Tenstorrent Unveils High-End Wormhole AI Processors, Featuring RISC-V
Tenstorrent launches Wormhole AI chips on RISC-V, emphasizing cost-effectiveness and scalability. Wormhole n150 offers 262 TFLOPS, n300 doubles power with 24 GB GDDR6. Priced from $999, undercutting NVIDIA. New workstations from $1,500.
Read original article
Tenstorrent has introduced its advanced Wormhole high-performance AI chips based on the RISC-V architecture, aiming to offer exceptional price-to-performance value. The CEO, Jim Keller, emphasizes the scalability and cost-effectiveness of the new AI processors, which come in two configurations: Wormhole n150 and n300. The n150 features 72 Tensix cores with 262 TFLOPS of FP8 performance, while the n300 doubles the computing power with 24 GB of GDDR6 memory. Tenstorrent also unveiled dedicated workstations like the TT-QuietBox and TT-LoudBox, incorporating AMD's EPYC and Intel's Xeon processors, respectively. The pricing for the Wormhole chips starts at $999 for the n150 and $1,399 for the n300, significantly undercutting competitors like NVIDIA. The workstations are priced at $1,500 for the TT-QuietBox and $6,000 for the TT-LoudBox. Tenstorrent's new products are positioned to benefit AI startups and individuals seeking cost-effective AI computing solutions with solid performance capabilities.
Related
Intel's Gaudi 3 will cost half the price of Nvidia's H100
Intel's Gaudi 3 AI processor is priced at $15,650, half of Nvidia's H100. Intel aims to compete in the AI market dominated by Nvidia, facing challenges from cloud providers' custom AI processors.
What Would You Do with a 16.8M Core Graph Processing Beast?
TSMC collaborates with Intel, MIT, and AWS to develop HIVE, a graph processing unit named PIUMA with 16.8 million cores. The advanced chip aims to efficiently process neural networks, featuring a custom RISC-based architecture and photonics interconnect for scalability.
Tenstorrent Launches Next-Gen Wormhole-Based Developer Kits and Workstations
Tenstorrent launches new Wormhole™ chip developer kits and workstations for scalable multi-chip development. Products include n150, n300 kits, TT-LoudBox workstation, and TT-QuietBox water-cooled AI workstation. Supported by TT-Buda and TT-Metallium software stacks. CEO Jim Keller excited about expanding product offerings.
Tenstorrent Wormhole Developer Kits Launched
Tenstorrent launches "Wormhole" PCIe developer kits with n150s and n300s cards, Tensix cores, GDDR6 memory, local SRAM caches, and high-speed Ethernet. Priced at $18K, offering cost-saving opportunities for AI developers.
Tenstorrent Wormhole RISC-V Based Workstations for AI Development
Tenstorrent unveils Wormhole dev kits for AI development, featuring royalty-free RISC-V processor ISA and scalable AI accelerators. Kits include n150 with 72 cores at $1,000 and n300 with 128 cores at $1,400. Workstations TT-LoudBox and TT-QuietBox range from $12,000 to $15,000. Future plans include Blackhole architecture for automotive AI.
11 commentsI did some 5 second digging into their code and it seems based on their Github that they took Berkeley's BOOM CPU (written in Chisel), and added support for the Vector 1.0 extension, and then they've called this core Ocelot. Seems like the core is open source at least.
https://github.com/tenstorrent/riscv-ocelot/blob/bobcat/READ...
Not sure about the driver/software situation though. Do you actually get to run whatever RISC-V code you want on these? Could be amazing if so.
How fast are the n300s compared to RTX 4090 or RTX 3090 cards?
At these prices it doesn't seem worthwhile to stick them into your own workstation instead of going for theirs.
A neat feather in RISC-V's cap. Though to be fair, RISC-V microcontrollers on PCI boards is nothing new; Nvidia's shipped their RISC-V Falcon GSP for years, and a lot of recent storage mediums rely on RISC-V microcontrollers.
https://tenstorrent.com/hardware/tt-loudbox
The real comparison for me is against a box using NVIDIA RTX’s. That’s what most use in the sub-$10,000 space for open-source models. They usually use $2,000-$5,000 worth of consumer GPU’s with maybe 24GB VRAM each. So, how does this compare on inference or training to 1-2 RTX’s with 24GB? And what performance for both small and large models?
If a single N300 is $1,399 and a quietbox has four then the quietbox can't be $1,500. Maybe the 1.5 is pricing of box minus cards?
We need more diversity on the hardware side and see what we can build on it.
Related
Intel's Gaudi 3 will cost half the price of Nvidia's H100
Intel's Gaudi 3 AI processor is priced at $15,650, half of Nvidia's H100. Intel aims to compete in the AI market dominated by Nvidia, facing challenges from cloud providers' custom AI processors.
What Would You Do with a 16.8M Core Graph Processing Beast?
TSMC collaborates with Intel, MIT, and AWS to develop HIVE, a graph processing unit named PIUMA with 16.8 million cores. The advanced chip aims to efficiently process neural networks, featuring a custom RISC-based architecture and photonics interconnect for scalability.
Tenstorrent Launches Next-Gen Wormhole-Based Developer Kits and Workstations
Tenstorrent launches new Wormhole™ chip developer kits and workstations for scalable multi-chip development. Products include n150, n300 kits, TT-LoudBox workstation, and TT-QuietBox water-cooled AI workstation. Supported by TT-Buda and TT-Metallium software stacks. CEO Jim Keller excited about expanding product offerings.
Tenstorrent Wormhole Developer Kits Launched
Tenstorrent launches "Wormhole" PCIe developer kits with n150s and n300s cards, Tensix cores, GDDR6 memory, local SRAM caches, and high-speed Ethernet. Priced at $18K, offering cost-saving opportunities for AI developers.
Tenstorrent Wormhole RISC-V Based Workstations for AI Development
Tenstorrent unveils Wormhole dev kits for AI development, featuring royalty-free RISC-V processor ISA and scalable AI accelerators. Kits include n150 with 72 cores at $1,000 and n300 with 128 cores at $1,400. Workstations TT-LoudBox and TT-QuietBox range from $12,000 to $15,000. Future plans include Blackhole architecture for automotive AI.