from SoJ and a Sega of America decided to split and create Project Mars, which became the 32x; other engineers joined Project Saturn, which ended up producing the Sega Saturn. Communication between the two projects was very poor, and the 32x and Saturn ultimately became competing products.
Joe Miller (head of Project Mars, from SoA) decided that no one would actually want a simply upgraded Genesis, and ordered that the 32x be built as an add-on instead. By the time Miller realized he was wrong, and build a few prototypes of Project Neptune, a combo Genesis and 32x, the Saturn was already shipping. Project Mars and Project Neptune were complete and utter failures: Project Saturn crushed them.
And this ends part 1. Read part 2 here.