Optimized and efficient access to the DRAM is critical to the performance of any chip. As the number of processing elements on a chip increases, so does the demand for data. With DRAM technology transitioning to DDR4 for infrastructure and LPDDR4 for mobile and consumer applications, not only does the frequency of DRAM operation increase significantly, but also the complexity of making the best use of the DRAM bandwidth to deliver high Quality of Service (QoS) at low power, increases. Managing the differing demands of multiple processing elements while delivering maximum DRAM bandwidth is the primary challenge addressed by the Dynamic Memory Controller (DMC).
CoreLink DMC-520 Dynamic Memory Controller
- Targeted at applications in server, networking and high-performance computing, using DDR4 and DDR3 memories Supports enterprise-class requirements for high-density DIMMs, error correction codes and reliability with ease-of-use Optimized for highest performance with the CoreLink CCN family
CoreLink DMC-620 Dynamic Memory Controller
- Builds on top of DMC-520 features to provide the best performance with RAS and end-to-end QoS support with CoreLink CMN-600.
- Reduces static pipeline latency by up to 50% compared to DMC-520. Expands support for 3DS DRAM with extended virtual rank support.
- Improves memory access latency for Arm Cortex v8-A processors significantly under common operating conditions.
Documents and blogs that will help users design Arm-based SoCs
Arm training courses and on-site system-design advisory services enable licensees to realize maximum system performance with lowest risk and fastest time-to-market.Arm training courses Open a support case
|Not answered||CHI protocol cache line states||0 votes||59 views||0 replies||Started 11 hours ago by S_Seth||Answer this|
|Not answered||STM32F769i-Discovery IP Camera Interface||0 votes||81 views||0 replies||Started yesterday by Kiran bhat||Answer this|
|Suggested answer||Store operations where the cache line is already cached (ACE protocol)||2 votes||6200 views||9 replies||Latest yesterday by het||Answer this|
|Not answered||Best most recent text on ARM arch||0 votes||127 views||0 replies||Started 4 days ago by d.ry||Answer this|
|Not answered||Readunique and cleanunique transactions in ACE protocol||0 votes||166 views||0 replies||Started 4 days ago by het||Answer this|
|Suggested answer||Raspberry pi 3 and .net 5 coreclr||1 votes||2076 views||2 replies||Latest 4 days ago by delinaty||Answer this|
|Not answered||CHI protocol cache line states Started 11 hours ago by S_Seth||0 replies 59 views|
|Not answered||STM32F769i-Discovery IP Camera Interface Started yesterday by Kiran bhat||0 replies 81 views|
|Suggested answer||Store operations where the cache line is already cached (ACE protocol) Latest yesterday by het||9 replies 6200 views|
|Not answered||Best most recent text on ARM arch Started 4 days ago by d.ry||0 replies 127 views|
|Not answered||Readunique and cleanunique transactions in ACE protocol Started 4 days ago by het||0 replies 166 views|
|Suggested answer||Raspberry pi 3 and .net 5 coreclr Latest 4 days ago by delinaty||2 replies 2076 views|