Habana Gaudi processor
E660956
The Habana Gaudi processor is a specialized AI training accelerator designed by Habana Labs (an Intel company) to deliver high-performance, scalable deep learning computation in data centers.
All labels observed (1)
| Label | Occurrences |
|---|---|
| Habana Gaudi processor canonical | 1 |
How this entity was disambiguated
This entity first appeared as the object of triple T7388251 — resolving that mention is where its identity was fixed. The disambiguator weighed these candidate entities and picked the highlighted one (or “None”, minting a new entity). This is how homonymy is resolved: the same surface form can point to different entities.
NED1
Entity disambiguation (via context triple)
gpt-5-mini-2025-08-07
Target entity: Habana Gaudi processor Context triple: [Goya inference processor, relatedProduct, Habana Gaudi processor]
-
A.
Apollo Lake platform
The Apollo Lake platform is Intel’s low-power system-on-chip family for entry-level PCs and embedded devices, built around the Goldmont CPU microarchitecture.
-
B.
Itanium
Itanium is a 64-bit server processor architecture developed by Intel (with early collaboration from HP) that was designed for high-end enterprise and technical computing but ultimately saw limited adoption and was discontinued.
-
C.
Snapdragon system-on-chip
The Snapdragon system-on-chip is a family of mobile processors widely used in smartphones and other devices, integrating CPU, GPU, modem, and other components to deliver high performance and power efficiency.
-
D.
A10 Fusion chip
The A10 Fusion chip is a 64‑bit ARM-based system-on-a-chip designed by Apple that significantly boosts performance and power efficiency for devices like the iPhone 7 and 7 Plus.
-
E.
Crusoe microprocessor
The Crusoe microprocessor is a low-power, x86-compatible CPU line from Transmeta that used code-morphing software to translate x86 instructions to an underlying VLIW architecture, targeting laptops and mobile devices.
- F. None of above. chosen
- G. Unsure - the case is ambiguous/there is not enough information to decide.
NED2
Entity disambiguation (via description)
gpt-5-mini-2025-08-07
Target entity: Habana Gaudi processor Target entity description: The Habana Gaudi processor is a specialized AI training accelerator designed by Habana Labs (an Intel company) to deliver high-performance, scalable deep learning computation in data centers.
-
A.
Apollo Lake platform
The Apollo Lake platform is Intel’s low-power system-on-chip family for entry-level PCs and embedded devices, built around the Goldmont CPU microarchitecture.
-
B.
Itanium
Itanium is a 64-bit server processor architecture developed by Intel (with early collaboration from HP) that was designed for high-end enterprise and technical computing but ultimately saw limited adoption and was discontinued.
-
C.
Snapdragon system-on-chip
The Snapdragon system-on-chip is a family of mobile processors widely used in smartphones and other devices, integrating CPU, GPU, modem, and other components to deliver high performance and power efficiency.
-
D.
A10 Fusion chip
The A10 Fusion chip is a 64‑bit ARM-based system-on-a-chip designed by Apple that significantly boosts performance and power efficiency for devices like the iPhone 7 and 7 Plus.
-
E.
Crusoe microprocessor
The Crusoe microprocessor is a low-power, x86-compatible CPU line from Transmeta that used code-morphing software to translate x86 instructions to an underlying VLIW architecture, targeting laptops and mobile devices.
- F. None of above. chosen
Statements (47)
| Predicate | Object |
|---|---|
| instanceOf |
AI accelerator
ⓘ
deep learning training processor ⓘ |
| announcedBy | Habana Labs NERFINISHED ⓘ |
| architecture | custom AI training architecture ⓘ |
| competesWith |
NVIDIA A100
NERFINISHED
ⓘ
NVIDIA H100 NERFINISHED ⓘ |
| connectivity |
RoCE-based Ethernet networking
ⓘ
high-speed on-chip fabric ⓘ |
| deploymentFormFactor |
PCIe card
ⓘ
mezzanine card ⓘ server system ⓘ |
| designer | Habana Labs NERFINISHED ⓘ |
| ecosystem | Intel AI portfolio ⓘ |
| feature |
HBM-based memory subsystem
ⓘ
matrix math engines ⓘ on-chip SRAM ⓘ programmable Tensor Processing Cores ⓘ |
| manufacturer | Habana Labs NERFINISHED ⓘ |
| optimizedFor |
TCO efficiency
ⓘ
large-scale training clusters ⓘ |
| ownedBy | Intel NERFINISHED ⓘ |
| parentCompany | Intel NERFINISHED ⓘ |
| powerOptimizedFor | performance-per-watt ⓘ |
| purpose |
AI training acceleration
ⓘ
deep learning training ⓘ |
| softwareStack | SynapseAI NERFINISHED ⓘ |
| successor | Habana Gaudi2 ⓘ |
| supports |
BF16
NERFINISHED
ⓘ
FP16 ⓘ INT8 ⓘ PyTorch NERFINISHED ⓘ TensorFlow NERFINISHED ⓘ data parallelism ⓘ distributed data loaders ⓘ large batch training ⓘ mixed-precision training ⓘ model parallelism ⓘ multi-node training ⓘ scale-out training ⓘ standard Ethernet networking ⓘ |
| targetMarket |
cloud service providers
ⓘ
enterprise data centers ⓘ |
| targetWorkloads |
NLP models
ⓘ
computer vision models ⓘ recommendation systems ⓘ |
| useCase |
cloud AI training
ⓘ
data center AI workloads ⓘ |
How these facts were elicited
The pipeline generated the facts above by prompting gpt-5.1 with this entity's name + description and the instruction below.
Instruction
You are a knowledge base construction expert. Given a subject entity and a description of it, return factual statements that you know for the subject as a JSON list of dictionaries(triples), where keys must be "subject", "predicate" and "object". The number of facts may be very high, between 25 to 50 or more, for very popular subjects. For less popular subjects, the number of facts can be very low, like 5 or 10. # Requirements - If you don't know the subject at all, return an empty list. - If the subject is not a named entity, return an empty list. - Include at least one triple where predicate is "instanceOf". - Do not get too wordy. - Separate several objects into multiple triples with one object.
Input
Subject: Habana Gaudi processor Description of subject: The Habana Gaudi processor is a specialized AI training accelerator designed by Habana Labs (an Intel company) to deliver high-performance, scalable deep learning computation in data centers.
Referenced by (1)
Full triples — surface form annotated when it differs from this entity's canonical label.