Create

Sign in to ReadmeX

or

How we handle your data: Privacy Policy

ReadmeX

Your AI reading companion

What would you like to catch up on?

Pick a briefing, or ask your own question.

Your community and people briefings stay personal to you.

Story

Phison details AI compute and mixed-media SSD designs

AI summary

At Supercomputer 2026, Phison detailed two enterprise SSD designs. One pairs a built-in NPU with 64GB of memory at 130 GB/s and supports running PyTorch, vLLM, OpenVINO and llama.cpp directly on the drive. The other combines pSLC cache with high-capacity QLC behind one controller. Phison said many of the products and features discussed remain in development, and no release dates were provided. The company said developers can use the computational-storage design without specialized programming tools, but did not disclose how a host accesses the drive's compute resources. Phison also cited 36,800 DMIPS for the integrated NPU; the report noted that DMIPS is a CPU integer benchmark, does not establish AI performance and likely refers to accompanying CPU cores. The memory type was not disclosed.

In-depth insight

The computational-storage design is intended to run AI code on data where it resides and return results, reducing the need to transfer whole datasets. The mixed-media design places cache and capacity within each drive, which may simplify system configurations and use drive bays more efficiently. Performance, host-access details and release timing remain unresolved.

Phison detailed two new enterprise SSD designs at Supercomputer 2026 (SC26) that aim to answer two separate problems: one with a built-in NPU and 64GB of 130 GB/s memory that supports running PyTorch, vLLM, OpenVINO, and llama.cpp directly on the drive, and another that merges fast pSLC cache with large QLC capacity behind just one controller. The company notes that “many of the products and features mentioned are still in development” and will arrive once finalized.

The computational-storage design’s premise is that a server can offload more to the SSD, which runs AI code on the data where it already resides and sends back results, instead of whole datasets. The pSLC-QLC design, meanwhile, puts the cache inside each drive, which may reduce system complexity, as every drive bay contributes usable capacity and none goes to dedicated cache drives.

Phison says that the integrated NPU is expected to deliver “36,800 DMIPS of compute performance,” or Dhrystone Million Instructions Per Second, a typical synthetic benchmark. This is admittedly an old CPU integer benchmark from the 1980s, while NPUs tend to be rated in TOPS, but that figure was not given. For reference, the Topaz NPU add-in card Phison showed at Computex carried eight NPUs rated at up to 40 TOPS each. The DMIPS figure therefore likely refers to, instead, the CPU cores that work alongside it, which we would expect to be based on an Arm Cortex design, adequate for running an SSD of this class but no measure of its AI capability.

#Phison#PyTorch#vLLM#OpenVINO#llama.cpp#X3#Topaz NPU

Original · 8 min readRead the original
Group chat

No comments yet. Start the conversation.