Clore GigaSPOT Launches as Next-Generation GPU Compute Power Trading Platform

Clore GigaSPOT, a cutting-edge GPU compute power trading system, officially launches on January 3, 2025. This platform builds on its predecessor, the SPOT market, introducing enhanced features aimed at maximizing income for GPU server owners while catering to clients with computing needs for interruptible tasks.

A Self-Regulated Compute Power Market

The GigaSPOT platform operates as a self-regulated, permissionless market, balancing the interests of server owners and clients. By enabling users to bid on machines based on workload requirements, GigaSPOT optimizes power usage and task allocation for maximum efficiency.

Clients can request specific power limits for GPUs, ensuring that workloads are tailored to the available resources. For instance, workloads that are VRAM bandwidth-limited can benefit from reduced power settings without affecting performance. The system dynamically reevaluates workloads during each billing interval to prioritize the most profitable tasks.

Key Features of GigaSPOT

Customizable Workload Settings:

• Supports Docker images or predefined configurations.
• Users can set overclocking parameters, including core and memory locks, offsets, and power limits.

Flexible Bidding:

• Compute power is priced in $CLORE/day.
• Environmental variables and GPU-specific settings can be adjusted for optimal performance.

Power Efficiency:

• Clients can define power limits to align with workload requirements, reducing unnecessary energy consumption.

Documentation and Access

Comprehensive documentation is available for users to get started with GigaSPOT:

• Overview: Clore GigaSPOT Overview
• API Documentation: GigaSPOT API Docs

Initial Setup and Future Expansion

The initial deployment connects three test machines, with plans to expand availability across all machines shortly. Clore GigaSPOT aims to redefine GPU compute power trading by creating a transparent, efficient, and scalable market.

As the platform goes live, it invites GPU server owners and clients to dive into its features and begin bidding.