Chip comparison
AMD Instinct MI250X vs Groq LPU (Language Processing Unit)
Side-by-side specs
Straight from each vendor's datasheet. Winner in bold: higher throughput, memory, or transistor count; lower power draw or price. Sparse (2:4) throughput roughly doubles the FP8/INT8 numbers shown.
| Field | AMD Instinct MI250X | Groq LPU (Language Processing Unit) |
|---|---|---|
| Process node | TSMC N6 | GlobalFoundries 14nm |
| TDP | 560 W | 375 W |
| Memory | 128 GB HBM2e | 0.23 GB SRAM (230 MB on-die) |
| Memory bandwidth | 3,200 GB/s | 80,000 GB/s |
| FP16 (dense) | 383 TFLOPS | 188 TFLOPS |
| BF16 (dense) | 383 TFLOPS | — |
| INT8 (dense) | 383 TOPS | 750 TOPS |
| Form factor | OAM | PCIe (GroqCard) / rack-scale GroqRack |
| Announced | 2021-11-08 | 2020-01-01 |
| Released | 2022-06-01 | 2021-06-01 |
Robots running AMD Instinct MI250X
No robots publicly running it yet.
Robots running Groq LPU (Language Processing Unit)
No robots publicly running it yet.
Data centers with AMD Instinct MI250X
No data centers publicly running it yet.
Data centers with Groq LPU (Language Processing Unit)
No data centers publicly running it yet.
Common questions
AMD Instinct MI250X vs Groq LPU (Language Processing Unit): which is faster for training?
AMD Instinct MI250X has 2.04x the dense FP16/BF16 throughput of the other (AMD Instinct MI250X: 383 TFLOPS; Groq LPU (Language Processing Unit): 188 TFLOPS). Real-world training speed also depends on memory bandwidth, interconnect, and software stack: see the spec diff for those inputs.
AMD Instinct MI250X vs Groq LPU (Language Processing Unit): which has more memory?
AMD Instinct MI250X carries more HBM (AMD Instinct MI250X: 128 GB HBM2e; Groq LPU (Language Processing Unit): 0.23 GB SRAM (230 MB on-die)). Memory capacity sets the largest model that fits on one chip without partitioning; memory bandwidth sets inference throughput.
AMD Instinct MI250X vs Groq LPU (Language Processing Unit): which is more power-efficient?
Dense FP16/BF16 TFLOPS per watt: AMD Instinct MI250X 0.68 (383 ÷ 560 W); Groq LPU (Language Processing Unit) 0.50 (188 ÷ 375 W). AMD Instinct MI250X wins at the die level; system-level efficiency also depends on cooling and interconnect.
See also: every chip comparison · AMD Instinct MI250X full page · Groq LPU (Language Processing Unit) full page.