Intel has finally revealed its next-gen AI Accelerator, the Gaudi 3, based on a 5nm process node and competing directly against NVIDIA's H100 GPUs.
22.03.2024 - 15:19 / wccftech.com / Hassan Mujtaba
NVIDIA's Blackwell B200 GPUs incorporate a brand new architecture compared to Hopper but also consume almost twice as much power.
When NVIDIA's CEO, Jensen Huang, announced Blackwell during the GTC 2024 keynote, the reveal lacked a lot of technical and architectural information. But during the next few days of GTC, NVIDIA shared slightly more details but still without going too much into the technical deep-dives that we are all awaiting. The new details were revealed by Jonah Albe (NVIDIA SVP & GPU Architect) and Ian Buck (NVIDIA VP of Hyperscale & HPC).
To start, we all knew that Blackwell was going to be a major architectural upgrade over Hopper & it looks like it's more than that with Jonah stating that Blackwell uses a completely different micro-architecture than Hopper.
What we do know about Blackwell is that it packs the 2nd Generation of Transformer Engine technology which adds FP4 and FP6 compute formats. These formats and new software optimizations are what make Blackwell the fastest AI chip of its kind on the planet but that has taken a toll on its standard FP64 compute which has only increased by 32% versus hopper. The reasoning is plain and simple, Blackwell is an AI chip first and that's its main target market. FP64 is not that important from an AI perspective and the lower you go, the faster the inferencing and training capabilities.
Also, the reason to go the chiplet (MCM) route happens to be the need to improve overall performance rather than improving the yields. It will be interesting to see how NVIDIA's first MCM approach works in the field since we are talking about two GPUs running on the same package. It's mentioned that CUDA does a fairly good job in handling the two GPUs & the different architecture, requiring no major changes to be made for programmers.
During the launch, there was a particularly big confusion surrounding all the Blackwell GPU and platform variants. Jensen stated that Blackwell isn't a GPU, it's an entire platform & the platform has a range of products but they are still based on GPUs. As of right now, NVIDIA has announced three official Blackwell GPU variants.
These include the flagship and full-spec B200 which is being used by the GB200 Superchip platforms. This chip has the highest-rated computing capabilities and has a maximum TDP of 1200W. This is 500 Watts more than the Hopper H100 which featured a 700W TDP. The entire
Intel has finally revealed its next-gen AI Accelerator, the Gaudi 3, based on a 5nm process node and competing directly against NVIDIA's H100 GPUs.
NVIDIA's next-gen GeForce RTX 5090 & RTX 5080 "Blackwell" GPUs are rumored to launch in the fourth quarter of 2024.
NVIDIA's board partners are reportedly increasing the prices of various GeForce RTX 40 & RTX 30 GPUs in China which is a stark contrast to what's happening in the US markets.
AMD's Zen 5 CPU core architecture might be shaping up to be a huge upgrade over the existing Zen 4 core as per a new rumor.
Intel's next-gen Arc Battlemage "Xe2-HPG" GPUs for gaming graphics cards have been confirmed within the latest shipment manifesto leaking spree by Momomo_US.
A few weeks back, PGL announced its hardware of choice for its upcoming CS2 Major Tournament which included systems with AMD Ryzen 7 7800X3D CPUs and NVIDIA GeForce RTX 4080 GPUs but it looks like things didn't go as planned as a driver crash associated with the GPU became the very reason of one team's chances of going into the playoffs being washed away.
Intel has just released its latest MLPerf v4.0 performance figures covering the Gaudi 2 Accelerators & 5th Gen Xeon "Emerald Rapids" CPUs, with the former showcasing strong performance per dollar values against NVIDIA's H100 GPU.
NVIDIA continues to push the AI envelope with its strong TensorRT-LLM suite, boosting the H200 GPUs to new heights in the latest MLPerf v4.0 results.
Spring is here and it's time for some GPU deals on some of the most popular AMD, NVIDIA & Intel GPUs out there for gamers.
NVIDIA has reportedly informed its partners that the supply of its existing GeForce RTX 40 "Ada" family will be reduced significantly moving forward as the company preps for RTX 50 "Blackwell" launch in the gaming segment.
NVIDIA is expecting a steady CoWoS packaging supply for its Blackwell AI GPUs as its CEO sees optimism in the supply chain.
AMD has announced its latest FidelityFX Super Resolution technology update in the form of FSR 3.1 which will enable Frame generation on 3rd party upscaling solutions.