Scoring methodology
GixCore scores are category-relative research signals. They help compare similar parts inside the current catalog; they are not universal benchmarks and do not predict an exact game, application, or build result.
Performance rank
CPUs, GPUs, RAM, and storage are evaluated on the axis they most directly control: compute, graphics rendering, memory throughput, or storage responsiveness. The public rank compares a part with other scorable parts in the same category. Equal underlying results share the same competition rank.
“Top 8% of CPUs,” for example, means the part's current catalog result ranks above roughly 92% of the scorable CPU population. It does not mean the processor is 92% faster than another product.
Benchmark and formula inputs
A clean benchmark-style value can lead CPU scoring when comparable coverage is broad enough across the category. If coverage is below 50%, the whole CPU population falls back to one specification-based method so benchmark and formula scales are not mixed unfairly.
Formula inputs vary by category. They can include cores, frequency, threading, graphics clock and bandwidth, memory generation and transfer rate, or storage class and rated speed. Architecture coefficients are category-level adjustments, not hand-tuned bonuses for individual products.
Capability is not the same as speed
Capacity and compatibility limits are kept separate from raw performance where possible. VRAM, RAM capacity, storage capacity, PSU wattage, sockets, and interfaces can determine whether a workload or build fits even when they do not make a component intrinsically faster.
Motherboards are enablers rather than compute devices, so their catalog signal is based on platform capabilities such as chipset tier, memory support, expansion, PCIe generation, storage slots, and connectivity—not a benchmark claim.
Value score
When reliable current pricing is available, value compares the part's category-specific capacity signal with its price and normalizes the result against the strongest current ratio in that category. A value score can change without the hardware changing because price and availability move.
Known limitations
- Catalog ranks change as products are added, corrected, or removed.
- Specification formulas cannot fully reproduce workload-specific benchmarks.
- Cooling, BIOS settings, memory configuration, drivers, and application behavior affect real systems.
- Scores from different component categories are not comparable to each other.