NVIDIA GeForce Go 6800

The NVIDIA GeForce Go 6800 is a mobile GPU from the GeForce Go 6 (Go 6000) generation, based on the NV41 chip and Curie architecture. TSMC fabricated the die on a 130 nm process, with 190 million transistors in a 225 mm² area, yielding a transistor density of 844.4K / mm². The release date is 2004-11-07, and the production status is end-of-life. In the benchmark database, the Go 6800 has no benchmark entries, an average benchmark score of 0, and a 50th percentile position among all GPUs. The nearestRivals list is empty.

Benchmark Performance

The most important result for the Go 6800 is the absence of results. The benchmark record contains no measured scores, so the average benchmark score of 0 represents a lack of samples rather than a measured level of performance. The 50th percentile is the only positional value in the database. That places the product at the midpoint of all GPUs in the database, but because no benchmark samples exist, the percentile cannot be tied to any actual frame-rate or compute result.

The hardware data provides a set of static throughput limits. The pixel rate is 2.400 GPixel/s, and the texture rate is 3.600 GTexel/s. These figures are produced by 12 TMUs and 8 ROPs. The memory side consists of 256 MB of GDDR3 on a 256 bit bus, with a memory clock of 550 MHz and an effective data rate of 1100 Mbps. The resulting bandwidth is 35.20 GB/s. Those values describe how quickly pixel, texture, and memory data can move, but the database does not contain an aggregate benchmark score that translates them into application performance.

No base, boost, or game clock is recorded for the GPU core. No shading-unit count is recorded, and no FP32 or FP16 throughput is recorded. As a result, the only compute-related numbers available are the pixel rate, texture rate, and memory bandwidth. The 2.400 GPixel/s and 3.600 GTexel/s figures are fixed-function processing ceilings. Without core clocks or shading units, the broader shader-level performance cannot be derived from this record.

The 50th percentile is a neutral ranking signal. It does not say that the Go 6800 is better or worse than any specific product. It simply indicates that the database places it in the middle of the distribution of all GPUs. Because the nearestRivals array is empty, there are no deltaPct values, no rival scores, and no rival names to cite. The benchmark performance of the Go 6800 must therefore be described as unmeasured in this database.

Power and Cooling

The thermal design power of the Go 6800 is 45 W. This is the only power figure in the record. The slot width is MXM Module, and the bus interface is MXM-III, indicating a modular mobile graphics implementation rather than a desktop expansion card. The power connectors field is None, so the module does not require external power cables. The database also lists no suggested PSU, meaning no recommended power-supply rating is available.

With no power connectors present, the MXM interface is responsible for both data and power delivery. The 45 W TDP is the limit that a host laptop design must accommodate. No dimensions are recorded for the module, and no cooler size can be inferred from the database. The power and cooling profile is thus defined by the 45 W envelope and the MXM form factor.

The absence of a suggested PSU is consistent with a mobile module. Desktop power-supply sizing is not part of the record. The display outputs are listed as Portable Device Dependent, which further indicates that the module is designed to be installed in a notebook whose display connections and power delivery are controlled by the host system.

How It Compares

The nearestRivals array in the database is empty. Therefore no rival comparisons can be made from the recorded benchmark data. There are no rival names, no rival scores, and no deltaPct percentages. The only contextual placement is the 50th percentile among all GPUs, which is too broad to identify a specific competitive relationship.

The product lineage provides two adjacent points. The predecessor is the GeForce FX Go 5, and the successor is the GeForce Go 7. The database does not list benchmark scores for either of these parts in the Go 6800 record, and neither is placed in nearestRivals. A generational relationship is therefore known, but a quantitative performance difference is not.

Because no nearestRivals are listed, no rival-by-rival performance paragraphs can be written. The empty rival list is itself a meaningful data finding: any statement that the Go 6800 is faster or slower than a named competitor would require benchmark data that this database does not contain.

Who Should Consider It

The Go 6800 is a 45 W MXM Module GPU with no auxiliary power connectors. That combination is aimed at notebook and portable systems where the MXM-III interface is available and where the host platform can supply power within a 45 W thermal envelope. The display outputs being Portable Device Dependent means the laptop determines the actual display configuration.

The performance guidance from the data is limited. The average benchmark score is 0, so no measured frame-rate or resolution recommendation can be made. The available throughput numbers are 2.400 GPixel/s, 3.600 GTexel/s, and 35.20 GB/s of memory bandwidth. Workloads that remain within a 256 MB memory footprint and fit within those fixed-function rates are the realistic candidates.

The 50th percentile ranking suggests a middle-of-the-database product, but with no scores to confirm what that means in real applications. Users should consider the Go 6800 if they need a modular MXM-III mobile GPU, if 256 MB GDDR3 is adequate for the intended workload, and if the absence of measured benchmark performance is acceptable. Users looking for current ray tracing acceleration or tensor acceleration should not consider it, because those features are not present in the record.

Ray Tracing and Feature Set

The Go 6800 has no ray tracing cores and no tensor cores recorded. It uses the Curie architecture, and its API support is DirectX 9.0c (9_3) with OpenGL 2.0 full and OpenGL 2.1 partial. Vulkan support is not listed, so the feature set ends before the Vulkan API generation.

The fixed-function units in the record are 12 TMUs and 8 ROPs. These are textured and pixel processing units, not dedicated ray traversal units or tensor units. The memory configuration of 256 MB GDDR3 on a 256 bit bus provides 35.20 GB/s of bandwidth, which feeds those units.

Without ray tracing cores, hardware-accelerated ray tracing is unavailable. Without tensor cores, hardware-accelerated tensor operations are unavailable. The recorded API support confirms the feature set belongs to the DirectX 9.0c and OpenGL 2.0/2.1 era. The partial OpenGL 2.1 entry means the implementation does not fully cover the OpenGL 2.1 specification.

The absence of Vulkan is also significant. Any application that requires Vulkan support cannot rely on this GPU. The feature set is entirely defined by the Curie-era API levels and the fixed-function units. For a modern feature evaluation, the Go 6800 is end-of-life, with no RT cores, no tensor cores, no Vulkan, and a maximum API level of DirectX 9.0c (9_3) plus partial OpenGL 2.1.