Large language models and the applications they power enable unprecedented opportunities for organizations to get deeper insights from their data reservoirs and to build entirely new classes of applications.
But with opportunities often come challenges.
Both on premises and in the cloud, applications that are expected to run in real time place significant demands on data center infrastructure to simultaneously deliver high throughput and low latency with one platform investment.
To drive continuous performance improvements and improve the return on infrastructure investments, NVIDIA regularly optimizes the state-of-the-art community models, including Meta's Llama, Google's Gemma, Microsoft's Phi and our own NVLM-D-72B, released just a few weeks ago.
Relentless Improvements Performance improvements let our customers and partners serve more complex models and reduce the needed infrastructure to host them. NVIDIA optimizes performance at every layer of the technology stack, including TensorRT-LLM, a purpose-built library to deliver state-of-the-art performance on the latest LLMs. With improvements to the open-source Llama 70B model, which delivers very high accuracy, we've already improved minimum latency performance by 3.5x in less than a year.
We're constantly improving our platform performance and regularly publish performance updates. Each week, improvements to NVIDIA software libraries are published, allowing customers to get more from the very same GPUs. For example, in just a few months' time, we've improved our low-latency Llama 70B performance by 3.5x.
NVIDIA has increased performance on the Llama 70B model by 3.5x. In the most recent round of MLPerf Inference 4.1, we made our first-ever submission with the Blackwell platform. It delivered 4x more performance than the previous generation.
This submission was also the first-ever MLPerf submission to use FP4 precision. Narrower precision formats, like FP4, reduces memory footprint and memory traffic, and also boost computational throughput. The process takes advantage of Blackwell's second-generation Transformer Engine, and with advanced quantization techniques that are part of TensorRT Model Optimizer, the Blackwell submission met the strict accuracy targets of the MLPerf benchmark.
Blackwell B200 delivers up to 4x more performance versus previous generation on MLPerf Inference v4.1's Llama 2 70B workload. Improvements in Blackwell haven't stopped the continued acceleration of Hopper. In the last year, Hopper performance has increased 3.4x in MLPerf on H100 thanks to regular software advancements. This means that NVIDIA's peak performance today, on Blackwell, is 10x faster than it was just one year ago on Hopper.
These results track progress on the MLPerf Inference Llama 2 70B Offline scenario over the past year. Our ongoing work is incorporated into TensorRT-LLM, a purpose-built library to accelerate LLMs that contain state-of-the-art optimizations to perform inference efficiently on NVIDIA GPUs. TensorRT-LLM is built on top of the TensorRT Deep Learning Inference library and leverages much of TensorRT's deep learning optimizations with additional LLM-specific improvements.
Improving Llama in Leaps and Bounds More recently, we've continued optimizing variants of Meta's Llama models, including versions 3.1 and 3.2 as well as model sizes 70B and the biggest model, 405B. These optimizations include custom quantization recipes, as well as efficient use of parallelization techniques to more efficiently split the model across multiple GPUs, leveraging NVIDIA NVLink and NVSwitch interconnect technologies. Cutting-edge LLMs like Llama 3.1 405B are very demanding and require the combined performance of multiple state-of-the-art GPUs for fast responses.
Parallelism techniques require a hardware platform with a robust GPU-to-GPU interconnect fabric to get maximum performance and avoid communication bottlenecks. Each NVIDIA H200 Tensor Core GPU features fourth-generation NVLink, which provides a whopping 900GB/s of GPU-to-GPU bandwidth. Every eight-GPU HGX H200 platform also ships with four NVLink Switches, enabling every H200 GPU to communicate with any other H200 GPU at 900GB/s, simultaneously.
Many LLM deployments use parallelism over choosing to keep the workload on a single GPU, which can have compute bottlenecks. LLMs seek to balance low latency and high throughput, with the optimal parallelization technique depending on application requirements.
For instance, if lowest latency is the priority, tensor parallelism is critical, as the combined compute performance of multiple GPUs can be used to serve tokens to users more quickly. However, for use cases where peak throughput across all users is prioritized, pipeline parallelism can efficiently boost overall server throughput.
The table below shows that tensor parallelism can deliver over 5x more throughput in minimum latency scenarios, whereas pipeline parallelism brings 50% more performance for maximum throughput use cases.
For production deployments that seek to maximize throughput within a given latency budget, a platform needs to provide the ability to effectively combine both techniques like in TensorRT-LLM.
Read the technical blog on boosting Llama 3.1 405B throughput to learn more about these techniques.
Different scenarios have different requirements, and parallelism techniques bring optimal performance for each of these scenarios. The Virtuous Cycle Over the lifecycle of our architectures, we deliver significant performance gains from ongoing software tuning and optimization. These improvements translate into additional value for customers who train and deploy on our platforms. They're able to create more capable models and applications and deploy their existing models using less infrastructure, enhancing th
Most recent headlines
05/01/2027
Worlds first 802.15.4ab-UWB chip verified by Calterah and Rohde & Schwarz to be ...
04/08/2026
Dalet, a leading technology and service provider for media-rich organizations, t...
04/07/2026
April 7 2026, 19:00 (PDT) Detective Conan: Fallen Angel of the Highway Opens in...
01/06/2026
January 6 2026, 05:30 (PST) Dolby Sets the New Standard for Premium Entertainment at CES 2026
Throughout the week, Dolby brings to life the latest innovatio...
12/05/2026
aconnic AG (ISIN: DE000A0LBKW6), Munich, has developed a modified fund raising p...
12/05/2026
Share
Copy link
Facebook
X
Linkedin
Bluesky
Email...
12/05/2026
Tyrell Corporation, specialists in high-end live sports and entertainment broadcasts, was tasked with delivering compelling broadcast coverage of premier equest...
12/05/2026
Registration is now open for IBC2026 as the global media, entertainment and technology community prepares to converge on the RAI Amsterdam from 11 14 September ...
12/05/2026
Ross Video, a global leader in live video production technology, will present its latest innovations and integrated production workflows at BroadcastAsia 2026, ...
12/05/2026
500 selected leaders from around the world across start-ups, corporates, and venture capital. Over 50bn in assets under management among attending investors, a...
12/05/2026
CVP, one of Europe's leading suppliers of professional video and broadcast solutions, has appointed Randy Koeman as Sales Manager, Netherlands, marking anot...
12/05/2026
Taiwan Television has expanded its HD content ingest workflow capabilities by investing in the PlayBox Neo Suite - a leading multi-channel multi-server UHD/HD/S...
12/05/2026
Jigsaw24 is strengthening its media team with the appointment of Tom Laker as Client Director and Ashish Chauhan as Project Manager. These strategic hires bring...
12/05/2026
Broadcast playout leader showcases hybrid IP infrastructure and workflow automation with Videoland deployment as regional benchmark...
12/05/2026
dzjinius, the all-in-one production management platform for the media and entertainment industry, will showcase its centralised approach to production workflows...
12/05/2026
Share
Copy link
Facebook
X
Linkedin
Bluesky
Email...
12/05/2026
Share
Copy link
Facebook
X
Linkedin
Bluesky
Email...
12/05/2026
Share
Copy link
Facebook
X
Linkedin
Bluesky
Email...
12/05/2026
Share
Copy link
Facebook
X
Linkedin
Bluesky
Email...
12/05/2026
NUGEN Audio will debut Version 1.1 of DialogCheck, its intelligent dialog intelligibility and compliance tool, at the 2026 Media Production & Technology Show (S...
12/05/2026
The reactive visuals can adapt to any stage or screen setup and comprise 50 layers of Notch content
Today, Disguise announced that its high performance GX 3 m...
12/05/2026
Snapgrid , Snapbag & Airglow Expand Creative Control for the ARRI Modular LED BarFollowing the launch of the ARRI Omnibar, DoPchoice introduces a dedicated li...
12/05/2026
Back to All News
Macarena Garc a and Carlos Cuevas star in El Acercamiento de l...
12/05/2026
Back to All News
Netflix Unveils Documentary on Norway's 26-Year Journey to...
12/05/2026
Back to All News
The Netflix Effect
Ted Sarandos
co-CEO
Social Impact
12 May 2026
Global
Link copied to clipboard
Download all assets
Ten years ago, Ne...
12/05/2026
This is my news content...
12/05/2026
From finance and procurement to supply chain and manufacturing, specialized AI agents are moving into the enterprise systems where business decisions are made, ...
11/05/2026
New production unit features not only wealth of Ross gear but also Sony cameras, Canon lenses, and a Calrec Argo S audio board...
11/05/2026
Solid State Logic (SSL) has announced TCA Tour, a portable fly-away audio production system built from System T components for broadcast, touring, and live prod...
11/05/2026
Riedel Communications has announced it will serve as Official Connectivity Integ...
11/05/2026
Guntermann and Drunck (G&D) and NETGEAR AV have announced a plugin that automates network configuration for G&D KVM-over-IP deployments on NETGEAR AV Line infra...
11/05/2026
The Famous Group (TFG) and Elite Edge, a production studio specializing in live-action production, show opens, set design, and fabrication for sports and live e...
11/05/2026
The native of S o Paulo, Brazil, will play a crucial role in the broadcaster'...
11/05/2026
Cobalt Digital will exhibit its IPMX/ST 2110 product lineup at BroadcastAsia 2026 (Stand 5D2-4), anchored by the COBALT blueCORE family of standalone signal pro...
11/05/2026
Interra Systems will exhibit at BroadcastAsia 2026 (Stand 5C1-10), demonstrating...
11/05/2026
T-Mobile is serving as a technology partner for the 2026 PGA Championship at Aronimink Golf Club in Newtown Square, Pennsylvania, deploying 5G connectivity and ...
11/05/2026
Vizrt has announced it has been named an Official Technical Supplier of the Eurovision Song Contest 2026, in partnership with host broadcaster ORF. The 70th edi...
11/05/2026
The WNBA and Amazon Web Services (AWS) have announced a multi-year partnership making AWS the Official Cloud and Cloud AI Partner of the WNBA. AWS is also joini...
11/05/2026
Austrian private broadcaster kronehit publicly awarded SLG Broadcast AG and Lawo...
11/05/2026
Behind The Mic provides a roundup of recent news regarding on-air talent, includ...
11/05/2026
As Eurovision fans gear up to celebrate the 70th anniversary of the iconic song ...
11/05/2026
Pela primeira vez na hist ria, o Brasil figura entre os oito maiores mercados de...
11/05/2026
New desktop instrument focused on tactile performance
Bastl Instruments have announced the Kalimba, a new desktop instrument that combines physical modellin...
11/05/2026
Compact sampling workstation gains new features
1010music have announced the blackbox 2, a new version of their compact standalone sampler designed for DAWl...
11/05/2026
Performance-focused sequencer with microtonal control
Berlin-based KOMA Elektronik have announced Monoplex, a new 42HP Eurorack sequencer designed for hands...
11/05/2026
Boasts wireless control from mobile devices
The latest iteration of IK Multimedia's compact amp and effects modelling pedal expands on the capabilities ...
11/05/2026
Spans traditional orchestral and experimental sounds
Sonora Cinematic have recently released the third instalment in their Transfigured Orchestra series, de...
11/05/2026
Our exhibition coverage
Watch all our SUPERBOOTH 2026 video coverage in one place. Check back to this page regularly as we will be updating with more video ...
11/05/2026
You're not meant to look at an Eclipse, but you won't be able to take yo...
Large language models and the applications they power enable unprecedented opportunities for organizations to get deeper insights from their data reservoirs and to build entirely new classes of applications.
