The Tidal Wave Of Rising GPU TAM Raises All Boats
The world has gone nuts for generative AI, and it is going to get a whole lot crazier. …
The world has gone nuts for generative AI, and it is going to get a whole lot crazier. …
There is nothing quite like great hardware to motivate people to create and tune software to take full advantage of it during a boom time. …
The exorbitant cost of GPU-accelerated systems for training and inference and latest to rush to find gold in mountains of corporate data are combining to exert tectonic forces on the datacenter landscape and push up a new Himalaya range – with Nvidia as its steepest and highest peak. …
In a world where allocations of “Hopper” H100 GPUs coming out of Nvidia’s factories are going out well into 2024, and the allocations for the impending “Antares” MI300X and MI300A GPUs are probably long since spoken for, anyone trying to build a GPU cluster to power a large language model for training or inference has to think outside of the box. …
For very sound technical and economic reasons, processors of all kinds have been overprovisioned on compute and underprovisioned on memory bandwidth – and sometimes memory capacity depending on the device and depending on the workload – for decades. …
Timing is a funny thing. The summer of 2006 when AMD bought GPU maker ATI Technologies for $5.6 billion and took on both Intel in CPUs and Nvidia in GPUs was the same summer when researchers first started figuring out how to offload single-precision floating point math operations from CPUs to Nvidia GPUs to try to accelerate HPC simulation and modeling workloads. …
Because they are in the front of the line for acquiring Nvidia datacenter GPUs, the hyperscalers and cloud builders are going to be the ones who benefit mightily from shortages of matrix math engines that can train AI models and run inference against them. …
If you had to sum up the second half of 2022 and the first half of 2023 from the perspective of the semiconductor industry, it would be that we made too many CPUs for PCs, smartphones, and servers and we didn’t make enough GPUs for the datacenter. …
Transitions in the datacenter take time.
It took Unix servers a decade, from 1985 through 1995, to supplant proprietary minicomputers and a lot of mainframe capacity that would have otherwise been bought. …
Heaven forbid that we take a few days of downtime. When we were not looking – and forcing ourselves to not look at any IT news because we have other things going on – that is the moment when Nvidia decides to put out a financial presentation that embeds a new product roadmap within it. …
All Content Copyright The Next Platform