
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 ...
17/11/2026
Fabric Insights and Fonn Group have announced a partnership to connect editorial...
17/11/2026
The National Association of Black Journalists is offering up a Thanksgiving Eve treat on Wednesday, Nov. 25, in the form of a live streaming event featuring for...
17/11/2026
Bridge Technologies has announced that its VB440 production probe now supports both monitoring and signal generation for Apple ProRes video streams. The additio...
17/11/2026
Canon U.S.A., Inc. recently welcomed 21 Jericho High School students to its Melv...
17/11/2026
Spectrum News has deployed the Amagi platform to support its live news operations. Amagi says the platform lets Spectrum News streamline newsroom workflows, red...
17/11/2026
Microsoft has expanded its use of Haivision's live video contribution ecosys...
17/11/2026
Setanta Sports has deployed Ateme's TITAN Live platform to support its OTT video delivery operations. Ateme's transcoding technology is integrated into ...
17/11/2026
Media Links has joined the KAIROS Alliance Partner program for Panasonic's KAIROS IT/IP live video production platform. The company completed interoperabili...
17/11/2026
Behind The Mic provides a roundup of recent news regarding on-air talent, including new deals, departures, and assignments compiled from press releases and repo...
17/11/2026
In live sports and in life, things go by in the blink of an eye. Savoring the mo...
17/11/2026
Satellite operators are adding a new generation of low-Earth-orbit (LEO) network...
17/11/2026
Clear-Com has delivered five customized test systems for electromagnetic securit...
17/11/2026
Stephen Arnold Music (SAM) has created the sonic identity for WTGL, the women's team golf league from TMRW Sports, created in partnership with the LPGA and ...
17/11/2026
The Professional Audio Manufacturers Alliance (PAMA) has introduced Give Sound a Voice, an initiative aimed at raising awareness of the value of sound among the...
17/11/2026
Fanatics Studios, the sports entertainment joint venture between Fanatics and OBB, has named Jamal Henderson its first President of Film & TV, announced Michael...
17/11/2026
The NFL's Thanksgiving Week games will feature halftime performances by U2, ...
17/11/2026
FanSided, a Minute brand with more than 250 fan communities, has launched a redesigned mobile app that combines team news, analysis, live game coverage, stats a...
17/11/2026
Major League Volleyball (MLV), the longest-running indoor professional volleybal...
17/11/2026
Prime Video will present a 15-hour slate of live sports on Black Friday, Nov. 27...
17/11/2026
Catherine Dior was more than a muse. A Resistance fighter, a survivor of deporta...
17/11/2026
Multiband dynamics plug-in boasts FT1 integration
Freqport's latest plug-in combines multiband upward and downward dynamics with the company's renow...
17/11/2026
Hi-fi experts launch two pro-audio models
Berlin-based hi-fi experts Teufel Audio have recently branched out into the pro-audio world with their first pairs...
17/11/2026
A year of ComposerCloud+ for $99
EastWest's annual Black Friday Sale is already live, and new users are being offered one year of the all-encompassing C...
17/11/2026
Pick your favourite gear from 2026
The annual SOS Awards are now underway, and as always, we're asking readers to help us choose the best pro-audio rele...
17/11/2026
As part of an ongoing series production program for a German automaker, production capacity has been increased at SGL Carbon's Austrian site in Ried im Innk...
17/11/2026
eds3_5_jq(document).ready(function($) { $(#eds_sliderM519).chameleonSlider_2_1({...
17/11/2026
Signiant has acquired the media intelligence technology, Ceivo, developed by Burnlab with founder Aaron Sloman joining the company as Vice President, Product Ma...
17/11/2026
Share
{alert(Link copied)}).catch(err => {console.error(Failed to copy: , err)}) data-platform=copy_link data-action= class=font-ui-heading py-2 text-sm tex...
17/11/2026
Share
{alert(Link copied)}).catch(err => {console.error(Failed to copy: , err)}) data-platform=copy_link data-action= class=font-ui-heading py-2 text-sm tex...
17/11/2026
Share
{alert(Link copied)}).catch(err => {console.error(Failed to copy: , err)}) data-platform=copy_link data-action= class=font-ui-heading py-2 text-sm tex...
17/11/2026
Share
{alert(Link copied)}).catch(err => {console.error(Failed to copy: , err)}) data-platform=copy_link data-action= class=font-ui-heading py-2 text-sm tex...
17/11/2026
Berklees 2027 Alumni Grammy Nominees: See the Full List From pop and R&B to jazz, musical theater, and video game scores, Berklee talent is represented across...
17/11/2026
Share
{alert(Link copied)}).catch(err => {console.error(Failed to copy: , err)}) data-platform=copy_link data-action= class=font-ui-heading py-2 text-sm tex...
17/11/2026
Share
{alert(Link copied)}).catch(err => {console.error(Failed to copy: , err)}) data-platform=copy_link data-action= class=font-ui-heading py-2 text-sm tex...
17/11/2026
LA JOLLA, CA-Two Scripps Research faculty, Assistant Professor Christian Diercks...
17/11/2026
Our Biggest Sale of the Year on Ivory 3!
From November 17th through November 30th, enjoy our best savings of the year on Ivory 3 Products!
Whether it's ou...
17/11/2026
Tuesday 17 November 2026
Sky announces major free upgrade to the Sky Glass and ...
17/11/2026
Tuesday 17 November 2026
Passion, Betrayal and Scandal: Sky Arts Reveals the Ma...
17/11/2026
Exact 3D Video, a web service based in Japan, has launched its English version, ...
17/11/2026
Today, Apple Podcasts named Baby, this is Keke Palmer its Show of the Year, hono...
17/11/2026
NVIDIA's Ian Buck kicked off the company's week at SC26 in Chicago on Monday with new tools for AI agents, software that can plan experiments, run code ...
16/11/2026
The ESPN Edge Innovation Conference once again gave key ESPN technology partners as a chance to take to the stage and discuss the future of sports fan experienc...
16/11/2026
As rights holders rethink who tells their stories and where, IMG is preparing th...
16/11/2026
From the sidelines, sonic savants advocate for greater emphasis on sound as sports and entertainment increasingly merge
SVG Live!, a one-day event on Nov. 10 i...
16/11/2026
Deal combines Yamaha's audio technology with Disguise's visual-productio...
16/11/2026
Creative effects designed to encourage experimentation
Berry Effects is the brainchild of Abe and Ralph, a pair of multi-instrumentalists and audio engineer...
16/11/2026
Physics-based synth gets first major update
Anukari have just released a new version of their innovative physics-based software synth, marking the instrumen...
16/11/2026
Both versions kitted out with new tools & improvements
Hit'n'Mix have just launched the latest major version of their innovative DAW software, with ...
16/11/2026
Dynamic bands, unmasking features & new stereo tools
Launched in 2021, Eventide's SplitEQ plug-in introduced an interesting new approach to EQ, with the...