ATI All-In-Wonder 9200 SE

The ATI All-In-Wonder 9200 SE is an AGP 8x card built by AMD around the RV280 chip on a 150 nm process at UMC. It belongs to the Rage 7 architecture and the All-In-Wonder (9000) generation. The card uses 128 MB of DDR memory on a 64-bit bus; the memory clock is 164 MHz and the listed effective data rate is 328 Mbps, producing 2.624 GB/s of bandwidth. The rendering pipeline contains 4 TMUs and 4 ROPs, with a pixel rate of 1.000 GPixel/s and a texture rate of 1.000 GTexel/s. The product is end-of-life and was released on 2004-01-25. The database assigns it a 50th-percentile rating across all GPUs and an average benchmark score of 0.

Benchmark Performance

The benchmark record for this card is sparse. The `benchmarks` array is empty, and the `nearestRivals` array is empty, so no exact percentage deltas can be computed from the supplied data. There is no competitor score, no workload result, and no relative performance gap to report. The only two performance-level values in the record are the average benchmark score and the percentile rating. An average benchmark score of 0 means the database contains no positive aggregate benchmark result for this product. That absence matters: a zero score is not a measured frame-rate figure, and it should not be read as proof of a specific level of speed. The 50th percentile placement puts the card at the midpoint of the all-GPU distribution tracked by the database. That median position, however, is not supported by individual benchmark entries, so it stands as a rank rather than a tested performance level.

The equal pixel and texture rates are a notable detail. With 4 ROPs and 4 TMUs, the card is listed at 1.000 GPixel/s and 1.000 GTexel/s, meaning the two fundamental rendering pipelines have the same nominal throughput. This is a balanced but low-width configuration. A wider pipeline would normally show different, and higher, rate values. The data supplies no percentage deltas, so any comparative statement such as one card being ahead of another would be unsupported. The percentile value of 50 is the only relative landmark. It places the All-In-Wonder 9200 SE in the middle of the database’s GPU distribution, but without any actual score entries, that mid-position should be interpreted with caution. The combination of a zero average score and an empty nearest-rival list makes this part a performance unknown rather than a measured contender.

Power and Cooling

The fact pack does not list a TDP, so the card’s thermal design power is not quantified here. What the data does specify is a suggested 200 W power supply, a single-slot form factor, and no power connector. The absence of a dedicated power connector implies that the board is intended to draw its operating current from the AGP 8x slot alone. A suggested 200 W system supply is a deliberate power-system limit: it points to a low total system power budget rather than a high-current graphics card. For cooling, the single-slot width is the only physical constraint given. No heatsink or fan specifications appear in the record, so the data cannot describe the cooling solution beyond that slot-width statement.

Physical chip data helps place the power requirement in context. The RV280 die contains 36 million transistors on a 98 mm² die, fabricated on a 150 nm process at UMC. That combination yields a transistor density of 367.3K per mm². These figures are modest compared with the broader GPU field in the database, and they are consistent with a board that does not list an auxiliary power connector and only suggests a 200 W power supply. The lack of a TDP value is a clear gap: an integrator would have only the suggested 200 W PSU and the absence of a power connector to estimate the system impact. The single-slot design also means the card does not require extra expansion-space headroom for cooling. In short, the power and cooling story is defined by a 200 W system-level suggestion, a single-slot width, and the absence of both a TDP figure and a power connector.

Ray Tracing and Feature Set

The API list is short. The card supports DirectX 8.1 and OpenGL 1.4, and Vulkan is not listed. The `rtCores` and `tensorCores` fields are both null, so the data contains no hardware ray tracing resources and no tensor acceleration resources. Ray tracing must therefore be set aside as a feature for this product. The card’s rendering capabilities are expressed through its 4 TMUs and 4 ROPs, with the listed 1.000 GTexel/s and 1.000 GPixel/s rates. These are conventional rasterization resources rather than dedicated compute or ray tracing blocks. The All-In-Wonder family name points toward multimedia and display features, and the output configuration of 1x VGA and 2x S-Video reinforces that interpretation. The feature set is bounded by DirectX 8.1-era applications and OpenGL 1.4 software; there is no Vulkan path in the record.

For a benchmark database, the absence of RT cores and tensor cores is categorical. There is no field value indicating any ray tracing capability. Similarly, the null Vulkan entry means the card cannot be expected to support modern Vulkan workloads. The feature set is therefore anchored to two older API levels. DirectX 8.1 and OpenGL 1.4 define the software horizon. The display outputs add a video-oriented component: 1x VGA for a monitor and 2x S-Video for video-out connections. This is not a feature set built around post-processing, compute shaders, or ray traced effects. It is a rasterization-oriented, multimedia-focused product whose APIs and core counts are all from the same narrow slice of the database’s history.

Who Should Consider It

Any purchasing decision should start with three hard facts from the data: an average benchmark score of 0, an end-of-life production status, and a 64-bit memory bus. These do not point to a card for modern high-resolution gaming. The 1.000 GPixel/s pixel rate and 2.624 GB/s memory bandwidth provide little headroom for demanding rendering workloads, and no benchmark record exists to suggest otherwise. The 50th percentile all-GPU rating is the only favorable relative data point, but because the average score is 0, that percentile has no measured performance behind it. For someone running DirectX 8.1-class software on an AGP 8x system with a 200 W power supply, the card is structurally appropriate. The 128 MB DDR frame buffer can handle modest texture loads, while the 64-bit bus places a firm limit on sustained throughput.

The 1x VGA and 2x S-Video outputs make this a plausible choice for a multimedia PC that needs to drive a monitor and send video output to secondary displays. Users who want to run titles that require more than DirectX 8.1 or OpenGL 1.4 will need to look beyond this part because the API list is fixed. The card is likewise not a match for users who need ray tracing or tensor acceleration, since those fields are null. The data suggests a narrow usage profile: legacy AGP systems, older API software, and video-output tasks. The 200 W suggested PSU and single-slot design are practical constraints that fit such a system. In the end, the absence of benchmark scores means the recommendation cannot be grounded in measured frame-rate data; it must be grounded in the chip’s small memory bus, low fill rates, and end-of-life status.

Memory Subsystem

The memory subsystem is the most fully specified part of the product. The card has 128 MB of DDR memory on a 64-bit bus, with a bandwidth of 2.624 GB/s. The memory clock is 164 MHz, and the listed effective transfer rate is 328 Mbps. The 64-bit interface is the defining constraint. For any high-resolution workload, the frame buffer must move color, depth, and texture data quickly, and 2.624 GB/s places a tight cap on that movement. The memory clock and effective-rate values indicate a double-pumped DDR-style transfer scheme, but the narrow bus width limits how much effective rate can reach the rendering pipeline. In addition, 128 MB is a finite pool for geometry buffers, color buffers, depth surfaces, and textures.

The relationship between the memory clock, effective rate, and final bandwidth is important. A 164 MHz memory clock with 328 Mbps effective transfer is a low per-pin data rate. On a 64-bit bus, that effective rate results in 2.624 GB/s. This is not a configuration designed to feed a wide, high-speed pixel pipeline. The card’s pixel rate of 1.000 GPixel/s is equally modest, so the memory interface and the pixel rate are aligned in scale. Higher-resolution settings are constrained first by the 64-bit bus and second by the small 128 MB frame buffer. Lower resolutions and reduced texture sizes are the only practical way to stay within the memory bandwidth. Since the database contains no benchmark scores for this card, the memory subsystem remains the strongest quantitative predictor of its behavior at high resolutions. The combination of 128 MB, a 64-bit path, and 2.624 GB/s leaves little room for large, high-detail rendering surfaces.