America: The Godfather of Game Development
Chapter 304

One Line of Code Made a Turing Award Laureate Swear

1,788 words6 min read1 view1d agoSource
Next

Silicon California, Silicon Valley.

Sutherland Company.

Ivan Sutherland sat at the table and gently opened a letter.

The sender's name was listed as [Nolan Bushnell].

This was one of his students when he was teaching at the University of Utah.

He was also the creator of Pong, the world's first commercially available arcade video game.

He was also the founder of Atari, the former video game empire.

[Dear Mr. Sutherland:]

[It's as if I'm seeing you in person. I've been busy with the operations of ChunkE. Cheese restaurant these days, so please excuse me for not being able to visit you in person.]

I heard you were awarded this year's Turing Award.

Please allow me to express my highest respect to you!

[Although it's the height of summer----]

Then came some words of comfort.

Ivan smiled slightly, carefully folded the letter, and put it in his filing cabinet.

This place is already neatly filled with greetings from people all over the world.

Clearly, he made quite a splash with this year's Turing Award.

"Mr. Sutherland, if you have finished your business—I hope we can begin as soon as possible."

Sitting across the Table was the president of Sun Microsystems.

He came here specifically today, clearly not to offer his condolences to Ivan.

As a highly paid senior researcher at Sun Microsystems, Ivan's receipt of the Turing Award was definitely a good thing for Sun.

"I know."

Ivan sat back down in his seat and reached for the disc the other person had brought.

"Biohazard"

"Mr. Sutherland, GAMENOVA produced this level of graphics using a Motorola that was far inferior to our SPARC processor."

The visual presentation of Biohazard was so stunning that even those of them working on workstations couldn't ignore it.

For Sun Microsystems, Gamenova was definitely one of the most troublesome companies.

Because they "got too close" to several of their competitors.

Silicon Graphics, Sun's most direct enemy, has completely eroded the graphics station market that Sun originally dominated in recent years thanks to its world-class graphics performance.

Nowadays, apart from SGI's graphics workstations, almost all of Hollywood's workstations are custom MicroNova machines, leaving them with absolutely no room to survive.

Motorola, once Sun's closest partner, has long since become a competitor after Sun developed its own SPARC processor.

To gain a foothold in the current chaotic processor market and break through the encirclement of giants like Intel, Texas Instruments, MIPS, and Motorola, SPARC must create a marketable ecosystem.

"SPARC's performance is absolutely unmatched in the industry."

He said.

This is by no means an exaggeration.

Last year, they developed the SPARC processor, which has a clock speed of 16MHz and uses a cutting-edge 0.8-micron manufacturing process.

Such terrifying processing power naturally outperformed all current CISC complex instruction set processors on the market in an instant.

but----

Ironically, there is one such peculiar company in the world that has been aggressively promoting RISC technology in recent years.

GAMENOVA!

Whether it was the special edition Z80 developed in collaboration with Zilog back then or Motorola today, GAMENOVA has always adhered to the RISC strategy and invested heavily in scientific research.

This really upset them.

In particular, GAMENOVA has a very close relationship with Motorola, with in-depth cooperation in various aspects.

I heard that they've been discussing a partnership with the NEXUS network recently.

That's disgusting.

SPARC should have dominated the entire market, bringing in hundreds of millions of dollars in revenue for Sun Microsystems.

But now it's as if I've been suddenly shoved by an invisible wall, stuck in a position far below what I expected, neither here nor there.

While SPARC is powerful, it is incompatible with most of Gamenova's peripheral technologies.

It is a great tragedy to have potential but not be able to realize it.

"I know."

Ivan repeated the sentence and placed the disc into an NX that was placed on the table.

This wasn't research equipment; he bought it with his own money.

It's impossible for a scholar specializing in computer graphics to have absolutely no interest in games.

"Humm... that's interesting."

Skip the opening animation quickly and enter the game.

Ivan was immediately captivated by the exaggerated scene.

They're right, this is indeed not the level of technology we have in this day and age.

The visuals are so exquisite they're almost unbelievable.

In this day and age, everyone is still Kung-Fu Master hard on 2D. Relying on the ever-improving performance of display chips to create images with higher resolution and richer colors is already considered a very advanced processing technology.

GAMENOVA is like someone who learned to run while learning to walk, and even won an Olympic gold medal along the way.

"This couldn't possibly be rendered in real time."

As a graphics expert who won this year's Turing Award, Ivan quickly drew his own conclusion after only a brief moment of astonishment.

He glanced at the enhancement chip inserted into the NX, then turned to the president of Sun in front of him.

"Has this chip been inspected?"

"It was sent to the lab immediately."

The other person nodded, their expression grave. "This is just a 16-bit processor, a standard Nova Radeon architecture, at most with a slightly higher frequency."

"Not even 32-bit?"

Even Ivan was starting to lose his composure.

If GAMENOVA had gone to the extreme of customizing a 32-bit next-generation processor to handle the 3D graphics, then it would be somewhat acceptable.

But this thing is actually just a standard 16-bit chip?

"Hmm—this isn't real-time rendering; that's the only thing I can confirm."

Ivan piloted Chris around the mansion, but couldn't discern anything amiss.

After a long while, he reluctantly turned off the game.

"GAMENOVA has always done a good job with encryption, so trying to figure out the secrets behind Biohazard by delving into its source code is probably unrealistic."

"But----"

Ivan paused, then tapped a few times on his PC. "We can work backwards from the technical implementation of DOOM, which GAMENOVA has already open-sourced, and NovaEngine."

"That's great! Get started now."

Sun's president was clearly getting impatient.

The longer it is delayed, the more serious SPARC's problems will become.

This is a truly serious and potentially damaging matter.

If we can use this opportunity to thoroughly understand GAMENOVA's homework, we might even have a fighting chance and overtake them on a curve.

Once their secret weapon, the 64-bit 0.18-micron processor UltraSPARCII, leaves the lab next year, it is expected to wipe out the entire market in one fell swoop.

Studying source code is no easy task.

Even with highly readable code written in the N Language, and even though Ivan himself was an expert in this field, it still consumed a great deal of his energy.

"That's strange."

Ivan frowned.

Gamenova is arguably the most unreasonable company in the industry.

Every step it took was beyond its own expectations.

But it somehow manages to do everything it wants in this completely unconventional way.

He did even better than other rule-abiding people!

That's infuriating!

"How?"

"Square root."

Ivan slowly uttered two words, and then fell into complete silence.

Whether in video games or business software, there exists a seemingly insignificant algorithm function.

Sqrt, which is the function for finding the square root.

Calculating spatial distances and motion states of characters, objects, and scenes requires frequent calls to square root calculations.

You could say this thing is the cornerstone of 3D games.

It is also often one of the most computationally expensive functions in a computer.

Because the low-level adders in computers simply cannot understand division.

Addition is a low-level logic function of the processor.

Since the advent of two's complement, subtraction can also be regarded as a kind of addition, just like 1-1=1+(-1).

As for multiplication, there are also many efficient and optimized algorithms, which can be easily achieved by simply repeating addition.

Division is the only operation that, when implemented, has a higher complexity than the other three operations combined.

The time required to perform a division operation can be several times that of a multiplication operation.

In a 3D environment, the presence of floating-point numbers and frequent spatial calculations make division calls unavoidable.

The implementation of square root operation is very similar to that of division, which makes processors overwhelmed when dealing with them.

This is also the main reason why 3D has been difficult to promote.

It's really too demanding on your health!

Ivan himself had done the calculations.

Given the current pace of hardware iteration, even with an upgrade to 32-bit, it will likely be difficult to achieve widespread adoption of 3D technology.

But----

GAMENOVA just wrote it like that.

It makes no attempt to avoid using square roots, yet its performance is surprisingly good.

This is absolutely unusual.

According to his calculations, if Gamenova truly implemented true 3D in this way, it would be impossible for a device like NX to run it.

"Hmm? This square root reciprocal is not a built-in instruction of the processor."

He suddenly noticed something unusual.

These function calls for square root reciprocals do not appear to originate from Motorola's own instruction set, but rather from a custom-written function.

Ivan's eyes lit up, and he immediately perked up.

The secret to everything, GAMENOVA's true technological trump card, lies here!

I am now infinitely close to the truth of the secret!

"Wait a minute—there are only 10 lines??"

Following the call path, the expected hundreds or thousands of lines of brilliant algorithm modules did not appear.

Aside from the function declarations, there are only 10 lines of code lying quietly inside.

"Newton's approximation method?"

Ivan recognized it at a glance; it was just a common square root operation, nothing special.

His eyes immediately dimmed.

"Etc----"

Just as he was about to close this section of code, he suddenly noticed a line of code that was completely nonsensical.

i = 05f3759df - (i >> 1);

"WHAT THE FUCK?"

It felt like something suddenly jumped onto his face and hit him with a cannonball in broad daylight.

Computer science is a precise science.

Ivan has always firmly believed this.

At least code is such an elegant thing that it should be able to express everything precisely.

GAMENOVA, on the other hand, inexplicably inserted an illogical number into its code.

0.5f3759df, converted to decimal, is 1597463007.

There's no logic to it.

But Ivan was certain that this thing was the key to GAMENOVA's high performance in fast square root reciprocal calculations.

He silently picked up the keyboard and typed a comment after the code.

This is his habit; whether reading or writing code, adding comments helps with learning.

But this time, he really didn't know what to write to explain this content.

//What the fuck?

All he could leave behind was a line of America elegance in the code.

End of Chapter
Prev
NovelSummaries
Next

Explore the wiki

Comments