● Covered by 1 source · 1 reportMedium impact1 neutral
Nvidia to Extend CUDA Support to RISC-V, Outlining Server-Grade Requirements
Key points
- Nvidia is extending CUDA support to RISC-V CPUs.
- RISC-V CPUs must meet server-grade specifications for CUDA integration.
- Requirements include RVA23 CPU, RAS features, and a security processor.
- ACPI support and PCIe coherency are also critical for CUDA on RISC-V.
CUDA Expansion to RISC-V
Nvidia is expanding its CUDA platform, which is widely used for GPU compute and machine learning, to include support for RISC-V CPUs. Currently, CUDA supports x86-64 and aarch64 architectures. This move will allow RISC-V processors to feed GPU compute tasks, broadening the ecosystem for Nvidia's GPU technology.
Server-Grade CPU Requirements
For CUDA integration, Nvidia requires RISC-V CPUs to meet server-grade specifications. This includes a RVA23 CPU and adherence to RISC-V's server SoC and server platform specifications. These specifications encompass features such as reliability, availability, and serviceability (RAS), along with a specialized security processor and other baseline functionalities essential for server environments.
Beyond Standard Specifications
Nvidia has additional requirements that extend beyond the standard RISC-V profile or platform specifications. These are necessary to ensure optimal CUDA software performance and prevent inefficiencies. The company emphasizes the need for performance-enhancing extensions, such as vector extensions with predication support, to avoid branches and ship efficient code. Without these, Nvidia would be forced to use less efficient code paths.
ACPI and PCIe Coherency
ACPI (Advanced Configuration and Power Interface) is another critical requirement. ACPI allows software to discover hardware capabilities and manage power, performance, and thermal aspects. While initial RISC-V hardware lacked ACPI, this issue has been resolved with the UEFI forum adding RISC-V ACPI support in 2025 and the ratification of the RISC-V BRS (Boot and Runtime Services) specification, which includes ACPI, last year.
Nvidia also requires PCIe coherency to prevent memory ordering problems. Without it, data inconsistencies can arise when the CPU writes data that remains in a cache, and a DMA request copies stale data from DRAM to the GPU. Similarly, when results are copied back from the GPU, the CPU could read stale data from its caches. PCIe coherency is considered a standard feature in server CPUs to avoid complex software cache invalidations.
✨ This summary was generated by AI from the outlets' reporting listed below. It is not independently verified and may contain errors — check the original sources. How BrevFeed works →
The daily brief
One email each morning: the day's tech stories, clustered across outlets and summarized. No account needed.
One email a day. Unsubscribe in one click, any time.
Today's brief
Spend a few minutes, get the whole day. Every topic's top stories in one hands-free rundown — listen, watch, or read the transcript.
▶ Play today's brief
New every morning, and the back catalogue is archived by date.
Primary sources
GitHub riscv-non-isa/riscv-server-platformReporting from
Nvidia plans to extend its CUDA platform to support RISC-V CPUs, enabling RISC-V processors to integrate with GPU compute applications. This expansion requires RISC-V CPUs to meet specific server-grade criteria, including adherence to server SoC and platform specifications, ACPI support, and PCIe coherency.