Thursday, December 11th 2025

NVIDIA Reaffirms Support for FP64, Next-Gen GPU to Bring HPC Improvements

NVIDIA's recent generations of GPUs, such as "Hopper" and "Blackwell," have been known for stalling their high-performance computing (HPC) oriented FP64 double-precision performance. However, the company is not inherently abandoning the HPC space and has confirmed for HPCWire that 64-bit floating point data is not going away. According to Dion Harris, the senior director of HPC and AI hyperscale infrastructure solutions at NVIDIA, the company is "definitely looking to bring some additional [FP64] capabilities in our future gen architectures. We are very serious about making sure that we can deliver the required performance to power those simulation workloads."

The acceleration of 64-bit floating-point data paths is crucial for the HPC community, particularly in the life sciences. When a workload demands sustained high-precision support, NVIDIA's recent generations have not met expectations. For comparison, NVIDIA's current most powerful B300 "Blackwell Ultra" accelerator achieves only 1.2 TeraFLOPS of FP64 performance. In contrast, the older H200 "Hopper" reaches an impressive 34 TeraFLOPS of FP64 compute at its peak. For FP8 low-precision, the B300 delivers 9 PetaFLOPS, while the H200 provides 3.958 PetaFLOPS.
This clearly indicates that NVIDIA has been focusing on optimizing for lower precision, beneficial for AI training and inference. For a while, the HPC community has been overlooked, forcing them to seek other brands for the desired FP64 performance. However, NVIDIA plans to enhance HPC performance in future accelerators. We might see NVIDIA walk down the same path as its closest competitor—AMD. In contrast, AMD's Instinct MI accelerator lineup has created distinct GPU series, with one enhancing AI performance and another boosting HPC performance. NVIDIA may be adopting a similar strategy, potentially developing GPU lineups optimized for low-precision AI and others focused on high-precision HPC performance.
Source: HPCWire
Add your own comment

3 Comments on NVIDIA Reaffirms Support for FP64, Next-Gen GPU to Bring HPC Improvements

#1
LastDudeALive
There's been a lot of major HPC contracts from the USFG and various European govs lately that have all gone to AMD, because their latest offerings have FP64. Nvidia's recognizing that HPC is still valuable. AI has its uses, but it's useless if there isn't high quality data for it to train on. And that only comes from HPC.
Posted on Reply
#2
john_
I guess Nvidia doesn't worry only about Instinct 450X, but also about instinct 430X.
Posted on Reply
#3
RejZoR
LastDudeALiveThere's been a lot of major HPC contracts from the USFG and various European govs lately that have all gone to AMD, because their latest offerings have FP64. Nvidia's recognizing that HPC is still valuable. AI has its uses, but it's useless if there isn't high quality data for it to train on. And that only comes from HPC.
It also comes from illegally scraping internet for free and pretending your Ai business can't exist without that...
Posted on Reply
Nov 18th, 2026 13:14 PST change timezone

New Forum Posts

Popular Reviews

TPU on YouTube

Controversial News Posts