Ashish Banga

ASIC vs FPGA – A Friendly Conversation Rahul: Bro, I’m confused! Should I go for ASIC or FPGA for my next project? Amit: Haha! That’s the classic debate. What’s your priority? Performance & cost or flexibility & quick prototyping? Rahul: I need high performance and low power consumption. Amit: Then ASIC is the way! Once fabricated, it’s super optimized. But remember, development takes time and is costly. Rahul: Hmm… What about FPGA? Amit: FPGA is like LEGO! 🧱 You can reprogram it, perfect for prototyping and low-volume production. But it consumes more power and is not as fast as ASIC. Rahul: Got it! So ASIC for mass production and FPGA for flexibility & development. Amit: Exactly! Both have their place in chip design. It depends on what you need. System-on-Chip (SoC): The Brain Behind Modern Tech! 💡 In today's fast-paced digital world, System-on-Chip (SoC) is powering almost everything around us - from smartphones and AI devices to automotive systems and IoT gadgets. 🔹 What is SoC? It’s a single chip that integrates CPU, memory, GPU, AI accelerators, and other essential components, making devices more compact, power-efficient, and high-performing. 🔹 Where is SoC used? Smartphones 📱 AI & Machine Learning 🤖 Automotive Electronics 🚗 Smart Wearables ⌚ IoT & Edge Computing 🌎 With the rise of AI, 5G, and autonomous systems, SoCs are redefining the future of modern technology. 🌐 Visit: https://provlogic.com/ 👉 👉 For more information do follow ProV Logic or Prasanthi Chanda #mentor #guidance #training #workshop #vlsi #internships #design #verification #masters #interview #physicaldesign #vlsidesign #coaching #chip #manufacturing #digital #analog #semiconductor

625

9 Comments

Important topics on SystemVerilog (SV) needed in frontend of VLSI (Basics): [0] Difference b/w Verilog and SystemVerilog. Why SV is chosen for Verfication. [1] Difference between Mailbox and Queue. Is it possible to connect Generator and Driver with a Queue instead of Mailbox. [2] How race condition is avoided in Verilog. And what enhancements are present in SystemVerilog to counter the race condition. [3] Difference between clocking block and modport. Why modport is needed. Difference between clocking and program block. [4] Different event regions of SV. Difference between active and reactive region. What is the use of observed region. [5] Execution flow of assertion in the testbench. Difference between sequence and property. What is the role of semaphore. [6] What is the difference between Static and Dynamic testbench. Give examples. Is it always necessary to build a dynamic testbench. [7] Default input and output Skew. What is the need of using them. What is Synchronous driving. What is Sampling. [8] How error injection is provided in SV. What are callbacks and what are the different methods associated with it. [9] What is Class and an Object. Concept of Inheritance, Encapsulation and Polymorphism. [10] What is the difference b/w Shallow Copy and Deep Copy. Concept of Abstract Class, Concrete Class and Interface Class. [11] Difference between Multiple and Multilevel Inheritance. Difference b/w implement and extend. What is Composition and when it is used. [12] What is Singleton Class and what are the different types of design pattern. What function used as friend function in SV. How garbage collection is taking care in SV. [13] What are virtual function and where it is used. What is Virtual class. What are local and global variables. [14] Operator precedence in SV. Difference b/w Struct and Union. What is a linked list. [15] Write the code for driver with asynchronous reset. [16] Different Array reduction methods. Sorting of arrays. What are Packed structure. [17] What is a Diamond Problem and how to avoid it. Difference between static and dynamic casting. Parameterized classes. [18] How to make hierarchial connection with scoreboard from monitor. How to model a out of order transaction and in order transaction scoreboard. [19] Difference b/w bounded and unbounded mailbox. Can we use queue in generator and driver instead of mailbox. [20] What is Solve x before y constraint and when it is used. How to override a constraint. [21] What are Hard and Soft Constraints. What are weighted distribution constraints. [22] What is Pre randomize and Post randomize. What are the effects of constraint on and off mode. [23] Unique and Priority statement in SV. What are verilog full case and Parallel case and how unique and priority statements are solving the issue. [24] What is the need of interface in SV. Can we use physical interface for dynamic testbench and if so how? [25] Difference between for and foreach loop in SV.

167

2 Comments

𝐈𝐧𝐝𝐢𝐚’𝐬 𝐒𝐞𝐦𝐢𝐜𝐨𝐧𝐝𝐮𝐜𝐭𝐨𝐫 𝐇𝐢𝐫𝐢𝐧𝐠 𝐒𝐮𝐫𝐠𝐞 𝐢𝐬 𝐑𝐞𝐝𝐞𝐟𝐢𝐧𝐢𝐧𝐠 𝐂𝐚𝐫𝐞𝐞𝐫𝐬! The semiconductor industry in India is experiencing an unprecedented hiring boom — gradually chipping away at IT placements and opening exciting opportunities for engineers across domains. 🔹 End-to-end hiring: RTL/SoC design, verification, embedded systems, manufacturing (process, packaging, yield). 🔹 Freshers’ salaries: ₹6–12 LPA (vs. ₹3–4 LPA in IT). 🔹 Aggressive campus recruitment from tier-2 & tier-3 colleges, not just IITs. 🔹 Internships: ₹40k–50k/month stipends to build hands-on chip design skills. 🔹 Hiring scale: ~12,000 freshers annually (up from 8,000 last year). 🔹 Career growth: Roles extend up to ₹20–120 LPA for architects & leaders. Fuelled by global supply chain shifts, EVs, telecom, defence, and government initiatives (DLI, PLI, ECMS), this hiring surge is reshaping student aspirations, especially in ECE and core engineering branches. With India projected to hit $103B semiconductor market by 2030, this is the right time for students and professionals to align their skills with VLSI, Embedded Systems, AI/ML hardware, and semiconductor manufacturing. The future is semiconductor — and India is powering ahead as a global hub! https://lnkd.in/gKtU5Gn2 #Semiconductor #VLSI #ChipDesign #EmbeddedSystems #Electronics #SoC #SemiconductorIndia #SemiconductorJobs #SemiconductorTalent #HiringTrends #Careers #ECE #Engineering #SemiconductorEcosystem #MakeInIndia #Innovation #ElectronicsManufacturing #FutureOfWork #AI #ML #Hardware #Manufacturing #IndiaSemiconductorMission

825

26 Comments

******* Multi Mode Multi Corner(MCMM) analysis ******* In ASIC Physical Design implementation, analyzing and fixing Timing violations across all scenarios are always been Challenging.! 🙂 In advanced nodes, with increasing complexity w.r.t Fabrication and design sizes, no. of Timing scenarios are increased significantly. to close Timing across multiple scenarios with only single scenario in PnR implementation is bit difficult (not saying impossible.!) as Timing closure is achieved at the cost of more turn-around time and further impacts on other signoff aspects such as Physical Verification and Reliability (EMIR). by using proper scenario combinations in PnR implementation helps to ease Signoff Timing closure and less turn-around time of overall Signoff closure activities. to understand more about "Multi Mode Multi Corner(MCMM)" and what all are the factors to be considered, Please go through below 👇 post: https://lnkd.in/deGJgbYH #ASIC #MCMM #SignoffClosure #ChipImplemenration #TimingChallenges

164

2 Comments

EMIR Checks in VLSI : As process nodes scale to 7nm, 5nm, and beyond, power delivery challenges become critical. EMIR (Electro-Migration & IR drop) analysis isn’t just a signoff formality—it’s the difference between reliable silicon and expensive respins. 🔹 IR Drop Static IR drop: Caused by average current → impacts DC voltage levels, reducing noise margin. Dynamic IR drop: Caused by instantaneous switching activity → leads to timing failures & setup/hold violations. 🔹 Electro-Migration (EM) High current density displaces metal atoms. Leads to void formation (open circuits) or hillocks (shorts). Reliability degradation follows Black’s Equation → failure rate ~ J^n * exp(-Ea/kT), where J is current density. 🔹 Why EMIR Checks Matter ✅ Ensure power integrity across standard cells & macros ✅ Capture voltage droop under realistic switching scenarios ✅ Validate reliability margins of interconnects & vias ✅ Provide lifetime projections for SoCs under actual workloads 🔹 Best Practices Use vector-based dynamic analysis for realistic IR stress conditions. Perform early in-design EMIR closure (not just signoff). Apply localized decap insertion & grid reinforcement where droop hotspots appear. Correlate RTL activity → gate-level simulation → signoff IR analysis for accuracy. With billion-transistor SoCs, EMIR is no longer just a backend task—it’s a design co-pilot. #VLSI #SoC #Semiconductors #ChipDesign #Reliability #EMIR

64

HW Bytes Post #6: BIRA & BISR in DFT - Memory Repair Essentials I was chatting with Copilot to refresh my memory on BIRA and BISR in DFT this week, and here are some basics I put together along the way:  #DesignForTestability #VLSI #SemiconductorTesting #ChipDesign #ASIC #SoC #ScanChain #BIST #BoundaryScan #ATPG

165

Why should Design Verification Engineers master protocols like AXI, APB, PCIe, USB, etc.? ✅ Design engineers integrate these protocols into SoCs to connect CPUs, memory, and peripherals. Their RTL must meet timing, area, and functionality specs. ✅ Design verification engineers must ensure this RTL implementation is correct, robust, and interoperable under all conditions. This means: Writing protocol-specific testbenches, checkers, and assertions Verifying corner cases, reset behavior, error responses, and throughput Debugging waveform-level transactions and confirming compliance with specs 📌 Who uses what? ARM, Qualcomm, Apple, MediaTek – AXI, APB (for interconnects and memory systems) Intel, AMD, NVIDIA – PCIe (for CPU-GPU and peripheral communication) Synopsys, Cadence, Rambus – USB, DDR, PCIe IP development and verification Broadcom, Marvell – Ethernet, PCIe, USB (for networking chips) Samsung, Micron, SK Hynix – LPDDR, DDR, UFS (for memory protocol stacks) 🔁 Protocol expertise = Real-world, job-ready verification skill. Don't just simulate logic – verify systems. #VLSI #DesignVerification #SystemVerilog #UVM #AXI #PCIe #USB #DDR #SoCDesign #SemiconductorJobs #ChipDesign #HardwareEngineering #RTLVerification #ProtocolVerification #ASIC #FPGA

118

1 Comment

🧩 Why do we still say VLSI and not ULSI? As integration levels increased, we moved from SSI to MSI to LSI… eventually reaching VLSI (Very Large Scale Integration). Technically, the next stage was ULSI (Ultra Large Scale Integration) — but the term never really took off in industry usage. Why? Because VLSI became the accepted standard, covering everything from thousands to billions of transistors. Instead of changing the terminology again and again, the industry chose to evolve under the same name. For students and beginners, it’s a great example of how terms stick not just for accuracy, but for consistency. #VLSI #Semiconductors #ChipDesign #EngineeringStudents #TechInsights #WhyVLSI #ULSI #IntegratedCircuits #LearningJourney

133

2 Comments

Direct Memory Access (DMA) is a critical feature in VLSI design that enhances data transfer efficiency between memory and I/O devices without burdening the CPU. ### What is DMA? DMA allows peripherals to access the main memory independently of the CPU, enabling faster data transfer rates and freeing up CPU resources for other tasks. This is particularly useful in applications requiring high-speed data transfer, such as video processing, audio streaming, and network communications. ### Key Features of DMA: 1. **Efficiency**: Reduces CPU involvement in data transfers, allowing it to perform other computations. 2. **Speed**: Enables higher data transfer rates compared to programmed I/O methods. 3. **Multitasking**: Facilitates concurrent operations, improving overall system performance. ### DMA Operations: Steps of DMA Transfers The DMA operation typically involves several key steps, which can be summarized as follows: 1. **Initialization**: - The CPU configures the DMA controller by setting parameters such as source address, destination address, and the number of bytes to transfer. - The DMA controller is also enabled to begin the transfer process. 2. **Requesting Transfer**: - The peripheral device that needs to transfer data sends a request to the DMA controller (often referred to as a "DMA request"). - Upon receiving the request, the DMA controller checks if the bus is available for data transfer. 3. **Bus Arbitration**: - If multiple devices request access to the bus, the DMA controller arbitrates to determine which device gets access first. - Once granted access, the DMA controller gains control of the system bus. 4. **Data Transfer**: - The DMA controller initiates the data transfer from the source (e.g., peripheral) to the destination (e.g., memory). - The transfer occurs without CPU intervention, allowing the CPU to execute other tasks simultaneously. 5. **Transfer Completion**: - After the specified amount of data has been transferred, the DMA controller sends an interrupt signal to the CPU to indicate that the transfer is complete. - The CPU can then check the status of the transfer and proceed with any required follow-up actions. 6. **Deactivation**: - The DMA controller releases the bus, allowing other devices or the CPU to use it for their operations. - The controller can be reconfigured for subsequent transfers as needed. ### Example Use Case Consider a scenario where a video camera captures frames and needs to transfer them to memory for processing. Using DMA, the following occurs: - The CPU initializes the DMA controller with the frame buffer's starting address and size. - The camera sends a DMA request to transfer the frame data. - The DMA controller gains control of the bus and transfers the frame data directly into memory. - Upon completion, the DMA controller interrupts the CPU, signaling that the frame is ready for processing. #dma #dmacontroller #bus #arbitration #vlsi

61

🚀 Most Asked RTL Question – Convergence in Clock Domain Crossing (CDC) Q: What is convergence? Convergence: Two signals are coming from different paths and getting used in one combinational logic gate. The output is getting synchronized in destination clock domine. Then here glitch may propagated into receiving clock domine, which can sample and cause a functional issue. Scenario : Imagine two signals coming from different clock paths feeding into a single combinational logic gate. The output of this gate eventually goes into a destination clock domain. If the signals are not properly synchronized, a glitch may propagate into the receiving clock domain. This glitch can be sampled by the destination flip-flop, leading to functional issues in your design — a classic CDC hazard. Why It Happens: Signals arrive at slightly different times. Their combination in the combinational logic can produce a short-lived incorrect pulse. The receiving clock domain may capture this pulse, causing unexpected behavior. 💡 Fix / Best Practice: Use synchronizers for each signal before combining. Avoid combining multiple asynchronous signals directly in one logic gate. Consider using handshaking protocols or CDC-aware FIFOs for multi-bit signals. Key Takeaway: Register the signal in the source domain The cleanest fix is to register the combinational signal in its own clock domain before crossing. Convergence issues are subtle but can break your RTL in real silicon, especially in high-speed multi-clock designs. Early CDC checks and careful synchronization are crucial for reliable RTL design. #RTLDesign #VLSI #ClockDomainCrossing #CDC #FIFO #DigitalDesign #ASIC #FPGA #SoCDesign #Verilog #SystemVerilog #DesignVerification #Engineering

92

7 Comments

🎯 Part 3: Compensation Done Right – When Theory Meets Tapeout “Every pole wants to win. Compensation teaches them teamwork.” Designing analog circuits at scale isn’t about “knowing” compensation – it’s about applying it smartly across different architectures, technologies, and PVT corners. Let’s decode each dimension of smart compensation. ⸻ 🔍 1. Where Do Poles Come From? Before you compensate anything, ask: • What are your dominant capacitors? • Who’s contributing slow dynamics? • Where’s the gain drop-off happening? 🧠 Example: • In a two-stage opamp, the 1st pole is at the output of the first stage, the 2nd pole is at the load of the output stage. • If you ignore that, you’ll get oscillations even with 50° phase margin on paper. ⸻ 📐 2. Design Checklist Before Applying Compensation ✅ Estimate open-loop gain and phase ✅ Identify all poles/zeros up to 3 GHz ✅ Run small-signal AC + transient sims ✅ Consider loading (pads, switches, next stages) ✅ Ask: “Can my compensation survive corners + mismatch?” 📎 Golden Rule: Always close the loop in simulation before you add compensation. Uncompensated behavior is your compass. ⸻ 🛠️ 3. Miller Compensation Design Tips • Use a feedback capacitor (Cm) between the output and intermediate node. • Size Cm such that first pole << second pole. • Add a series resistor (nulling resistor) to move the RHP zero back to left-half plane. 📏 Rule of thumb: Cm = 1/gm of 2nd stage → yields decent 45–60° phase margin ⸻ 🔀 4. Nested Miller for 3-stage Design 3-stage opamps often suffer from: • Extra pole • Inter-stage loading • Long tail swing paths 🧩 Solution: • Nest 2 compensation caps: one between stage 2 and 1, another from stage 3 to 1. • Add gm boosters or buffers to manage parasitics. 📌 Best used in: • Pipeline ADC opamps • High-gain regulators with PVT swings ⸻ ⚡ 5. Feedforward is Great, But Use with Caution Feedforward is tempting for: • Slew boost • Fast settling in switched-cap circuits But: • You lose DC gain • It can increase peaking in frequency response 💡 Smart Trick: Use common-mode feedforward — helps in differential architectures without hurting differential-mode gain. ⸻ 🧱 6. Ahuja (Cascode) Buffers Are Clean Fixes When layout symmetry and parasitic matching are critical (think high-resolution SAR ADC drivers), Ahuja saves the day: ✔️ Maintains high gain ✔️ Avoids RHP zero ✔️ Keeps layout predictable 📍 Tip: Place the buffer carefully in layout to minimize mismatch paths. ⸻ 🔁 7. Simulate Your Stability – But Do It Right Too many designers: • Skip phase margin checks in corners • Ignore load-dependent poles • Only simulate at TT, no mismatch corners 🔍 Must check: • AC response • Phase margin vs load cap sweep • Monte Carlo on RC extraction ⸻ 🚧 Common Pitfalls ❌ Overcompensating → slow response ❌ Forgetting layout parasitics → wrong phase margin ❌ Relying only on AC sims → transient may tell a different story

76

4 Comments

We presented an updated device engineering of GCT DRAM device in VLSI 2025 T20-4, where a BE-SONOS charge-trapping gate stack is integrated in the previous GCT (gate-controlled thyristor) so that the device can perform both DRAM-mode and NVM-mode simultaneously. The data transfer from NVM to DRAM mode can be instantly using normal read mode thus we call it "recall free". This kind of "persistent memory" separates the NVM operations from the DRAM operations so that DRAM mode has unlimited endurance in R/W, while NVM still preserves excellent data retention. Thus we does not need to trade-off the materials to balance high endurance with NVM retention. Reference to nvSRAM, this novel device has much higher memory density. Very excellent I-V hysteresis window is provided in GCT for DRAM R/W modes. When injected electrons in BE-SONOS it will eliminate the hysteresis (always read "1"), while injecting holes in BE-SONOS can restore back to normal hysteresis (GCT DRAM mode R/W 0/1).

96

3 Comments

DDR5 DRAM MODULE TECHNOLOGY DDR5 DRAM technology, the fifth generation of Double Data Rate Synchronous Dynamic Random Access Memory, offers significant performance improvements over DDR4 with higher bandwidth, lower power consumption, and enhanced data integrity, including on-die error correction code (ECC). Key Features and Advantages: 1. Increased Bandwidth: DDR5 delivers up to twice the bandwidth compared to DDR4, enabling faster data transfer rates and improved system performance. 2. Lower Power Consumption: DDR5 operates at a lower voltage (1.1V) and incorporates on-module power management integrated circuits (PMICs) for better energy efficiency. 3. Enhanced Data Integrity: DDR5 features on-die ECC, which detects and corrects errors before they reach the CPU, leading to more robust and reliable memory performance. 4. Dual 32-bit Subchannels: DDR5 utilizes two independent 32-bit subchannels on a single module, allowing for increased data transfer efficiency and improved scaling performance. 5. Increased Banks and Burst Length: DDR5 doubles the number of banks (16 to 32) and the minimum burst length (8 to 16), further enhancing data bus efficiency and reducing the number of reads/writes to access the same cache data line. 6. Scalable Memory: DDR5 is designed to support higher capacities and speeds, making it suitable for demanding applications like gaming, content creation, and data centers. 7. Improved Performance: DDR5 offers a significant performance boost for CPUs and platforms that require more cores and memory bandwidth. 8. On-die ECC: All DDR5 chips have on-die error-correction code, that detects and corrects errors before sending data to the CPU, to improve reliability and allow denser RAM chips which lowers per-chip defect rate. 9. Power Management Integrated Circuits (PMIC): Regulates the power required by the various components of the memory module. 

113

2 Comments

Are you interested in understanding bare-metal development for microcontrollers? This mini-cookbook, authored by EmbeddedExpertIO, provides step-by-step instructions for writing bare-metal peripheral drivers for the STM32F4 family of microcontrollers. It covers essential topics like GPIO, timers, ADC, and UART. Each section includes practical code examples tested on the STM32F411 Nucleo . Save it for future use or reference 🤓 #Embedded_C #EmbeddedDev #EmbeddedSoftware

143

7 Comments

🔍 Why Is VLSI Recruitment So Challenging? 🤔 If you've ever tried hiring in the VLSI domain, you know it’s no walk in the park. From sourcing to onboarding, it’s a tough game—and here’s why: 💡 1. Niche Skill Set: VLSI isn't just another tech buzzword—it’s a highly specialized field requiring deep knowledge in areas like digital design, verification, physical design, and semiconductor fundamentals. Not every engineer can make the cut. 📉 2. Limited Talent Pool: The number of engineers trained in VLSI (especially hands-on experience) is much smaller compared to general software roles. Even fewer have experience with industry tools like Synopsys, Cadence, or Mentor. 🏫 3. Gap Between Academia & Industry: Universities often focus on theoretical aspects, while the industry demands practical expertise. This disconnect makes it harder for fresh grads to be job-ready, slowing down the hiring funnel. ⚙️ 4. Long Training Time: Even if you find good candidates, ramping them up takes months. Companies are cautious because hiring the wrong candidate in a chip design role is costly—literally. 📍 5. High Competition: Big players, startups, and fabs alike are all fishing from the same pond. Good candidates are often snapped up before you even get through the interview process. 🎯 What's the way forward? Partner closely with academic institutions for curriculum alignment Offer internships & co-op programs to create a future talent pipeline Invest in upskilling passionate engineers from adjacent domains Streamline and speed up your hiring process Let’s talk about how we can make VLSI hiring more sustainable and scalable. Are you facing the same challenges? Drop your thoughts below! 👇 #VLSI #Semiconductors #HiringChallenges #TalentAcquisition #ChipDesign #EDA #VLSIHiring #Recruitment

35

*** Changing Role of Verification Engineers in a Chiplet World *** In a monolithic SoC world, a verification engineer’s job was mostly: o Ensuring functional correctness of individual IP blocks. o Running full-chip verification across a single die. o Debugging issues inside the same physical domain. But in a chiplet-based SoC, the role expands: o Understanding inter-chiplet protocols and system-wide behavior. o Modeling interactions across dies, not just within a single die. o Debugging failures where the root cause might be in the package, not just the chiplets themselves. Chiplet-based SoCs blur the lines between: o Digital and physical verification—since interconnect and package-level effects are inseparable. o Pre-silicon and post-silicon validation—because real-world behavior isn’t always captured in traditional simulations. o Design and verification responsibilities—since partitioning choices impact how easily a system can be tested. Verification is no longer just about "checking correctness"—it’s about understanding the full system at every level. How do we train the next generation of verification engineers for this shift?

52

2 Comments