中文 English

PCI Express 5.0 Takes Center Stage For Data Centers


The demands on servers at the heart of data centers continue an inexorable rise. Responding to these demands, new platforms keep coming that deliver greater computing performance, have more memory and use faster interconnects. On the way at the end of this year and early 2022 are new server platforms that will take performance to a new level. These new platforms will transition to DDR5 DIMMs fo... » read more

HBM2E Raises The Bar For AI/ML Training


The largest AI/ML neural network training models now exceed an enormous 100 billion parameters. With the rate of growth over the last decade on a 10X annual pace, we’re headed to trillion parameter models in the not-too-distant future. Given the tremendous value that can be derived from AI/ML (it is mission critical to five of six of the top market cap companies in the world), there has been ... » read more

CXL: Sorting Out The Interconnect Soup


In the webinar Hidden Signals: Memory and Interconnect Decisions for AI, IoT and 5G, Shane Rau of IDC and Rambus Fellow Steven Woo discussed how interconnects were a critical enabling technology for future computing platforms. One of the major complications was the “interconnect soup” of numerous and divergent interface protocols. The Compute Express Link (CXL) standard offers to sort out m... » read more

Five Key Changes Coming With DDR5 DIMMs


On July 14th of last year, JEDEC announced the publication of the DDR5 SDRAM standard. This signaled the nearing industry transition to DDR5 server dual-inline memory modules (DIMM). DDR5 memory brings a number of key enhancements that will bring great performance and power benefits in next generation servers. Scaling Data Rates to 6.4 Gb/s You can never have enough memory bandwidth, and DD... » read more

Pushing The Envelope With HBM2E Memory


In September, Rambus announced the achievement of reaching 4 gigabits per second (Gbps) operation with our HBM2E memory interface. This milestone was demonstrated in silicon and required mastering substantial signal integrity and power integrity (SI/PI) challenges. The 4 Gbps mark represents a 20% rise from the previous maximum data rate of 3.2 Gbps for HBM2E. To date, the industry’s faste... » read more

The Expanding Universe Of MIPI Applications


It’s hard to imagine today, but there was a time when mobile phones had no cameras and displays were tiny monochrome LCDs capable of displaying a phone number and not much more. The iconic Nokia 3310 announced Sept. 1, 2000, had an 84 x 48 pixel monochrome display and went on to sell 126 million units worldwide. You may still have one in your junk drawer. By the time of the original iPhone... » read more

GDDR6 Memory For Life On The Edge


With the torrid growth in data traffic, it is unsurprising that the number of hyperscale data centers has grown apace. According to analysts at the Synergy Research Group, in July of this year there were 541 hyperscale data centers worldwide. That represents a doubling in the number since 2015. Even more striking, there are an additional 176 in the pipeline, so the breakneck growth in hyperscal... » read more

PCIe 5.0: A Key Interface Solution For The Evolving Data Center


A great many developments are shaping the evolution of the data center. Enterprise workloads are increasingly shifting to the cloud, whether these be hosted or colocation implementations. The nature of workload traffic is changing such that data centers are architected to manage greater east-west (within the data center) communication. New workloads, with AI/ML (artificial intelligence/machine ... » read more

Scaling AI/ML Training Performance With HBM2E Memory


In my April SemiEngineering Low Power-High Performance blog, I wrote: “Today, AI/ML neural network training models can exceed 10 billion parameters, soon it will be over 100 billion.” “Soon” didn’t take long to arrive. At the end of May, OpenAI unveiled a new 175-billion parameter GPT-3 language model. This represented a more that 100X jump over the size of GPT-2’s 1.5 billion param... » read more

An Expanding Application Space For GDDR6 Memory


The origins of graphics double data rate (GDDR) memory can be traced to the rise of 3D gaming on PCs and consoles. The first GPUs used single data rate (SDR) and double data rate (DDR) DRAM, the same memory used for CPU main memory. The quest for higher frame rates at higher resolutions drove the need for a graphics-workload specific memory solution. The commercial success of gaming PCs and con... » read more

← Older posts