An Analysis of Machine Learning Hardware from Tenstorrent
Bibliographic record
Abstract
Machine learning is growing at an exponential rate in industry today, with contemporary machine learning models having parameter counts in the high hundred millions to trillions paired with a huge computing requirement. As the field continues to grow and the models get larger and more computationally expensive, there is a greater need for specialized hardware that can accelerate the inference and training computations that are required of these machine learning models. The design of a custom hardware device specifically made to accelerate machine learning models is very intricate, with software and hardware development that needs to work together at the cutting edge. The following paper will cover these hardware accelerators and focus specifically on the Grayskull e150, a chip optimized for machine learning inference developed by Tenstorrent. Tenstorrent is a hardware company out of Toronto, Canada. This paper will detail the hardware design choices in the Grayskull that make it optimal for machine learning inference, along with the software design that facilitates compatibility with a large number of machine learning frameworks and models. The paper will end with tests of the Grayskull card on various machine learning models, comparing its performance against a Nvidia GPU and an Intel CPU.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.005 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.001 | 0.002 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.003 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.015 | 0.003 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".