←Search results

Search “NVIDIA”

By relevance · Total 2010 results Clear search

Thursday

Tuesday

Monday

Sunday

  1. Chips and Cheese

    Hot Chips 2026: XCENA and Samsung's Near-Memory Computing CXL Device MX1

    XCENA partnered with Samsung to launch the CXL memory expansion device MX1, which can host up to 2 TB of DDR5 memory, connects to the host via a PCIe 6/CXL 3.2 x8 interface with 64 GB/s bandwidth in each direction, and additionally provides 8 downstream PCIe 6 lanes for connecting SSDs. The MX1 chip integrates 3072 RISC-V cores, divided into 24 subsystems of 128 cores each, is manufactured on Samsung's 4nm process, draws 40 W for the chip and 90 W for the whole board, and features a subsystem-level custom vector processing engine with FP32/FP16 dot-product throughput of about 3 TFLOPS.

    XCENA partnered with Samsung to launch the CXL memory expansion device MX1, which can host up to 2 TB of DDR5 memory, connects to the host via a PCIe 6/CXL 3.2 x8 interface with 64 GB/s bandwidth in each direction, and additionally provides 8 downstream PCIe 6 lanes for connecting SSDs. The MX1 chip integrates 3072 RISC-V cores, divided into 24 subsystems of 128 cores each, is manufactured on Samsung's 4nm process, draws 40 W for the chip and 90 W for the whole board, and features a subsystem-le

    Matched fieldsBody

    #hot-signal#Product updates#English text is being prepared.#English text is being prepared.

Friday

  3. Ant Ling: Developer BlogSelected

    Ant Ling releases finance-enhanced model Ling-3.0-flash-Fin

    Ant Ling releases its first finance-enhanced model, Ling-3.0-flash-Fin, which continues the Ling-3.0-flash architecture, maintaining 124B total parameters and 5.1B activated parameters. Through continued pre-training on financial corpora, domain post-training, and tool-use optimization, it strengthens the handling of complex financial content such as annual reports and financial workbooks. The model was tested on benchmarks including FinFIRST, FinSearchComp Verified, Finance Agent, APEX-Agents, SpreadsheetBench, and τ³-Banking, with its AA Intelligence Index v4.1.1 score rising from 38 to 41. The model will offer a one-month limited-time free API on OpenRouter, and its weights will be open-sourced next week.

    English text is being prepared.

    Ant Ling releases its first finance-enhanced model, Ling-3.0-flash-Fin, which continues the Ling-3.0-flash architecture, maintaining 124B total parameters and 5.1B activated parameters. Through continued pre-training on financial corpora, domain post-training, and tool-use optimization, it strengthens the handling of complex financial content such as annual reports and financial workbooks. The model was tested on benchmarks including FinFIRST, FinSearchComp Verified, Finance Agent, APEX-Agents,

    Matched fieldsBody

    #engineering#Model releases#Agents#Reasoning

Thursday