More from Jorge Arango
Last week, I spoke with a business leader who’s excited about AI. But as we talked, it became clear that there’s a wide gulf between enthusiasm and creating value for a particular business. Most people’s impression of AI is based on limited use and media hype. Take the recent Ghibli-fication mania: millions are smitten with the idea of seeing themselves as a Miyazaki character. (I’m one of them!) And it’s understandable: the outputs are impressive. But AI can do more than make beautiful drawings or write compelling essays. As I’ve argued before, these aren’t the best uses for AI. Instead, we should use it to augment our abilities. But how? It’s hard to see beyond the outputs — especially since doing so entails getting more abstract. I’m still thinking about how to explain it, but three ideas are key: Businesses consist of information flows. Information exists to support decisions. Information can be optimized for better decision-making. Let’s unpack them. First, your business consists of information flows. Whatever your business is, it runs on information: how it’s captured, produced, shared, and processed. A proposal? Information. The request behind it? Also information. A standup meeting? An exchange of information. Your business creates value when it uses information effectively. Sure, that’s not the only way it creates value: the things you make and services you provide are key. But information is essential. Why? Because of the second point: information is in service to decision-making. The proposal helps the prospect decide whether to work with you. Research helps you decide whether to enter a new market. The meeting helps determine next steps. Third, information can be optimized. When I say “information,” you may think spreadsheets and databases. But that’s structured information. Most business information — conversations, documents, emails — is unstructured. Consider that meeting. It may have some structure: an agenda, list of attendees, start and end time. But the stuff you care about — what people say — isn’t structured. Even if you transcribe it, you must still think about what it means for you. AI can help tame the messy information flows that make up real work. Efforts to formalize them often kill spontaneity, nuance, and context. And even if they didn’t, there’s so much information that it’s been hard to make sense of it. But now we have AI. Don’t let the charismatic drawings distract you. That’s only a superficial application — and commodifying art is bad for our souls. Instead, focus on using AI for tasks that were previously impossible or impractical: working with vast amounts of unstructured information, playing out what-if scenarios at scale, and augmenting your team’s expertise. Information architects can help. We’ve been mapping information flows and making sense of unstructured information for decades. If you’re exploring how AI could create real business value — not just flashy outputs — let’s talk.
This week, I sampled one of the world’s great spiritual texts: the Qur’an. I also read poems by Rumi and saw a movie that led me to question my understanding of both. Readings As with many other people in the West, my understanding of Islam mostly comes from the news — often, in negative light. Also, I was raised Catholic and can’t help but parse these teachings through that tradition. Which is to say, I tried to read objectively, but aware that I’m likely biased. That said, I relished the opportunity to experience this book firsthand. The Qur’an has an interesting structure. Like many other books, it’s divided into chapters (surahs.) But instead of a chronological or topical organization, chapters are roughly sequenced from longest to shortest. (Except the first, which is a short prayer.) Gioia recommended fifteen surahs. As I often do after reading classic texts, I asked ChatGPT for a summary. Here it is verbatim: 1. Al-Fatiha (The Opening): A short prayer for guidance, mercy, and the straight path. Central to every unit of Muslim prayer. 2. Al-Baqara (The Cow): Longest surah; covers law, ethics, worship, community life, and stories of earlier prophets. Establishes Islam’s moral and legal foundations. 3. Al-Imran (The Family of Imran): Addresses Christians and Jews, emphasizes the unity of divine messages, and recounts the stories of Mary and Jesus with warnings against disbelief. 4. An-Nisa (The Women): Focuses on justice, women’s rights, inheritance laws, and proper conduct in war and peace; warns against hypocrisy and betrayal. 5. Al-Ma’ida (The Table Spread): Deals with dietary laws, contracts, and relations with Jews and Christians. Emphasizes loyalty to God’s covenant and the finality of Muhammad’s message. 12. Yusuf (Joseph): A narrative surah telling the story of Joseph in a literary and emotional style. A tale of patience, betrayal, forgiveness, and divine providence. 17. Al-Isra (The Night Journey): Takes its name from Muhammad’s mystical journey to Jerusalem and heaven. Emphasizes the Qur’an’s authority and the consequences of moral choices. 18. Al-Kahf (The Cave): Contains allegorical stories (Sleepers of the Cave, Moses and the Wise Man, Dhul-Qarnayn) about faith, knowledge, and the trials of life. 32. As-Sajda (The Prostration): Stresses the importance of divine revelation and resurrection. Calls for humility and recognition of God’s signs in creation. 36. Ya-Sin: Considered the “heart of the Qur’an”; focuses on God’s power, the resurrection, and human accountability, often recited for the dying. 55. Ar-Rahman (The Most Merciful): A poetic celebration of God’s mercy, creation, and judgment—famous for the refrain: “Which of your Lord’s favors will you deny?” 67. Al-Mulk (The Sovereignty): Reflects on God’s dominion over life and death. Encourages awe and contemplation of the natural world as a sign of divine power. 103. Al-Asr (Time): A short, stark reminder that time is running out—success lies in faith, good works, truth, and patience. 112. Al-Ikhlas (Sincerity): A concise declaration of pure monotheism: God is one, eternal, without equal or offspring. As may be evident here, the Qur’an blends spiritual teachings with social instruction, proposing norms, values, and institutions for a society centered on God. If Christianity’s core message is love, the core here is belief — specifically, belief in one supreme God. All else falls in line behind that idea. Reading Rumi deepened my understanding and appreciation. Through his poems, I perceived signals of what it might be like to live in a world saturated with God: ecstasy at the sight of a flower or the moon’s reflection in a still pond. Still, he reminds us these are only words: This is how it always is A great silence overcomes me, Gioia recommended a selection of Rumi’s poems, but I’ve enjoyed them enough that I’ve decided to read a whole collection. Rather than speedruning through it, I’ll read a bit at a time during my morning journaling sessions throughout the rest of the year. Audiovisual Music: Nusrat Fateh Ali Khan plus music from Tunisia. I was familiar with Ali Khan; his music was everywhere in the 1990s. He’s also one of the performers in the LAST TEMPTATION OF CHRIST soundtracks I highlighted earlier this year. Revisiting his music now gave me a new appreciation; like Rumi, he seems to express the ecstasy of living in a world soaked in divinity. Art: Gioia recommended looking at Islamic architecture. I’ve had the privilege of visiting two of the world’s great Islamic buildings, Hagia Sophia in Istanbul and the Alhambra in Granada. So that’s where my mind went. Hagia Sophia is one of the most impressive spaces I’ve ever been in; it must be experienced firsthand. That said, this NOVA program explains the building’s historical, cultural, and structural significance: This short presentation helped me understand the spiritual significance of sequencing and structure in Islamic religious buildings: The Alhambra is impressive for different reasons. For one thing, it’s not a religious structure. It’s arguably most famous for its beautiful mosaics. This video focuses on the palace’s hydraulic network, one of the world’s great engineering feats: This TEDEd presentation helped me understand the significance of the astonishing geometric decorations in these buildings: Both of these buildings also highlight Islam’s complex relationship with Christianity. Hagia Sophia started as a Christian church before being converted into a mosque, and the Alhambra stands as a reminder that part of Europe was under Islamic rule for a time. Cinema: rather than selecting a movie related to this week’s readings, I chose one I’d wanted to watch for a long time: Hal Ashby’s 1979 satire BEING THERE, starring the great Peter Sellers. By chance, a sheltered, single-minded, TV-obsessed, illiterate gardener named Chance is unwittingly thrust into the highest echelons of economic and political power. Due to his genteel demeanor, his naive utterances are continually misinterpreted as profundities. He remains aloof and uninvolved even as he becomes incredibly powerful. Rather, he becomes powerful because he is aloof and uninvolved. He has no awareness of (or concern for) how his words are interpreted, so he remains cool and neutral. This amplifies his image as a sage, feeding his growing myth. He and his interlocutors live in diverging worlds, reading into each others’ utterances meanings that aren’t there. Reflections Or are they? The movie suggests words don’t have inherent meaning; the meaning is in how we interpret them. Context informs understanding. Lacking an official backstory, people parse Chance through their preconceptions. In the first act, he introduces himself as “Chance the gardener.” But the idea that this seemingly refined gentleman is a mere gardener seems so unlikely that his interlocutor misinterprets his name as the more upper class Chancey Gardiner. And so it remains. When Chance(y) says “I like to watch,” he means something very different than what others hear, leading to particular and often unexpected actions. And it’s actions that ultimately matter. The Talmud puts it succinctly: “We do not see things as they are, we see them as we are.” I think about this whenever I engage with ancient scriptures. These texts were produced by and for very different cultures. The words mean different things to me than they likely meant to their original audiences. It’s not just that they were written in a language I don’t understand: they were written in a world I don’t understand. And yet, they endure. Why? We make the meaning. More precisely: meaning emerges as we parse the text through our particular understandings. And meaning adds value to our lives. What I took from my reading is that Islam positions itself as the culmination of monotheism. Characters, stories, and themes from the Jewish and Christian scriptures are present here too, but with a different focus. Somewhere along the line, Judaism and Christianity softened on the idea that there’s only one God. Islam aims to correct course. The idea of single-mindedly centering your life on something other than yourself is appealing. (Post-Freud, we’d say ego is the enemy.) The concept isn’t unique to Islam, of course. (And again, I’m likely reading into it.) What seems different is the Qur’an’s confident assertion of its message’s universal relevance. I can see how this all-encompassing vision can be seen as incompatible with secular post-modern societies. The Abrahamic traditions are also closely related, which can foster mutual understanding but also intensify disagreement. But I also sense the conflict emerges from literal readings, which we shouldn’t do. Again, these words emerge from — and speak to — different contexts. And I see the transcendent beauty in Rumi’s poems, the patterns of the Alhambra, and Nusrat Fateh Ali Khan’s singing, much as I do in Chartres Cathedral, the Mass in B Minor, and the Divine Comedy. All remind me that words can connect people to higher orders of being and to each other, elevating us all. Does God come in through the words or through what we bring to the words? Does it matter? It does. As Nietzsche warned, once we lose sense of the words’ specialness, we approach them differently — much like Chancey’s friends would hear him differently if they knew he was mere Chance. Some might say truth is what ultimately matters: that Chance is just a gardener who is, in the words of a character who knows, “stuffed with rice pudding between th’ ears. Shortchanged by the Lord, and dumb as a jackass.” But another character who also knows opts to not reveal the “truth” to avoid destroying his dying friend’s comforting illusion. An act of mercy? Delusion? Both? And then there’s the film’s last shot, which I won’t spoil. Let’s just say it casts doubts on whether any of these characters — or us voyeurs in this world of Chance — really knows what is in front of our eyes. Notes on Note-taking I used ChatGPT more extensively this week than at any previous point in the course. I started a ChatGPT project with the following system prompt: You are an expert tutor in literature, history, music, cinema, and more – in short, the humanities. I am in a year-long program to read through (and listen to and watch) the most important humanities works from across the world. You’ll provide summaries, pointers, and advice. Respond concisely but helpfully. Whenever possible, contextualize content by pointing out possible parallels with the modern day. Within this context, I asked ChatGPT to bring perspective to what I was reading — especially some of the most difficult passages and parallels between the other two Abrahamic traditions. I found parts of surah 4 challenging — especially those that address warfare and justice. I asked ChatGPT for a charitable interpretation of the scripture, and the LLM helpfully responded with explanations that helped me understand the text in its broader context. ChatGPT didn’t recommend it, but this video from Bishop Robert Barron also reminded me that the Judeo-Christian scriptures also have similarly difficult passages and that we must interpret them in the context for which they were written: After watching the TEDEd video about geometry in Islamic architecture, I asked ChatGPT, “How does the repetition through tessellation evident in Islamic decoration reflect the religion’s spiritual values?” I’ll quote its answer verbatim, since I found it both insightful and exemplary of the value I’m getting from using AI in this way: Beautiful question—and one that gets at the visual theology of Islam. Islamic decoration, especially through geometric tessellation, is not just ornamental—it’s deeply expressive of spiritual truths. Here’s how repetition, pattern, and abstraction in Islamic art reflect the core values of the faith: ⸻ 1. Unity in Multiplicity (Tawhid) The oneness of God (tawhid) is Islam’s central belief. Repeating geometric patterns express how divine unity underlies all diversity—a single pattern unfolds into infinite forms, just as all creation reflects the One. Spiritual message: Beneath complexity lies a single divine source. ⸻ 2. Infinity and the Transcendent Tessellations can theoretically extend forever—they have no beginning or end. This evokes infinity, a symbol of God’s eternal, boundless nature. The viewer is drawn into a contemplative state by the endless repetition. Spiritual message: God is beyond time, form, and limit. ⸻ 3. Non-figurative, Non-idolatrous Representation Islamic tradition discourages figural imagery in sacred contexts to avoid idolatry. Abstract, geometric art avoids representing God and instead points toward Him by evoking beauty, harmony, and transcendence. Spiritual message: God cannot be pictured—but can be reflected in patterns of perfection. ⸻ 4. Contemplation and Remembrance (Dhikr) Like repeated prayer or names of God, visual repetition becomes a form of remembrance. Islamic art invites stillness, reflection, and mental surrender—a visual dhikr. Spiritual message: Still the ego, contemplate the divine, dissolve into the pattern. ⸻ 5. Mathematics as Sacred Language Islamic artists used advanced geometry and symmetry, often guided by spiritual cosmology. The precision of the art mirrors the divine order in the cosmos, hinting that the universe is intelligible, ordered, and meaningful. Spiritual message: Creation is a sign (ayah) pointing to its Creator. The LLM included a helpful table: Artistic Feature Spiritual Reflection Repetition / Tessellation Unity, infinity, divine order No images of God or people Transcendence, anti-idolatry Symmetry and precision Balance, harmony, divine perfection Infinite extension God’s boundlessness and eternal presence Meditative patterning Encourages contemplation and surrender It concluded: Islamic art, in this sense, functions like visual theology: it doesn’t depict the divine, but invites you to perceive it—not through representation, but through pattern, balance, and beauty. Would you like to see some examples or diagrams that illustrate how these patterns are constructed mathematically or spiritually? I was intrigued by this last question, especially given that OpenAI just released improved ChatGPT image rendering capabilities. So I responded, “Yes, please show me an example diagram that illustrates these principles.” At this point, ChatGPT generated this image: Up Next We’re moving into the second quarter of the course. Gioia recommends books 1 and 2 of the Aeneid, book 1 of Ovid’s Metamorphoses, and a few other Roman poems, aphorisms, and proverbs. So we’re on our way back to Rome! Again, there’s a YouTube playlist for the videos I’m sharing here. I’m also sharing these posts via Substack if you’d like to subscribe and comment. See you next week!
In Episode 6 of the Traction Heroes podcast, Harry and I explored Chesterton’s fence — a simple yet profound idea that has important implications for leaders navigating complex, high-stakes changes. The gist: when change is needed, don’t start by destroying what you don’t understand. Assume things are the way they are because of reasons. Once you understand the reasons, you’re more likely to avoid unintended consequences when making changes. Here’s the passage I read from Chesterton’s The Thing: In the matter of reforming things, as distinct from deforming them, there is one plain and simple principle; a principle which will probably be called a paradox. There exists in such a case a certain institution or law; let us say for the sake of simplicity, a fence or gate erected across a road. The more modern type of reformer goes gaily up to it and says, “I don’t see the use of this; let us clear it away.” To which the more intelligent type of reformer will do well to answer: “If you don’t see the use of it, I certainly won’t let you clear it away. Go away and think. Then, when you can come back and tell me that you do see the use of it, I may allow you to destroy it.” This paradox rests on the most elementary common sense. The gate or fence did not grow there. It was not set up by somnambulists who built it in their sleep. It is highly improbable that it was put there by escaped lunatics who were for some reason loose in the street. Some person had some reason for thinking it would be a good thing for somebody. And until we know what the reason was, we really cannot judge whether the reason was reasonable. It is extremely probable that we have overlooked some whole aspect of the question, if something set up by human beings like ourselves seems to be entirely meaningless and mysterious. There are reformers who get over this difficulty by assuming that all their fathers were fools; but if that be so, we can only say that folly appears to be a hereditary disease. Catastrophic outcomes happen for many reasons. One of the worst is what Harry called stupidity: “a result of a series of actions that lead to an outcome that’s the opposite of what you say you want, under conditions of self-deception.” Perhaps if more people knew about Chesterton’s fence there would be less suffering caused by stupidity. As always, I learned a lot from bouncing these ideas off Harry. Among other things, he responded with an intriguing followup book. Perhaps that will be the subject of a future episode. Stay tuned for more!
Louis Rosenfeld interviewed Harry Max and me for his Rosenfeld Review podcast. The subject? Harry and my podcast, Traction Heroes. We recorded this conversation late in 2024, before we’d shared the first episode. This interview lays out Traction Heroes’s backstory. It’s fitting that we shared it in Lou’s show, since he published Harry’s book Managing Priorities and my Living in Information and Duly Noted. The Rosenfeld Review Podcast (Rosenfeld Media) · Traction Heroes with Harry Max & Jorge Arango Listen on SoundCloud
More in technology
Most people think of machine instructions as the fundamental steps that a computer performs. However, many processors have another layer of software underneath: microcode. With microcode, instead of building the processor's control circuitry from complex logic gates, the control logic is implemented with code known as microcode, stored in the microcode ROM. To execute a machine instruction, the computer internally executes several simpler micro-instructions, specified by the microcode. In this post, I examine the microcode ROM in the original Pentium, looking at the low-level circuitry. The photo below shows the Pentium's thumbnail-sized silicon die under a microscope. I've labeled the main functional blocks. The microcode ROM is highlighted at the right. If you look closely, you can see that the microcode ROM consists of two rectangular banks, one above the other. This die photo of the Pentium shows the location of the microcode ROM. Click this image (or any other) for a larger version. The image below shows a closeup of the two microcode ROM banks. Each bank provides 45 bits of output; together they implement a micro-instruction that is 90 bits long. Each bank consists of a grid of transistors arranged into 288 rows and 720 columns. The microcode ROM holds 4608 micro-instructions, 414,720 bits in total. At this magnification, the ROM appears featureless, but it is covered with horizontal wires, each just 1.5 µm thick. The 90 output lines from the ROM, with a closeup of six lines exiting the ROM. The ROM's 90 output lines are collected into a bundle of wires between the banks, as shown above. The detail shows how six of the bits exit from the banks and join the bundle. This bundle exits the ROM to the left, travels to various parts of the chip, and controls the chip's circuitry. The output lines are in the chip's top metal layer (M3): the Pentium has three layers of metal wiring with M1 on the bottom, M2 in the middle, and M3 on top. The Pentium has a large number of bits in its micro-instruction, 90 bits compared to 21 bits in the 8086. Presumably, the Pentium has a "horizontal" microcode architecture, where the microcode bits correspond to low-level control signals, as opposed to "vertical" microcode, where the bits are encoded into denser micro-instructions. I don't have any information on the Pentium's encoding of microcode; unlike the 8086, the Pentium's patents don't provide any clues. The 8086's microcode ROM holds 512 micro-instructions, much less than the Pentium's 4608 micro-instructions. This makes sense, given the much greater complexity of the Pentium's instruction set, including the floating-point unit on the chip. The image below shows a closeup of the Pentium's microcode ROM. For this image, I removed the three layers of metal and the polysilicon layer to expose the chip's underlying silicon. The pattern of silicon doping is visible, showing the transistors and thus the data stored in the ROM. If you have enough time, you can extract the bits from the ROM by examining the silicon and seeing where transistors are present. A closeup of the ROM showing how bits are encoded in the layout of transistors. Before explaining the ROM's circuitry, I'll review how an NMOS transistor is constructed. A transistor can be considered a switch between the source and drain, controlled by the gate. The source and drain regions (green) consist of silicon doped with impurities to change its semiconductor properties, forming N+ silicon. (These regions are visible in the photo above.) The gate consists of a layer of polysilicon (red), separated from the silicon by a very thin insulating oxide layer. Whenever polysilicon crosses active silicon, a transistor is formed. Diagram showing the structure of an NMOS transistor. Bits are stored in the ROM through the pattern of transistors in the grid. The presence or absence of a transistor stores a 0 or 1 bit.1 The closeup below shows eight bits of the microcode ROM. There are four transistors present and four gaps where transistors are missing. Thus, this part of the ROM holds four 0 bits and four 1 bits. For the diagram below, I removed the three metal layers and the polysilicon to show the underlying silicon. I colored doped (active) silicon regions green, and drew in the horizontal polysilicon lines in red. As explained above, a transistor is created if polysilicon crosses doped silicon. Thus, the contents of the ROM are defined by the pattern of silicon regions, which creates the transistors. Eight bits of the microcode ROM, with four transistors present. The horizontal silicon lines are used as wiring to provide ground to the transistors, while the horizontal polysilicon lines select one of the rows in the ROM. The transistors in that row will turn on, pulling the associated output lines low. That is, the presence of a transistor in a row causes the output to be pulled low, while the absence of a transistor causes the output line to remain high. A schematic corresponding to the eight bits above. The diagram below shows the silicon, polysilicon, and bottom metal (M1) layers. I removed the metal from the left to reveal the silicon and polysilicon underneath, but the pattern of vertical metal lines continues there. As shown earlier, the silicon pattern forms transistors. Each horizontal metal line has a connection to ground through a metal line (not shown). The horizontal polysilicon lines select a row. When polysilicon lines cross doped silicon, the gate of a transistor is formed. Two transistors may share the drain, as in the transistor pair on the left. Diagram showing the silicon, polysilicon, and M1 layers. The vertical metal wires form the outputs. The circles are contacts between the metal wire and the silicon of a transistor.2 Short metal jumpers connect the polysilicon lines to the metal layer above, which will be described next. The image below shows the upper left corner of the ROM. The yellowish metal lines are the top metal layer (M3), while the reddish metal lines are the middle metal layer (M2). The thick yellowish M3 lines distribute ground to the ROM. Underneath the horizontal M3 line, a horizontal M2 line also distributes ground. The grids of black dots are numerous contacts between the M3 line and the M2 line, providing a low-resistance connection. The M2 line, in turn, connects to vertical M1 ground lines underneath—these wide vertical lines are faintly visible. These M1 lines connect to the silicon, as shown earlier, providing ground to each transistor. This illustrates the complexity of power distribution in the Pentium: the thick top metal (M3) is the primary distribution of +5 volts and ground through the chip, but power must be passed down through M2 and M1 to reach the transistors. The upper left corner of the ROM. The other important feature above is the horizontal metal lines, which help distribute the row-select signals. As shown earlier, horizontal polysilicon lines provide the row-select signals to the transistors. However, polysilicon is not as good a conductor as metal, so long polysilicon lines have too much resistance. The solution is to run metal lines in parallel, periodically connected to the underlying polysilicon lines and reducing the overall resistance. Since the vertical metal output lines are in the M1 layer, the horizontal row-select lines run in the M2 layer so they don't collide. Short "jumpers" in the M1 layer connect the M2 lines to the polysilicon lines. To summarize, each ROM bank contains a grid of transistors and transistor vacancies to define the bits of the ROM. The ROM is carefully designed so the different layers—silicon, polysilicon, M1, and M2—work together to maximize the ROM's performance and density. Microcode Address Register As the Pentium executes an instruction, it provides the address of each micro-instruction to the microcode ROM. The Pentium holds this address—the micro-address—in the Microcode Address Register (MAR). The MAR is a 13-bit register located above the microcode ROM. The diagram below shows the Microcode Address Register above the upper ROM bank. It consists of 13 bits; each bit has multiple latches to hold the value as well as any pushed subroutine micro-addresses. Between bits 7 and 8, some buffer circuitry amplifies the control signals that go to each bit's circuitry. At the right, drivers amplify the outputs from the MAR, sending the signals to the row drivers and column-select circuitry that I will discuss below. To the left of the MAR is a 32-bit register that is apparently unrelated to the microcode ROM, although I haven't determined its function. The Microcode Address Register is located above the upper ROM bank. The outputs from the Microcode Address Register select rows and columns in the microcode ROM, as I'll explain below. Bits 12 through 7 of the MAR select a block of 8 rows, while bits 6 through 4 select a row in this block. Bits 3 through 0 select one column out of each group of 16 columns to select an output bit. Thus, the microcode address controls what word is provided by the ROM. Several different operations can be performed on the Microcode Address Register. When executing a machine instruction, the MAR must be loaded with the address of the corresponding microcode routine. (I haven't determined how this address is generated.) As microcode is executed, the MAR is usually incremented to move to the next micro-instruction. However, the MAR can branch to a new micro-address as required. The MAR also supports microcode subroutine calls; it will push the current micro-address and jump to the new micro-address. At the end of the micro-subroutine, the micro-address is popped so execution returns to the previous location. The MAR supports three levels of subroutine calls, as it contains three registers to hold the stack of pushed micro-addresses. The MAR receives control signals and addresses from standard-cell logic located above the MAR. Strangely, in Intel's published floorplans for the Pentium, this standard-cell logic is labeled as part of the branch prediction logic, which is above it. However, carefully tracing the signals from the standard-cell logic shows that is connected to the Microcode Address Register, not the branch predictor. Row-select drivers As explained above, each ROM bank has 288 rows of transistors, with polysilicon lines to select one of the rows. To the right of the ROM is circuitry that activates one of these row-select lines, based on the micro-address. Each row matches a different 9-bit address. A straightforward implementation would use a 9-input AND gate for each row, matching a particular pattern of 9 address bits or their complements. However, this implementation would require 576 very large AND gates, so it is impractical. Instead, the Pentium uses an optimized implementation with one 6-input AND gate for each group of 8 rows. The remaining three address bits are decoded once at the top of the ROM. As a result, each row only needs one gate, detecting if its group of eight rows is selected and if the particular one of eight is selected. Simplified schematic of the row driver circuitry. The schematic above shows the circuitry for a group of eight rows, slightly simplified.3 At the top, three address bits are decoded, generating eight output lines with one active at a time. The remaining six address bits are inverted, providing the bit and its complement to the decoding circuitry. Thus, the 9 bits are converted into 20 signals that flow through the decoders, a large number of wires, but not unmanageable. Each group of eight rows has a 6-input AND gate that matches a particular 6-bit address, determined by which inputs are complemented and which are not.4 The NAND gate and inverter at the left combine the 3-bit decoding and the 6-bit decoding, activating the appropriate row. Since there are up to 720 transistors in each row, the row-select lines need to be driven with high current. Thus, the row-select drivers use large transistors, roughly 25 times the size of a regular transistor. To fit these transistors into the same vertical spacing as the rest of the decoding circuitry, a tricky packing is used. The drivers for each group of 8 rows are packed into a 3×3 grid, except the first column has two drivers (since there are 8 drivers in the group, not 9). To avoid a gap, the drivers in the first column are larger vertically and squashed horizontally. Output circuitry The schematic below shows the multiplexer circuit that selects one of 16 columns for a microcode output bit. The first stage has four 4-to-1 multiplexers. Next, another 4-to-1 multiplexer selects one of the outputs. Finally, a BiCMOS driver amplifies the output for transmission to the rest of the processor. The 16-to-1 multiplexer/output driver. In more detail, the ROM and the first multiplexer are essentially NMOS circuits, rather than CMOS. Specifically, the ROM's grid of transistors is constructed from NMOS transistors that can pull a column line low, but there are no PMOS transistors in the grid to pull the line high (since that would double the size of the ROM). Instead, the multiplexer includes precharge transistors to pull the lines high, presumably in the clock phase before the ROM is read. The capacitance of the lines will keep the line high unless it is pulled low by a transistor in the grid. One of the four transistors in the multiplexer is activated (by control signal a, b, c, or d) to select the desired line. The output goes to a "keeper" circuit, which keeps the output high unless it is pulled low. The keeper uses an inverter with a weak PMOS transistor that can only provide a small pull-up current. A stronger low input will overpower this transistor, switching the state of the keeper. The output of this multiplexer, along with the outputs of three other multiplexers, goes to the second-stage multiplexer,5 which selects one of its four inputs, based on control signals e, f, g, and h. The output of this multiplexer is held in a latch built from two inverters. The second latch has weak transistors so the latch can be easily forced into the desired state. The output from the first latch goes through a CMOS switch into a second latch, creating a flip-flop. The output from the second latch goes to a BiCMOS driver, which drives one of the 90 microcode output lines. Most processors are built from CMOS circuitry (i.e. NMOS and PMOS transistors), but the Pentium is built from BiCMOS circuitry: bipolar transistors as well as CMOS. At the time, bipolar transistors improved performance for high-current drivers; see my article on the Pentium's BiCMOS circuitry. The diagram below shows three bits of the microcode output. This circuitry is for the upper ROM bank; the circuitry is mirrored for the lower bank. The circuitry matches the schematic above. Each of the three blocks has 16 input lines from the ROM grid. Four 4-to-1 multiplexers reduce this to 4 lines, and the second multiplexer selects a single line. The result is latched and amplified by the output driver. (Note the large square shape of the bipolar transistors.) Next is the shift register that processes the microcode ROM outputs for testing. The shift register uses XOR logic for its feedback; unlike the rest of the circuitry, the XOR logic is irregular since only some bits are fed into XOR gates. Three bits of output from the microcode, I removed the three metal layers to show the polysilicon and silicon. Circuitry for testing Why does the microcode ROM have shift registers and XOR gates? The reason is that a chip such as the Pentium is very difficult to test: if one out of 3.1 million transistors goes bad, how do you detect it? For a simple processor like the 8086, you can run through the instruction set and be fairly confident that any problem would turn up. But with a complex chip, it is almost impossible to design an instruction sequence that would test every bit of the microcode ROM, every bit of the cache, and so forth. Starting with the 386, Intel added circuitry to the processor solely to make testing easier; about 2.7% of the transistors in the 386 were for testing. The Pentium has this testing circuitry for many ROMs and PLAs, including the division PLA that caused the infamous FDIV bug. To test a ROM inside the processor, Intel added circuitry to scan the entire ROM and checksum its contents. Specifically, a pseudo-random number generator runs through each address, while another circuit computes a checksum of the ROM output, forming a "signature" word. At the end, if the signature word has the right value, the ROM is almost certainly correct. But if there is even a single bit error, the checksum will be wrong and the chip will be rejected. The pseudo-random numbers and the checksum are both implemented with linear feedback shift registers (LFSR), a shift register along with a few XOR gates to feed the output back to the input. For more information on testing circuitry in the 386, see Design and Test of the 80386, written by Pat Gelsinger, who became Intel's CEO years later. Conclusions You'd think that implementing a ROM would be straightforward, but the Pentium's microcode ROM is surprisingly complex due to its optimized structure and its circuitry for testing. I haven't been able to determine much about how the microcode works, except that the micro-instruction is 90 bits wide and the ROM holds 4608 micro-instructions in total. But hopefully you've found this look at the circuitry interesting. Disclaimer: this should all be viewed as slightly speculative and there are probably some errors. I didn't want to prefix every statement with "I think that..." but you should pretend it is there. I plan to write more about the implementation of the Pentium, so follow me on Bluesky (@righto.com) or RSS for updates. Peter Bosch has done some reverse engineering of the Pentium II microcode; his information is here. Footnotes and references It is arbitrary if a transistor corresponds to a 0 bit or a 1 bit. A transistor will pull the output line low (i.e. a 0 bit), but the signal could be inverted before it is used. More analysis of the circuitry or ROM contents would clear this up. ↩ When looking at a ROM like this, the contact pattern seems like it should tell you the contents of the ROM. Unfortunately, this doesn't work. Since a contact can be attached to one or two transistors, the contact pattern doesn't give you enough information. You need to see the silicon to determine the transistor pattern and thus the bits. ↩ I simplified the row driver schematic. The most interesting difference is that the NAND gates are optimized to use three transistors each, instead of four transistors. The trick is that one of the NMOS transistors is essentially shared across the group of 8 drivers; an inverter drives the low side of all eight gates. The second simplification is that the 6-input AND gate is implemented with two 3-input NAND gates and a NOR gate for electrical reasons. Also, the decoder that converts 3 bits into 8 select lines is located between the banks, at the right, not at the top of the ROM as I showed in the schematic. Likewise, the inverters for the 6 row-select bits are not at the top. Instead, there are 6 inverters and 6 buffers arranged in a column to the right of the ROM, which works better for the layout. These are BiCMOS drivers so they can provide the high-current outputs necessary for the long wires and numerous transistor gates that they must drive. ↩ The inputs to the 6-input AND gate are arranged in a binary counting pattern, selecting each row in sequence. This binary arrangment is standard for a ROM's decoder circuitry and is a good way to recognize a ROM on a die. The Pentium has 36 row decoders, rather than the 64 that you'd expect from a 6-bit input. The ROM was made to the size necessary, rather than a full power of two. In most ROMs, it's difficult to determine if the ROM is addressed bottom-to-top or top-to-bottom. However, because the microcode ROM's counting pattern is truncated, one can see that the top bank starts with 0 at the top and counts downward, while the bottom bank is reversed, starting with 0 at the bottom and counting upward. ↩ A note to anyone trying to read the ROM contents: it appears that the order of entries in a group of 16 is inconsistent, so a straightforward attempt to visually read the ROM will end up with scrambled data. That is, some of the groups are reversed. I don't see any obvious pattern in which groups are reversed. A closeup of the first stage output mux. This image shows the M1 metal layer. In the diagram above, look at the contacts from the select lines, connecting the select lines to the mux transistors. The contacts on the left are the mirror image of the contacts on the right, so the columns will be accessed in the opposite order. This mirroring pattern isn't consistent, though; sometimes neighboring groups are mirrored and sometimes they aren't. I don't know why the circuitry has this layout. Sometimes mirroring adjacent groups makes the layout more efficient, but the inconsistent mirroring argues against this. Maybe an automated layout system decided this was the best way. Or maybe Intel did this to provide a bit of obfuscation against reverse engineering. ↩
If it's all just electromagnetic waves, why is electricity in a conductor moving slower than visible light?
Everyone should pull one great practical joke in their lifetimes. This one was mine, and I think it's past the statute of limitations. The story is true. Only the names are redacted to protect the guilty. My first job out of college was a database programmer, even though my undergraduate degree had nothing to do with computers and my current profession still mostly doesn't. The reason was that the University I worked for couldn't afford competitive wages, but they did offer various fringe benefits, and they were willing to train someone who at least had decent working knowledge. I, as a newly minted graduate of the august University of California system, had decent working knowledge at least of BSD/386 and SunOS, but more importantly also had the glowing recommendation of my predecessor who was being promoted into a new position. I was hired, which was their first mistake. The system I was hired to work on was an HP 9000 K250, one of Hewlett-Packard's big PA-RISC servers. I wish I had a photograph of it, but all I have are a couple bad scans of some bad Polaroids of my office and none of the server room. The server room was downstairs from my office back in the days when server rooms were on-premises, complete with a swipe card lock and a halon system that would give you a few seconds of grace before it flooded everything. The K250 hulked in there where it had recently replaced what I think was an Encore mini of some sort (probably a Multimax, since it was a few years old and the 88K Encores would have been too new for the University), along with the AIX RS/6000s that provided student and faculty shell accounts and E-mail, the bonded T1 lines, some of the terminal servers, the massive Cabletron routers and a lot of the telco stuff. One of the tape reels from the Encore hangs on my wall today as a memento. The K250 and the Encore it replaced (as well as the L-Class that later replaced the K250 when I was a consultant) ran an all-singing, all-dancing student information system called CARS. CARS is still around, renamed Jenzabar, though I suspect that many of its underpinnings remain if you look under the table. In those days CARS was a massive overlay that was loaded atop the operating system and database, which when I started were, respectively, HP/UX 10.20 and Informix. (I'm old.) It used Informix tables, screens and stored procedures plus its own text UI libraries to run code written variously as Perform screens, SQL, C-shell scripts and plain old C or ESQL/C. Everything was tracked in RCS using overgrown Makefiles. I had the admin side (resource management, financials, attendance trackers, etc.) and my office partner had the academic side (mostly grades and faculty tracking). My job was to write and maintain this code and shortly after to help the University create custom applications in CARS' brand-spanking new web module, which chose the new hotness in scripting languages, i.e., Perl. Fortuitously I had learned Perl in, appropriately enough, a computational linguistics course. CARS also managed most of the printers on campus except for the few that the RS/6000s controlled directly. Most of the campus admin printers were HP LaserJet 4 units of some derivation equipped with JetDirect cards for networking. These are great warhorse printers, some of the best laser printers HP ever made. I suspect there were line printers other places, but those printers were largely what existed in the University's offices. It turns out that the READY message these printers show on their VFD panels is changeable. I don't remember where I read this, probably idly paging through the manual over a lunch break, but initially the only fun things I could think of to do was to have the printer say hi to my boss when she sent jobs to it, stuff like that (whereupon she would tell me to get back to work). Then it dawned on me: because I had access to the printer spools on the K250, and the spool directories were conveniently named the same as their hostnames, I knew where each and every networked LaserJet on campus was. I was young, rash and motivated. This was a hack I just couldn't resist. It would be even better than what had been my favourite joke at my alma mater, where campus services, notable for posting various service suspension notices, posted one April Fools' Day that gravity itself would be suspended to various buildings. I felt sure this hack would eclipse that too. The plan on April Fools' Day was to get into work at OMG early o'clock and iterate over every entry in the spool, sending it a sequence that would change the READY message to INSERT 5 CENTS. This would cause every networked LaserJet on campus to appear to ask for a nickel before you printed anything. The script was very simple (this is the actual script, I saved it): The ^[ was a literal ASCII 27 ESCape character, and netto was a simple netcat-like script I had written in these days before netcat was widely used. That's it. Now, let me be clear: the printer was still ready! The effect was merely cosmetic! It would still print if you sent jobs to it! Nevertheless, to complete the effect, this message was sent out on the campus-wide administration mailing list (which I also saved): At the end of the day I would reset everything back to READY, smile smugly, and continue with my menial existence. That was the plan. Having sent this out, I fielded a few anxious calls, who laughed uproariously when they realized, and I reset their printers manually afterwards. The people who knew me, knew I was a practical joker, took note of the date, and sent approving replies. One of the best was sent to me later in the day by intercampus mail, printed on their laser printer, with a nickel taped to it. Unfortunately, not everybody on campus knew me, and those who did not not only did not call me, but instead called university administration directly. By 8:30am it was chaos in the main office and this filtered up to the head of HR, who most definitely did know me, and told me I'd better send a retraction before the CFO got in or I was in big trouble. That went wrong also, because my retraction said that campus administration was not considering charging per-page fees when in fact they actually were, so I had to retract it and send a new retraction that didn't call attention to that fact. I also ran the script to reset everything early. Eventually the hubbub finally settled down around noon. Everybody in the office thought it was very funny. Even my boss, who officially disapproved, thought it was somewhat funny. The other thing that went wrong, as if all that weren't enough, was that the director of IT — which is to say, my boss's boss — was away on vacation when all this took place. (Read E-mail remotely? Who does that?) I compounded this situation with the tactical error of going skiing over the coming weekend and part of the next week, most of which I spent snowplowing down the bunny slopes face first, so that he discovered all the angry E-mail in his box without me around to explain myself. (My office partner remembers him coming in wide-eyed asking, "what did he do??") When I returned, it was icier in the office than it had been on the mountain. The assistant director, who thought it was funny, was in trouble for not putting a lid on it, and I was in really big trouble for doing it in the first place. I was appropriately contrite and made various apologies and was an uncharacteristically model employee for an unnaturally long period of time. The Ice Age eventually thawed and the incident was officially dropped except for a "poor judgment" on my next performance review and the satisfaction of what was then considered the best practical joke ever pulled on campus. Indeed, everyone agreed it was much more technically accomplished than the previous award winner, where someone had supposedly gotten it around the grounds that the security guards at the entrance would be charging a nominal admission fee per head. Years later they still said it was legendary. I like to think they still do.
We know what you’re waiting for - this isn’t it. Today, we’re back with more tales of our adventures in Kentico’s Xperience CMS. Due to it’s wide usage, the type of solution, and the types of enterprises using this solution
Mia Sato writing for The Verge: Elon Musk’s $1 Million Handout Winners Are Connected to Republican Causes On Sunday, a few thousand people in Green Bay, Wisconsin, gathered to hear Elon Musk speak — and give away two giant cardboard checks for $1 million. Attendance at the event