Limitless Dream
[132] Misrill Network
‘It's a pity. ’
A subtle regret. Sometimes I feel it when I look at the tartarus.
Kerberos' own data processing speed is astonishingly fast. The process of loading, calculating, and storing data ends quickly.
However, the rate at which information is exchanged between different Kerberos is not the same. because it is limited by existing network circuit speeds.
Tartaros is conceptually more of a server cluster (where multiple servers are bundled together and built like a single server). It's a system built around 10,000 Kerberos.
The network equipment that connects each Kerberos had to use the old one, which brought about the speed limit.
“At this rate, 50 exaflops. If you install the right OS and go to the good network equipment, how much performance will it increase? ”
Programming capabilities are still lacking, and insights about OS are not refreshing.
However, semiconductors are different. My insights show that I have already amazed myself. That's why insights work better.
Looking at the schematics of the server network relay device, Seojin focused his mind.
‘Come on, show me. ’
A golden glow appeared on the monitor.
It was the light that led to the truth, which he could see.
“Someone thought the lab here was your personal studio. ”
Early in the morning, Park Hyo-san said a word when he saw touching the fair equipment. No hard feelings, just a little teasing.
Han Seo-jin responded favorably.
“Come on, I fixed the EPR-2 problem. ”
“Yes, but I still haven't received the research money I promised to give you. That's why I didn't get anything. ”
“I notarized it, but you didn't give it to me? Is that possible?"
“What if the person who needs to pay doesn't pay? First of all, Choi has said that it must be enforced from the beginning of next year, but I'm still waiting for it. ”
Han Seo-jin reminded me of the confident face of Lee. She's the one who's impeding the administration of the research fund.
‘I don't think this is going to help Jinsung Electronics. ’
If you can inherit the group, does that mean it doesn't matter if it's harmful to the group? Han Seo-jin didn't understand her thoughts.
Then a phone call came from Song Hana.
“Uh, just one. What's going on?”
I took a vacation just now.
“Now?”
Not today, not now. Looks like the vacation just ended.
I'm bored, what are you doing?
“Me? I'm working on a prototype I talked to you about. ”
"Wow, already? It's only been two days.
“I had something in mind. I'm just a little ahead of schedule. ”
Then can I see it too?
“You? Making a semiconductor won't be fun.... ”
I want to see it too. I wonder. "
Since there was no reason to refuse, Seojin nodded after thinking for a moment.
“Then come. This is the Korean University Semiconductor Institute. ”
I'll be right there.
Hanging up the phone, Song Na arrived less than 40 minutes later. With a tour ticket, she smiled as Han Seo-jin left.
“You're early. ”
“I don't have a lot of traffic. Wow, this is where you work. ”
“No, this is not a workplace, this is a school subsidiary. ”
Han Seo-jin thought for a moment and added.
“By the way, all the prototypes are being built here. It's almost like a workshop. ”
Park Hyo-San seems not to be very wrong to ask if this is a private workshop. Suddenly, I was a little sorry to the other guys.
When I entered Park Hyo-San Lab with one song, the air changed in an instant. Not only Ahn Hong-chul, who likes to rumble, but also Choi Taegyu, who was calm, looked surprised.
“Who's your girlfriend? ”
“It's not like that. I know her. ”
“Hi, I'm Songhana. ”
Song Yang bows his head, clasped. Professor Park Hyo-San whistled lightly.
“Seojin is doing well these days. I have this beautiful girlfriend. ”
“I mean, she's my sister. ”
“It must be now. Well, excuse me for denying it too much in front of the party, man. ”
I was stabbed for no reason, so I turned around. Song Na stared at me with a beautiful eye. Like, "Why? 'I feel a little relieved.’ You're not conscious, are you?
“The mask is up, so I'm going to take off the wafer right now. ”
“What's a wafer? ”
“Like a semiconductor plate. It's usually made of silicone. Oh, you know silicone? ”
“… I know. ”
Song Na lowers her head slightly, slightly shy, then says a little.
“I didn't do that. ”
“Huh?"
Seojin stopped wanting to know what this meant, but only realized it and stopped missing.
“No, it's not that. It's silicon. Semiconductors are made of silicon. It's not what you imagine. ”
Song One's face turned red underneath his ears. Han Seo-jin thought it was cute.
“The main ingredient is silicone and other compounds are added, but I'm going to take it anyway. Watch this."
Han Seo-jin operated the machine remotely. Song watches the robot arm move beyond the reinforced glass, shining its eyes.
Slice the wafer column thinly, examine the laser over the photo mask, cut the wafer with a circuit…….
I came next to Professor Park Hyo-San and watched interesting.
“I don't think it's a computational device, what are you building? ”
“Network chips. ”
“What are you going to do with scobrianium or misrels? You're not trying to expose the Misrelli Semiconductor, are you? ”
“What's wrong with ADSC and patent licensing? The Misrill Semiconductor hasn't been released in years. It's just a wiring circuit. ”
“Why? I just use gold. It'll be less resistant. ”
“There's a reason. You'll see.”
In fact, I didn't know why. I'm just imitating it because insight tells me to use a missile as a wiring circuit.
‘Common sense is to use gold or copper. ’
Why do you want me to use a missile? I want to, but I've never had a bad experience doing what my insights tell me to do, so I can repeat it without saying a word.
So the prototype was finalized. Song Hwang looked at the side and tilted.
“It's so small. ”
“It's a relay chip. This is the beginning.”
“Wasn't it over? ”
“Now we have to build a network substrate. ”
Seojin once again became obsessed with his work.
What was finally completed was a board shaped substrate. The centre was equipped with communication relay chips for data exchange and 30 dedicated slots for plugging in the Schnauzer.
Hyosan Park asked curiously.
“Why not Schnauzer? ”
Perhaps because he is conscious of what he hears, he does not take out the name Cerberus.
“I'm going to try to commercialize it with network equipment. That's why we test it with Schnauzer. ”
“Well, if the performance is right, we can implant it into Z7. ”
Han Seo-jin plugged 30 Schnauzers into the substrate and connected the monitor and the power source. The screen reminded me of the state of all the Schnauzer connections.
He was satisfied with the optimal results.
“No connection status problems. I'll measure network speed. ”
As I worked longer, Song Na looked a little bored, but I was still watching her with interest.
Thirty Schnauzers behaved like a single computing device, connected via a network relay chip. Han Seo-jin measured the velocity by raising the computational load to 70%.
“Huh?"
His complexion turned bad. Park Hyo San also asked curiously.
“There's nothing good about it. It's not bad, but it doesn't look much different from the old equipment. ”
I took the ball and made it, but there's no speed improvement at all. Seojin was embarrassed.
“What's wrong with this? This can't be happening. ”
“Don't you think Schnauzer is a defective product? Try something else. ”
“I have to.”
I tried to replace all 30 Schnauzer, but the same was true. Network operations were normal, but they were performing at the same level as existing equipment.
“Is this about the substrate? ”
Han Seo-jin took an extra substrate, plugged in the Schnauzer separately, and connected to the monitor and power source. However, the network speed was still the same.
Park Hyo San tapped my shoulder lightly.
“Well, genius can make mistakes sometimes. But it works just fine. ”
“This can't be.... ”
“I think it's using a missile. Why don't you trade it for gold and try making it again? ”
“Gold?”
Han Seo-jin bit his lip. Should I replace it with gold and rebuild it? But a design that violates the instructions of the insight?
‘If one goes wrong, everything will go wrong. ’
Insights reveal the design structure that borrows the power of the ether. For those who don't know the meaning of design, it's just a pile of circuitry that they don't understand. So competitors, like Wintel, are struggling not to analyze the Schnauzer circuit design.
The design itself could not function properly just because it was converted to gold.
Song Hana, who was watching while yawning, suddenly pointed to a part and asked.
“But, brother, it said 30 before, so why 60 now? ”
“Huh? Day 60. There are only 30 TPU slots. ”
Seojin was embarrassed to check that part.
The monitor showed that 60 Schnauzers were connected. We only have 30 on board.
“What's going on here? ”
“Isn't that a mistake? ”
“Wait a minute. If you look at the computational operations, there's not 30 of them. 60 of them are right. ”
Unlike the constant network speed, the computational processing capability was correct for the performance that 60 could have seen.
Is this supposed to double Schnauzer's abilities? ’
Suddenly, while thinking about it, Seojin touched another substrate. It was the substrate that first measured performance. When he saw the 30 Schnauzers embedded in it, he slapped his knee.
“Wait! I think I know! ”
“What? What is it? ”
“Let's test it. ”
Han Seo-jin brought in all the spare substrates. He plugged all the Schnauzer into a total of 10 substrates and connected the power source to eight free substrates.
There were 300 Schnauzers on the monitor. Then I realized what happened to Hyosan Park, and was frightened.
“Is this a wireless network? No radio transmissions? Is this even possible? ”
“Looks like it was mainly wired to a missile. Without radio transmissions, the missile wiring seems to be connected to each other virtually. ”
“Phew... This is ridiculous. ”
Wireless connections are made by exchanging radio waves with each other. But there's no such device, and they're wirelessly connected.
It's physically nonsense. If this becomes known to the world, it will flip over.
“This will be very convenient to build a cluster. You don't have to connect the cable separately, and the size of the network equipment is reduced. ”
“The network speed is intact, but this is a huge discovery. We can make the Z7 smaller. ”
“I still struggled with the space that network equipment eats when I was making supercomputers, but this is an amazing discovery. ”
Large-scale system building is accompanied by a large network facility connection. However, this will dramatically reduce its size.
“This is a 100% Nobel Prize. Seojin, you finally did it. ”
“If we miniaturize the substrate to a chip, we could convert the tartarus into something smaller than it is today. ”
The miniaturization of the Nobel Prize and the equipment, the two of them inflated their big dreams.
Song Hana, who was watching the whole time, carefully opened her mouth.
“Come on, brother. ”
“Huh? Ah, just one. It can be hard to understand what this is. In a nutshell, this is individual parts of a computer, or bulk connections between computers……. ”
“I think I know what it is. It's a cell phone antenna. ”
“ ……. ”
The excited professor and disciple were silent in an instant.
Reviews of works
Successfully crafted the sacred collar.
Do not use frequency.
The telecom company and the government hate this invention.
= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =