HBM TSV Bus Switching for Bandwidth Scaling and Timing Sync
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing high bandwidth memory (HBM) devices face challenges in increasing bandwidth while maintaining memory array timing, TSV bus timing, and DQ bus timing synchronization, especially at tight timing margins, leading to potential power consumption increases and reduced efficiency.
Innovation Solution
Implementing a new timing parameter tCCDBG and dividing bank groups into multiple sets with separate TSV buses to accommodate increased bandwidth, allowing for relaxed timing schedules and efficient data transmission without raising power consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If bandwidth is increased in HBM devices, then data transmission capacity is improved, but timing synchronization between memory array, TSV bus, and DQ bus becomes difficult to maintain
Solution Approach 1:
The patent divides bank groups into multiple sets (first bank group set, second bank group set, etc.) where each set can be accessed independently through separate TSV buses. This segmentation allows concurrent access to different bank group sets, effectively increasing bandwidth while maintaining timing synchronization within each set. The memory controller can issue commands to different bank group sets without interference, resolving the contradiction between bandwidth and timing synchronization.
Solution Approach 2:
The patent introduces dynamic timing parameters (tCCDBG, tCCDLS, tCCDRS) that can be adjusted based on the access pattern and bank group set configuration. These dynamic timing parameters allow the system to optimize timing synchronization for each specific access scenario, enabling bandwidth scaling while maintaining reliable timing margins across different operating conditions.
2Productivity
If timing margins are reduced to increase bandwidth, then data transmission speed is improved, but power consumption increases
Solution Approach 1:
By segmenting bank groups into multiple sets that can be accessed concurrently, the patent achieves higher effective bandwidth without needing to reduce timing margins on individual access paths. Each bank group set maintains adequate timing margins while the overall system throughput increases through parallel access, thus avoiding the power consumption penalty associated with tight timing margins.
Solution Approach 2:
The patent employs periodic timing cycles (tCCDBG, tCCDLS, tCCDRS) that are optimized for each bank group set access pattern. This periodic structure allows the system to maintain consistent, reliable timing margins while achieving high bandwidth through efficient utilization of multiple access cycles, thereby avoiding the need for reduced timing margins that would increase power consumption.
3Productivity
If multiple bank groups are accessed concurrently to increase bandwidth, then data transmission capacity is improved, but timing synchronization between bank groups becomes complex
Solution Approach 1:
The patent simplifies concurrent access control by segmenting bank groups into distinct sets, where each set can be independently addressed and accessed. The memory controller maintains separate timing parameters for each bank group set, which simplifies the synchronization logic compared to managing timing for all bank groups uniformly. This segmentation reduces timing synchronization complexity while enabling high-capacity concurrent data transmission.
Solution Approach 2:
The patent introduces bank group set identifiers and intermediate timing parameters (tCCDBG, tCCDLS, tCCDRS) that act as mediators between the memory controller and individual bank groups. These intermediaries simplify the control logic by providing a structured way to manage concurrent accesses, reducing the complexity of timing synchronization while maintaining high data transmission capacity.
4Device complexity
If traditional single TSV bus per channel is used, then device structure is simple, but bandwidth scaling is limited
Solution Approach 1:
The patent segments the traditional single TSV bus per channel into multiple TSV buses, with each bus serving a specific bank group set. This segmentation enables bandwidth scaling by allowing concurrent data transmission through multiple TSV buses while keeping each individual bus structure simple and manageable. The overall device structure remains relatively simple due to the modular organization of bank group sets and their corresponding TSV buses.
Solution Approach 2:
The patent transitions from a single-dimension bus architecture (one TSV bus per channel) to a multi-dimensional architecture where bank groups are organized in sets that can be accessed through multiple TSV buses simultaneously. This dimensional change enables bandwidth scaling without proportionally increasing device complexity, as the additional TSV buses are organized in a structured, manageable manner through the bank group set hierarchy.
Data Source
AI summary
A system-in-package (SiP) device that includes a base substrate and a processing unit. The SiP also includes a high bandwidth memory (HBM) device that is electrically coupled to the processing unit. The HBM device includes a plurality of bank group sets associated with a same channel or a same pseudo channel of the HBM device, where each bank group set includes one or more bank groups with each bank group having one or more banks with memory arrays. The HBM device includes a plurality of TSV buses, where each TSV bus is associated with a respective bank group set. The HBM device also includes a DQ bus and a bus switching circuit configured to select a TSV bus from the plurality of TSV buses and communicatively couple the DQ bus to the selected TSV bus based on a command from a host device.


